THE IMPACTS OF ADOLESCENT NONCOMPLIANCE ON CAREER PATHS IN TRANSITIONAL SOCIETIES
A DISSERTATION SUBMITTED TO THE DEPARTMENT OF SOCIOLOGY AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
Heili Pals September 2006
© Copyright by Heili Pals 2006 All Rights Reserved
ii
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.
________________________________ (Nancy B. Tuma) Principal Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.
________________________________ (Morris Zelditch Jr.)
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.
________________________________ (Daniel A. McFarland)
Approved for the University Committee on Graduate Studies.
iii
ABSTRACT In this study, I examine the effects of adolescent noncompliance on labor market outcomes in rapidly changing, unpredictable societies. Contrary to conventional wisdom, I argue that in societies with a high level of unpredictability, adolescent noncompliance to school authorities' rules has positive effects on labor market outcomes in adulthood. In particular, I argue that it tends to increase occupational status (in particular, the likelihood of being a manager) and the likelihood of being engaged in entrepreneurial activities. I also argue that in a society where rules and norms are fluctuating often and considerably, those who were noncompliant in adolescence fare better than those who were compliant because compliant youths find it more difficult to operate in a society with changing rules. As an example of rapidly changing, unpredictable societies, I examine several regions of the former Soviet Union during and after the collapse of the Soviet regime. I consider several labor market outcomes: 1) entrepreneurship before and after the collapse of the Soviet Union, 2) working in the private sector rather than in the public sector, 3) occupational status, 4) being unemployed, and 5) occupational status. I analyze these outcomes using longitudinal data from the Paths of a Generation study from five former Soviet regions: Belarus, Estonia, Latvia, Kharkiv, and Kurgan regions (sample size varies from 1,200 to 2,100 per region). These data allow adolescent noncompliance to be measured more than a decade before the labor market outcomes in adulthood. The results of my analyses indicate that the positive effects of
iv
adolescent noncompliance on adult labor market outcomes in unpredictable societies are largely due to the disadvantage of being compliant in adolescence. I also perform a variety of tests comparing the effects of adolescent noncompliance in settings with different levels of unpredictability. Adolescent noncompliance does not increase the likelihood of becoming a manager before the collapse of the Soviet Union. However, it does have a positive effect the likelihood of being a manager after the collapse of the Soviet Union. Similarly, adolescent noncompliance has a positive effect on occupational status in the private sector (a sector characterized by high unpredictability) but not in the state sector (a sector in which unpredictability is relatively low). These results are confirmed by a regional comparison: adolescent noncompliance has more consistent positive effects in the rapidly changing Baltic states than in the more slowly changing regions in Russia, Ukraine, and Belarus.
v
ACKNOWLEDGEMENTS In completing this piece of research, I would like to acknowledge the kind assistance of the following people: Nancy B. Tuma, my adviser, for her critical, yet always supportive guidance. For her invaluable comments on my numerous drafts, for teaching and advising me on quantitative methods, and for her editorial efforts. Daniel A. McFarland for his willingness to discuss my work for hours and for his extensive comments on my drafts. Morris Zelditch Jr. for his extensive help in developing the theoretical argument and getting started at the early stages of this research. David Grusky and Rebecca Sandefur for their role as my proposal defense committee members. Mikk Titma and Nancy B. Tuma for their generous allowance of this unique data. Without Paths of a Generation data this research would not have been possible. I am very thankful for all those people, especially regional leaders, who throughout the years worked hard to produce this fantastic dataset and wealth of information. Kristen Backor for her invaluable help with English and editing. My cohort, including Irena Stepanikova, Stefanie Möllborn, Jennifer van Stelle, Cynthia Brandt, Sean Everton, and Gabe Ignatow, for being there through the whole doctoral student years. In particular, I would like to thank Irena and Jen for
vi
helping me to get started and for their encouragement and support throughout the whole process of dissertation writing. Emily Borom and Norman Nie for their continued academic, financial, and day-to-day support. My gratitude goes out to Emily for making me feel well-cared for, especially during the final home-stretch. Staff at the Sociology Department for helping with literally any kinds of problems that came up during my doctoral studies. And finally, but most importantly, I want to thank all my friends and family who have supported me throughout my research. My gratitude goes out to my family for their never-ending patience and love from far away: Kodused – suur tänu teie alatise toetuse eest [In Estonian]. I thank Oscar for his continuous support from the very beginning until the completion of my dissertation. I am grateful for all my friends who supported and comforted me during the time of dissertation writing.
vii
TABLE OF CONTENTS CHAPTER 1: INTRODUCTION .......................................................................................1 Importance.....................................................................................................3 Goals of the Study .........................................................................................6 Organization of the Dissertation....................................................................6 CHAPTER 2: NONCOMPLIANCE AND LIFE CONSEQUENCES: A LITERATURE REVIEW ..10 Learning to Labor in Changing Society ........................................................10 Noncompliance and Related Concepts.......................................................... 14 The Concept of Noncompliance ...................................................................21 Noncompliance and Outcomes...................................................................... 24 Effects of Noncompliance Depend on the Social Context ............................28 CHAPTER 3: NONCOMPLIANCE AND LABOR MARKET OUTCOMES IN FORMER SOVIET REGIONS .....................................................................................................................32 Societies in Transition: The Former Soviet Regions..................................... 32 Generalizability ............................................................................................40 Noncompliance to School Authorities’ Rules and New Types of Employment.42 Noncompliance to School Authorities’ Rules and Occupational Status.............48 Status Differences in the Effects of Adolescent Noncompliance..................49 Structural Explanation: Noncompliance to School Authorities’ Rules......... 54 CHAPTER 4: DATA AND MEASURES ............................................................................60 Data................................................................................................................60 Adolescent Noncompliance to School Authorities’ Rules ............................62 Adult Outcome Variables ..............................................................................73 Control Variables...........................................................................................77
viii
Analysis Methods ..........................................................................................78 CHAPTER 5: EFFECTS OF NONCOMPLIANCE ON LABOR MARKET OUTCOMES .............80 Entrepreneurship Before and After the Collapse of the Soviet Union ..........80 Private vs. State Sector ..................................................................................97 Occupational Status .......................................................................................101 Adolescent Noncompliance and Not Having a Job.......................................107 Nonlinearity of the Effects of Adolescent Noncompliance...........................113 Dependence of Effects of Adolescent Noncompliance on Status .................121 Conclusion: Labor Market Outcomes ...........................................................126 CHAPTER 6: STRUCTURAL DIFFERENCES IN THE EFFECTS OF ADOLESCENT NONCOMPLIANCE .......................................................................................................129 Over-Time Differences: Occupational Status of the First Job ...................... 129 Differences in the Effects of Adolescent Noncompliance by Economic Sector 132 Differences in Effects of Adolescent Noncompliance across Regions..............138 Conclusion: Structural Differences ...............................................................149 CHAPTER 7: CONCLUSION AND DISCUSSIONS .............................................................151 Conclusion and Discussions ..........................................................................151 Summary of Results ......................................................................................154 Strengths of the Study ...................................................................................159 Further Research and Applications ............................................................... 161 APPENDIX I: English Translation of the Questions ...................................................165 APPENDIX II: Note about Multinomial Logistic Regression .....................................168 REFERENCES ...............................................................................................................172
ix
LIST OF TABLES Table 3.1. Summary of Hypotheses ..........................................................................59 Table 4.1. Measures of Noncompliance to School and Family Authorities’ Rules .. 63 Table 4.2. Adolescent Noncompliance by Region ....................................................64 Table 5.1. ML Estimates of Binary Logistic Regression Model of Participation in the Second Economy Before 1992 ............................................................................82 Table 5.2. ML Estimates of Binary Logistic Regression Model of Owning a Firm in 1997-1998 and Ever Being Self-employed since 1992 .......................................84 Table 5.3. ML Estimates of Multinomial Logistic Regression Model of Economic Activity in 1997-1998, Effects Relative to “Employees” ...................................90 Table 5.4. ML Estimates of Ordered Logistic Regression Model of the Number of Types of Entrepreneurial Activities Participated in ............................................94 Table 5.5. ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, Effects Relative to the State Sector................................... 99 Table 5.6. Estimated Effect of Adolescent Noncompliance on the Odds of Working in the Private Sector.................................................................................................100 Table 5.7. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998, Effects Relative to Workers ..............................................103 Table 5.8. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998 Controlling for Occupational Status of First Job before 1992, Relative to Workers.............................................................................................105 Table 5.9. Break-down of No Job Category..............................................................109
x
Table 5.10. ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, Effects Relative to the State Sector................................... 110 Table 5.11. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998, Relative to Workers...........................................................112 Table 5.12. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Binary Logistic Regression Models of Owning a Firm in 1997-1998 and Ever Being Self-employed since 1992.........................................................................116 Table 5.13. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Economic Activity in 1997-1998, Effects Relative to “Employees” ......................................................................... 117 Table 5.14. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, Effects Relative to State Sector ...........................................................................118 Table 5.15. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Occupational Status of First Job before 1992, Relative to Workers........................................................................119 Table 5.16. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998, Effects Relative to Workers................................................................................. 120 Table 6.1. ML Estimates of Multinomial Logistic Regression Model of Occupational Status of the First Job before 1992, Relative to Workers....................................131 Table 6.2. ML Estimates of Multinomial Logistic Regression Model of Occupational Status by Economic Sector, Relative to Workers................................................134
xi
Table 6.3. Summary of Regional Differences in the Effects of Noncompliance....139 Table 6.4. ML Estimates of Binary Logistic Regression Models of Entrepreneurial Activities, by Region Group................................................................................141 Table 6.5. ML Estimates of Multinomial Logistic Regression Model of Economic Activity in 1997-1998, by Region Group, Effects Relative to “Employees”......142 Table 6.6. ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, by Region Group, Effects Relative to State Sector...........144 Table 6.7. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in the Baltic States, Relative to Workers. ................................................. 147 Table 6.8. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in Belarus, Kharkiv, and Kurgan, Relative to Workers ............................148 Table 7.1. Summary of Hypotheses ..........................................................................155 Table A1. Hausman-McFadden and Small-Hsiao Tests for the Independence of Irrelevant Alternatives .........................................................................................169
xii
LIST OF FIGURES Figure 3.1. The Effects of Adolescent Noncompliance on Labor Market Outcomes, As Dependent on Structural Conditions ........................................................40 Figure 4.1. Item Characteristic Curves...................................................................... 71 Figure 4.2. Total Information of the Adolescent Noncompliance Scale ................... 72 Figure 4.3. Distribution of the Adolescent Noncompliance Scale ............................73 Figure 5.1. Predicted Probability of Participation in the Second Economy and Owning a Firm in 1997-1998, by Adolescent Noncompliance...................................87 Figure 5.2. Predicted Probability of Self-Employment after 1992, by Adolescent Noncompliance..............................................................................................88 Figure
5.3.
Predicted
Probability
of
Economic
Activity,
by
Adolescent
Noncompliance..............................................................................................91 Figure 5.4. Predicted Probability of the Number of Entrepreneurial Activities, by Adolescent Noncompliance........................................................................... 96 Figure 6.1. Predicted Probability of Occupational Attainment by Economic Sector in 1997-1998 and Adolescent Noncompliance..................................................136 Figure 6.2. Distribution of Occupational Status in 1991 and 1997-1998..................145
xiii
THE IMPACTS OF ADOLESCENT NONCOMPLIANCE ON CAREER PATHS IN TRANSITIONAL SOCIETIES In this study, I take a new look at noncompliance to school authorities’ rules and the effects of that noncompliance on the life course. The sociological and socialpsychological literature routinely connects adolescent noncompliance with negative life outcomes, such as dropping out of school, low educational and occupational outcomes, and criminal and antisocial behavior in adulthood. In contrast, this study connects noncompliance to school authorities’ rules1 with positive labor market outcomes in adulthood. The main argument of this research is that the effects of noncompliance to authorities’ rules on life outcomes may vary depending on the fluctuations in society. The common view about the negative effects of noncompliance and related concepts applies to stable societies where the conditions and future are relatively predictable. I argue that adolescent noncompliance can have the opposite effects in societies with a high level of societal changes. In particular, I argue that conditions in rapidly changing societies are suitable for the positive effects of adolescent noncompliance because structure, norms, and laws are continually changing. Under these societal conditions, noncompliance to authorities’ rules may operate as a proxy for being relatively
1
Here and after, I use "noncompliance" to refer to noncompliance to rules of school
authorities or parents unless otherwise noted. I do not use "noncompliance" to refer to violation of laws or rules of political authorities.
© Heili Pals September 2006
1
Chapter 1: Introduction independent of society’s conventions, allowing an individual to adapt more readily to societal changes. I use the former Soviet states during the transition from a command economy as an example of societies in which those who were noncompliant as adolescents are hypothesized to fare better than those who were compliant as adolescents. I focus on several labor market outcomes, including participating in entrepreneurial activities and working in the private or state sectors, as well as occupational status more generally. In the following sections, when referring to labor market outcomes, I mean all of these activities. I analyze the effects of noncompliance to school authorities’ rules on labor market outcomes from both an individual and a structural perspective. On the individual level, I expect stronger positive effects of noncompliance to authorities’ rules for some labor market outcomes than others. In particular, I hypothesize that noncompliance is more advantageous for less regulated career paths such as new types of employment. I compare the effects of noncompliance to authorities’ rules in different regions and in different economic sectors of former Soviet states, because the extent of change has differed across economic sectors and regions. I argue that the positive effects of noncompliance to authorities’ rules are stronger in societies with faster and more extensive changes. I also compare the effect of noncompliance for high and low status people.
© Heili Pals September 2006
2
Chapter 1: Introduction IMPORTANCE The argument that the effects of noncompliance depend on social context is important in multiple domains. First, the argument developed here clarifies the micromacro connection between the structure of society and individual behavior; it recognizes variations in attainment based on social context. Second, the argument developed here explains attainment mechanisms particular to transitional societies and contributes to the life course literature by linking early experiences with adult outcomes. Furthermore, this study alerts deviance researchers to the crucial importance of separating different types and levels of noncompliance because they can have very different impacts on adult outcomes. The importance of this research lies partly in its contribution to the discussion about the interconnectedness of macro structures and individual behavior. I demonstrate that the relationship between noncompliance to school authorities’ rules and later life outcomes is greatly affected by structural factors. In particular, I argue against the general belief, challenged only by a few (Becker 1963; Phillips & Zuckerman 2001), that noncompliance to authorities’ rules is something inherently negative that always has dampening effects on adult outcomes, independent of the structural context. I demonstrate that attainment mechanisms are not the same in a stable society with a relatively high level of predictability and in a society that is undergoing rapid social changes and that has fluctuating norms and open economic niches. Generalizations from this study of transitional societies might also help us to understand the attainment processes in other types of “innovative” societies, such as societies with an abundance of open niches and vacancies, and societies or fields with
© Heili Pals September 2006
3
Chapter 1: Introduction less institutionalized structures, where the rules and norms are not fully fixed. Such structures could include new industries, new firms, new economic areas, rapidly developing countries, and so on. In this study, I apply arguments about noncompliance in different contexts to the former Soviet societies. An extensive literature has studied attainment in transitional societies moving from a command economy (e.g., Böröcz & Róna-Tas 1995; Eyal, Szelényi, & Townsley 1998; Futo, Hoggett, & Kallay 1997; Gerber 2001a, 2001b, 2002; Róna-Tas 1994; Shkaratan 1992; Slomczynski & Tomescu-Dubrow 2005; Titma, Tuma, & Silver 1998; Titma, Tuma, & Roosma 2003; Verhoeven, Jansen, & Dessens 2005). This literature focused mostly on traditional attainment factors, such as education, parental background, gender, school type, and urbanization. Less traditional attainment mechanisms particular to these types of societies are considered less often.2 This study tries to fill this gap in the literature and consider other attainment mechanisms particular to these types of societies. This research adds to the existing literature on the life course in several other ways. First, one of the major aims of the life course literature has been to evaluate whether and how much someone’s life depends on his or her earlier experiences and behaviors. As in other life course studies, this research claims that experiences and
2
Communist Party membership is one exception that is relatively well-studied during
the Soviet Union (Titma, Tooding, & Tuma 2004), as well as after its collapse (Wong 2004, 2002; Hanley & Treiman 2005; Titma, Tuma, & Roosma 2003; Gerber 2000, 2001a, 2002; Kryshtanovskaya & White 1996; Rona-Tas 1994).
© Heili Pals September 2006
4
Chapter 1: Introduction behaviors from adolescence do affect life in adulthood. I find that a person’s life is interconnected over time. Secondly, starting with Elder (1974, 1986, 1987), the life course literature is concerned not only with the interconnectedness of early and later life, but to a greater extent with the interconnectedness of time, place, and individual life outcomes. Contributing to this body of literature, I demonstrate that the effects of early behavior on later labor market outcomes are shaped by the structural conditions in which a person lives. I find that structural conditions may affect not only the existence of this relationship, but also the nature of the relationship between early and late experiences. Thus, I find that adolescent noncompliance to school authorities’ rules can have positive effects during times of rapid social change, as opposed to the dampening effects under stable structural conditions found by previous literature. Thirdly, like the literature on deviance and criminology, the life course literature has been concerned about the effects of deviant behavior on individual social development. The general finding has been that engagement in any deviant behavior has negative effects on various adult outcomes. In this research, I argue, however, that not every type of deviant behavior in adolescence results in lower attainment in adult life. Noncompliance to the rules of authorities can result in positive labor market outcomes in adulthood if society is undergoing extensive societal changes. Therefore, life course studies should differentiate between deviance and noncompliance, as their effects on later life can be quite different. This applies to the research on deviance and criminology, where it is common to concentrate solely on more serious types of deviance and not to distinguish them from noncompliance to the rules of authorities.
© Heili Pals September 2006
5
Chapter 1: Introduction Thus, this research makes multiple contributions, both to theoretical thinking and to the specific fields of transitional societies, the life course, and deviance research. GOALS OF THE STUDY This research is motivated by different fields in sociology, including research on attainment mechanisms, deviance, the life course, adolescent development, social psychology, and transitional societies. The aim of this study is to demonstrate that attainment mechanisms can differ in different structural conditions. I do this by developing a theoretical framework for studying the interconnectedness of society and the individual. Using the example of transitional post-communist societies, I consider the types of societal changes that influence individual attainment. I then propose a theoretical model that identifies the relationship between the level of social fluctuations and the effects of individual noncompliance to school authorities’ rules on labor market outcomes. The empirical part of this research deals directly with regions of the former Soviet Union and labor market outcomes in these regions. It contributes to the research on attainment in post-communist transitional societies and provides a variety of empirical tests of the argument proposed in this study that the effects of noncompliance depend on the social context. ORGANIZATION OF THE DISSERTATION This study is divided into seven chapters. In Chapter 2, I develop a theoretical framework for conceptualizing noncompliance and lay the groundwork both for the development of my theoretical argument and the actual empirical measurement of
© Heili Pals September 2006
6
Chapter 1: Introduction noncompliance. The conceptualization of noncompliance is followed by an overview of the various ways that types of noncompliance and related concepts have been defined in the human development and deviance literature. The third section of Chapter 2 reviews the literature connecting different types of noncompliance and nonconformity with life outcomes in both sociological and social psychological research. In the final section of Chapter 2, I outline a general framework for the argument that noncompliance depends on social context and connect the level of societal changes with the effects of noncompliance on individual life. I propose that the effects of noncompliance to authorities’ rules on life outcomes are not the same across societies with different levels of societal changes. Chapter 3 applies the general argument of effects of noncompliance to a particular context: the transitional former Soviet societies that are changing from a command economy. First, I describe the changes in these societies in detail. I divide the changes into three major dimensions that directly contribute to the effects of noncompliance to school authorities’ rules on later life: 1) fluctuations in the legal structure, 2) changes in socially accepted norms, and 3) changes in the opportunity structure. Second, I state hypotheses about the effects of noncompliance on work in new types of employment opportunities, those that were only loosely formed when the transition began. I follow this with hypotheses regarding the effects of noncompliance on occupational status in general. Next, I propose hypotheses regarding the effect of noncompliance based on the status of the individual. The chapter concludes with hypotheses about the structural aspects of the effects of noncompliance; I compare the
© Heili Pals September 2006
7
Chapter 1: Introduction relationship between noncompliance and attainment in structures with different levels of societal changes. The empirical work of this research is introduced in Chapter 4; it starts with a description of the data used in this study and continues with a description of the measurement of variables. I describe the empirical conceptualization of the index of noncompliance to the rules of authorities, introduce the labor market outcome variables, and list the various control variables used in the study. Chapter 4 concludes with a brief overview of the methods of analysis used in this research. In Chapter 5, I discuss the empirical tests of the hypotheses about the effects of noncompliance on different labor market outcomes. First, I discuss the effects of noncompliance on new types of employment opportunities, such as entrepreneurship before the collapse of the Soviet Union and various measures of entrepreneurship after the regime changed. Secondly, I compare the effects of noncompliance on the chances of working in the state versus the private sector. I continue to consider the hypotheses regarding the general effects of noncompliance on occupational status. I test the effects of noncompliance on occupational attainment both with and without controls for initial attainment. Furthermore, I discuss the potential negative effects of noncompliance. I find empirically that noncompliance to authorities’ rules does not increase unemployment during the post-Soviet transition. Finally, I discuss and examine the linearity of the effects of noncompliance to school authorities’ rules and consider the interaction between status and noncompliance. Chapter 6 reports the results of three direct empirical tests of how the effects of noncompliance depend on societal context. The first test considers time: I estimate the effects of adolescent
© Heili Pals September 2006
8
Chapter 1: Introduction noncompliance on the occupational status of the first job before the collapse of the Soviet Union. These results are compared to the results from Chapter 5 using the occupational status after the Soviet Union's collapse as a dependent variable. Secondly, I compare the effects of noncompliance to school authorities’ rules on occupational status in the state and private sectors. As a third test, I compare the effects of noncompliance on labor market outcomes in different regions of the former Soviet Union. In Chapter 7 I review the strengths and limitations of this study, summarize the empirical findings of chapters 5 and 6, and offer a discussion and conclusion. I propose various further applications of the argument that effects of noncompliance depend on the social context.
© Heili Pals September 2006
9
CHAPTER 2: NONCOMPLIANCE AND LIFE CONSEQUENCES: A LITERATURE REVIEW In this chapter, I conceptualize noncompliance and give an overview of various definitions of similar concepts in the literature. I also give an overview of the previous literature connecting nonconformity and noncompliance to life outcomes. The chapter concludes with the development of a general argument that the effect of noncompliance to authorities’ rules depends on social context. The aim in this section is to analyze whether the social context influences the effects of noncompliance on life outcomes, and, if so, to what extent. LEARNING TO LABOR IN CHANGING SOCIETY In his classic work about 1970s secondary school boys in Wales and England, Willis (1977) shows how rejection of school authorities leads to the reproduction of the working class. He describes working class students who, in their struggle to escape their class status, resist the rules of school authorities; through this noncompliance, they create a so-called counter-school culture. They miss homework, argue with teachers, play tricks on teachers and school administrators, drink alcohol, talk and brag about having sex, and find other ways to oppose the school’s institutionalized rules. They oppose the rules of school authorities in an effort to show that they are not worse or lower in status than their teachers. The irony is, according to Willis (1977), that the lads (the term working-class students use to describe themselves) are bright enough to succeed in life and gain positions better than working-class jobs. However, since they
© Heili Pals September 2006
10
Chapter 2: Noncompliance and Life Consequences: A Literature Review have chosen to resist school authorities’ rules, they end up with less education and working-class jobs similar to those of their parents. The question emerges whether a similar pattern of class reproduction would occur in a context of major societal change. Willis describes working-class values in a stable capitalist society. In this society, those opposing institutionalized rules are often penalized and end up with lower level positions. Consequently, working class youths who rebel in order to exert autonomy in educational settings end up reproducing their parents’ working-class position. However, what would happen if England went from a capitalist to a communist society? In such a case, noncompliance to rules in the capitalist society might win the lads a higher status position. But what if England was already communist? Then such opposition is no longer rebellious. Perhaps they would need to embrace Western culture to win position in transition from a communist to a capitalist society. The change in the socio-economic system of the society could bring about two important changes for those who are noncompliant. First, it would change the meaning of their noncompliance. While in a stable society their noncompliance is not valued, the values of working-class students’ noncompliance to the rules of school authorities might become heroic and even prophetic in a way in a society that is in transition. Thus, the value of their noncompliance would change from negative to positive. Second, by resisting the norms of the prior system, the lads are more prepared to adapt to the norms and values of the new system.
© Heili Pals September 2006
11
Chapter 2: Noncompliance and Life Consequences: A Literature Review An analysis of noncompliance to school authorities’ rules in the former region of the Soviet Union would give a proximate answer to the hypothetical questions posed above. However, before parallels can be drawn between Willis’ working-class rebellion in English secondary schools and noncompliance to school authorities’ rules in the former Soviet Union, noncompliance needs to be placed in context. The Soviet educational system, including secondary education, was centralized and governed by clear and consistent rules. These rules ranged from a centralized curriculum to uniforms and to clear norms of behavior. Schools were highly hierarchical, and the rules of school authorities were not challenged openly. Uniforms were required at every level of school, including primary, middle, and secondary school. If a student missed an item (e.g., left a hat at home) a punishment was expected (the student was sent back home to get the hat). Similarly, skipping school or being late was not tolerated. Students were not permitted to speak up in class or argue with teachers; they were usually allowed to speak only at the teachers’ request. If a student wanted to speak, either by her own initiative or as a response to a teacher’s question, she had to signal this by lifting her arm. She was then allowed to speak only after being given a clear acknowledgement by the teacher. The teacher was an authority figure. As a representative of the institution, the teacher was always correct. All of the rules were essentially suppressing students’ individuality. Such suppression went as far as regulating students’ movements during breaks between classes. Often schools required students to move in a circle during the break under the supervision of a teacher: two students at a time, like fish swimming in schools.
© Heili Pals September 2006
12
Chapter 2: Noncompliance and Life Consequences: A Literature Review Thus, noncompliance to school authorities’ rules in Soviet times incorporates different aspects of opposition. It is a noncompliance to authorities’ rules, to the higher-ups setting boundaries for students. It is an opposition to a collective mentality and preference for individualism. It is an opposition against structure and obedience. It is an opposition to hierarchy and clear institutionalized borders. Noncompliance to school authorities’ rules as described here was not and cannot be classified as noncompliance to the rules and laws of the communist system. It is likely that students who were noncompliant to the rules of school authorities in the Soviet Union never thought about the socioeconomic system of a communist society or defined their own actions as actions against the ruling regime. Nevertheless, their noncompliance was targeted at the institutionalized authority of teachers and schools. In this sense, students not complying with school authorities’ rules during Soviet times were expressing their individuality. They were looking to abolish institutionalized borders and rules and set their own. They were looking for a time when they would not be told what to do, but could choose their own activities. Thus, it was a partial rejection of the rules of established institutions. The collapse of the Soviet system led to the collapse of many of the similar authority structures and rule sets that students opposed in secondary school. In turn, their noncompliance to the rules of authorities prepared them for the disappearance or weakening of authority and rule structures after the collapse of the Soviet regime. With the collapse of the Soviet regime, noncompliance to authorities’ rules became the very characteristic that was valued.
© Heili Pals September 2006
13
Chapter 2: Noncompliance and Life Consequences: A Literature Review Even if individualism was not a conscious aim of noncompliant adolescents, it did prepare them for the collapse of authority structures and confusion in rules regarding what was allowed and what was not. In this sense, one could draw parallels to Willis’ description of working-class lads and could imagine what might happen if, after lads graduated from school, their society experienced a revolution that changed the value of their prior opposition. If the subsequent change in society would have changed the value attached to their opposition from something generally negative to something generally positive, then the lads’ outcomes might have changed as well. It might have been easier for the lads to gain higher outcomes in this situation, rather than maintain their working-class status. NONCOMPLIANCE AND RELATED CONCEPTS To further analyze whether adolescent noncompliance to school authorities’ rules has positive effects in a changing society, I first discuss different usages of noncompliance and related concepts in the existing literature. I then conceptualize noncompliance as I use it in this research. After that, I give an overview of the literature discussing adolescent noncompliance and life outcomes. The last section of this chapter links the notion of noncompliance to changing societies by describing the general argument that the effect of noncompliance depends on social context. Willis describes the lads’ opposition to authority as a rebellion expressed as resentment toward the rules and institutions that stops just short of direct confrontation. Opposition often takes the form of behaving in an informal fashion when formal relations are the norm. The lads did not want to be excluded from the
© Heili Pals September 2006
14
Chapter 2: Noncompliance and Life Consequences: A Literature Review school system, but they did want to make clear that they opposed the system and would not comply with its rules. The lads opposed not only teachers and school administrators, but also those conforming to the institutionalized school culture, the so-called ear-holes. Willis’ conceptualization of noncompliance is not the only one in the sociological and psychological literatures. Both fields have reported a considerable amount of research on noncompliance and related concepts. In this section, I briefly describe the various concepts and definitions of deviation from norms used in different fields and approaches. In the next section, I give an overview of research focusing on the long-term effects on the life course of adolescent noncompliance to social norms and rules. The terminology used to denote deviation from rules has been inconsistent. Terms such as nonconformity, rebellion, deviance, problem behavior, and noncompliance have often been used with little clarity of definition and without much justification for the choice of one term rather than another. I first review the social psychological literature on nonconformity, and then discuss various terms denoting deviation from rules in the sociological literature in general and youth research in particular. Nonconformity and Social Psychological Literature. The term nonconformity has been widely used in the social psychological literature. Early social psychological literature often conceptualized nonconformity as noncompliance to the opinions of others rather than as noncompliance to rules. Classic examples are the experimental studies by Asch (1952; 1956), Sherif (1936), Crutchfield (1955), and Milgram (1969)
© Heili Pals September 2006
15
Chapter 2: Noncompliance and Life Consequences: A Literature Review on obedience to an authority and agreement with a group. Asch (1952) developed a situation in which subjects had to provide a judgment while being confronted with consistently incorrect statements from group members. Conformity scores were obtained by counting the number of times the subject “went with the flow” and gave the obviously wrong answer supported by the group. Although this experiment took slightly different forms over the years (Crutchfield 1955; Milgram 1969), the essence of the experiment and the conceptualization of nonconformity remained constant. Milgram’s (1969) study went so far as to show that some participants would continue to agree with an authority figure, even when they believed they were physically harming or even killing other humans. Several studies connected nonconforming behavior with certain personality traits. Conformers showed intolerance for ambiguity, while nonconformists claimed to enjoy going against the rules. Nonconformists also rated higher in leadership ability and intellectual competence (Crutchfield 1955). Crowne and Liverant (1963) found that conformists have less self-confidence in an evaluative situation. My focus, however, is on noncompliance to rules of institutionalized authorities. Therefore, the conceptualization of nonconformity in the social psychological literature does not fit the logic of this study well. Nonconformity and Sociological Literature. In the sociological literature, nonconformity is usually defined as deviance from commonly held norms and rules rather than noncompliance to a group’s view as defined by social psychologists. Thus, this conceptualization is closer to the concept used here. Conformity and nonconformity have been a focus of sociological research starting with Durkheim
© Heili Pals September 2006
16
Chapter 2: Noncompliance and Life Consequences: A Literature Review (1915). Similar concepts were developed further by Merton (1968) and his contemporaries. Durkheim asserted that deviance is an integral part of all societies and serves four major functions: 1) affirming cultural values and norms, 2) clarifying moral boundaries, 3) promoting social unity, and 4) encouraging social change. The last function refers to rebels and a society’s flexibility to accommodate change initiated by those opposing the existing norms of a society. Merton (1968) described five different types of deviation that occur most often when there is a discrepancy between the socially structured capacities of members of the group to act in accord with them and cultural norms and goals. Two forms of deviance described by Merton are relevant here: innovation and rebellion. He defined innovation as accepting cultural goals (e.g., money), but rejecting legitimate means to those goals in favor of illegitimate means (e.g., theft, embezzlement, fraud). Thus, Merton’s innovation is similar to the more general definitions of deviance used in the criminology and sociological research. Merton described rebellion as the intention to change the norms: it involved rejecting both legitimate means and legitimate goals. Alienated youth gangs or subcultures are examples of rebellion. If rebellion “becomes endemic in a substantial part of society, it provides a potential for revolution” (Merton 1968: 245). Thus, both Durkheim and Merton discuss deviance and rebelliousness in terms of intent to change, but not necessarily in a positive direction. The functionalist theory that grew out of the works of Durkheim and Merton, however, argued that those going against the rules are a driving force of society (see Scull 1988 for a review). The main reasoning is that positive and negative deviance exist, both of which ignore the rules
© Heili Pals September 2006
17
Chapter 2: Noncompliance and Life Consequences: A Literature Review of the majority. Negative deviance is widely analyzed in criminology and in the deviance literature (Cleckley 1976; Wilson & Herrnstein 1985), while the positive aspects of deviance have been considered less often. The notion of positive deviance (Heckert & Heckert 2002; Spreitzer & Sonenshein 2004; West 2004) emerged in the late 1980s. It is one of the few lines of research that has focused on the positive aspects of deviance. The literature on positive deviance suggests that both negative and positive deviance exists, forming the tails of a continuum of normative behavior. The positive tail of the deviance continuum is conceived of as extreme conformism. Ultra-conformity refers to following the rules of society to a greater extent than expected. An example of this kind of positive deviance is Mother Teresa, who took the norm of helping others to an extreme. Another type of continuum defines deviance based on the type of sanctions that could be applied: depending on the social structure, deviance might draw negative or positive sanctions (Spreitzer & Sonenshein 2004). That is, criminal behavior draws negative sanctions and usually has negative outcomes, while entrepreneurialism or intentional drive to change society might draw positive sanctions and yield positive outcomes. Phillips and Zuckerman (2001) applied the framework of conformity to Silicon Valley firms and analysts making recommendations to their company CEOs. They argued that conformity is related to status. They claimed that middle-status actors are more likely than low- or high-status actors to be conformists. Because high-status actors feel confident in their acceptance, they can deviate with little fear of sanctions. Low-status actors deviate because they feel excluded regardless of sanctions.
© Heili Pals September 2006
18
Chapter 2: Noncompliance and Life Consequences: A Literature Review Terminology in Adolescent Literature. The literature described above focuses on nonconformity in general. Sociological research on adolescent rule-breaking has used a wide variety of terms noting different aspects and levels of deviation from norms. Jessor, Donovan, and Costa (1991) developed a psycho-social framework. They used Problem-Behavior Theory to analyze whether involvement in problem behavior in adolescence determines outcomes in young adulthood. They conceptualized problem behavior as involvement in problem drinking, marijuana use, use of other illicit drugs, smoking, and other forms of deviant behavior such as damaging public or private property, giving false information, shoplifting, starting fights, lying, writing bad checks, breaking into places, and stealing. Their conceptualization is rather similar to common definitions of adolescent deviance. A very similar conceptualization of deviance has been used by criminologists, sociologists, and adolescent researchers, who have considered it to include stealing, inflicting damage to property, under-age drinking, drug use, being arrested by police, and so on. Both problem behavior and deviance, as defined above, often focus on criminal activity and marked deviation from norms. While both problem behavior and deviance involve going against norms, both also involve an attempt to hide one’s actions from public knowledge. The terms rebellion and resistance, in contrast to conceptualizations of deviance and problem behavior, involve a more open disobedience to the rules of authority. While the extent of the disobedience is less (i.e., resistance usually does not involve criminally punishable actions), the actor usually does not try to hide his or her disobedience from authorities. Rettig (1979) distinguished between rebellion and
© Heili Pals September 2006
19
Chapter 2: Noncompliance and Life Consequences: A Literature Review deviance. He identified deviance based on whether youngsters have been officially apprehended by the justice system. Thus, delinquency is activity that has come to the attention of social control agencies and resulted in legal attention to the offender (legal-behavioral perspective). Based on Rettig (1979), rebellion is rule breaking that has not led to legal consequences. Rafky (1979) measured school rebellion based on the frequency of self-reported school vandalism and confrontations (participation in student strikes or walkouts, participation in a school sit-in or riot, physical confrontation with a teacher, damaging school property in excess of $100). Stinchcombe (1964) used the terms high school rebellion and nonconformity interchangeably. He measured rebellion by an index of skipping school, getting a flunk notice, and having been sent out of class. McFarland (2004) described resistance in a classroom setting. This concept is closely related to school rebellion and adolescent nonconformity. Acts of resistance that McFarland (2004) recognized include: 1) passive resistance, making jokes and voicing complaints or criticism about the teacher or tasks, 2) active resistance that goes beyond passive criticism and is made without regard for the teacher’s authority, such as overt challenges voiced in class that breach school norms. McFarland (2004) therefore defined resistance in terms of intentional classroom interactions aimed at the transformation of the interaction between the student and the teacher. In a separate piece, McFarland (2001) examined the causes of student resistance and found that students with high social standing (i.e., popular students) were more likely to manifest student resistance. Students with high academic status (i.e., high GPA) were less likely to manifest student resistance.
© Heili Pals September 2006
20
Chapter 2: Noncompliance and Life Consequences: A Literature Review THE CONCEPT OF NONCOMPLIANCE Adolescent Noncompliance. Clearly, the conceptualization of noncompliance depends on the context. There are two clear reasons why I concentrate on adolescent noncompliance rather than adult noncompliance or nonconformity in this research. The first relates to the context. The two institutions serving as authority structures for adolescents, the school and the family have relatively universal rules and they define opposition in a relatively similar fashion across schools and families. Adult roles and the norms guiding adult roles are much more varied. Furthermore, adult roles are less often governed by a universal authority structure as school and family are for children. This fact makes a relatively uniform definition of noncompliance much harder for adults than for children and youths. Therefore, to be more parsimonious, I have elected to focus on the concept of adolescent noncompliance. The second reason for focusing on adolescent noncompliance is to keep the predictor and the outcomes temporally separate. I am studying labor market outcomes in adult life. By focusing on and measuring noncompliance in adolescence, I have identified the time order: adolescent noncompliance clearly precedes the outcomes of interest. There are a variety of terms one could use to denote the opposition to school or family authorities and rules, including resistance, nonconformity, rebellion, noncompliance, and others. Here and after, I have chosen to use adolescent noncompliance as it captures resistance to authorities’ rules. As such, I define noncompliance as disobedience to school rules or rules stated by authority figures such as parents, teachers, and school administrators. I conceptualize noncompliance through four dimensions: 1) context, 2) relationship between the actor and the
© Heili Pals September 2006
21
Chapter 2: Noncompliance and Life Consequences: A Literature Review institution upholding the norms, 3) intentionality, and 4) the extent of deviation from the rules. First, adolescents operate in contexts that often differ widely from each other. Both home and school are environments where adults are authorities, and have rules that differ considerably from the rules of youths’ peer groups. The roles of a student and a child in a family are heavily regulated by society and incorporate clear expectations for adolescent behavior. While conformity to peers is an interesting topic, it does not address the issue of obedience to the rules of authorities. Therefore, I identify noncompliance in relation to the norms accepted by the wider society: that is, noncompliance to the rules stated by school and family authorities and the noncompliance to the rules and expectations from above. Noncompliance to school or family rules may be manifested in a variety of ways. A person defying school rules, skipping classes, not doing homework, or confronting teachers and parents is not complying with school norms. At the same time, a student in school behaving as if he is a diplomat, wearing a business suit and carrying business cards can be also seen as deviating from widely accepted norms. Is there a difference in these types of deviance? Having conflicts with teachers and skipping school is a more confrontational type of noncompliance, as it is confrontational with regard to school authorities. This type of noncompliance is clearly more likely to be negatively sanctioned by school authorities. It is unknown whether different types of deviance have different effects on life outcomes, but confrontational noncompliance to the rules of authorities seems likely to have stronger
© Heili Pals September 2006
22
Chapter 2: Noncompliance and Life Consequences: A Literature Review and more lasting effects on adult outcomes. Therefore, I focus on confrontational types of noncompliance, such as arguing with teachers and skipping school. Noncompliance also depends on the larger context and on time. The same kind of activity may be seen as noncompliance in one setting and as compliance in another. For example, wearing make-up can be a direct defiance of rules in a school that has a rule against wearing make-up, whereas the same behavior is not viewed as noncompliance in other schools. Therefore, I need to take care to measure noncompliance in more universal terms and avoid types of noncompliance that can depend on local school norms. Rather, noncompliance should be identified through relatively universal rule-breaking (such as skipping school or having conflicts with teachers and parents). Finally, I need to consider the extent of deviation from the rules. Skipping school represents a relatively small deviation from school rules compared to hitting a teacher. Similarly, disobeying school authorities is not as great a deviation as deliberately causing someone bodily injury. The extent of deviation is often also measured by sanctions: whether or not the actions result in legal punishment. I focus on noncompliance to school authorities’ rules that differs from the usual understanding of deviance in its extent and exclusion of legal punishment.3 Therefore, I
3
Noncompliance to school authorities’ rules may bring sanctions from the legal
system if it is performed on a regular basis. A student who consistently skips school or has frequent conflicts with teachers may eventually be considered by the legal system as a candidate for a special boarding school.
© Heili Pals September 2006
23
Chapter 2: Noncompliance and Life Consequences: A Literature Review conceptualize adolescent noncompliance as opposition to rules of school and family authority measured in universal terms that does not involve criminal deviation from rules. NONCOMPLIANCE AND OUTCOMES The effects of adolescent noncompliance on adult outcomes have been analyzed in two strands of contemporary research. Psychologists consider such effects from the perspective of adolescent risk-taking, while sociological and youth development research concentrates on the life course effects of noncompliance. More often than not, however, noncompliance is conceptualized as deviance (and therefore as yielding negative sanctions and outcomes), and not as a sign of opportunism and innovativeness. Risk-taking as Opportunity. Psychologists have analyzed adolescent risktaking from two different theoretical perspectives: 1) “risk-taking as trouble,” believed to lead to dropping out of school, violence, and crime, and 2) “risk-taking as opportunity,” related to experimentation, autonomy, and identity development (Lightfoot 1997). The latter perspective recognizes that some level of risk-taking in adolescence can be advantageous for personal development. Constructive risk-taking is related to agentic qualities of adolescents whose risk-taking is both future- and goaloriented. Adolescent development researchers have also noted that some risk-taking activities (e.g., early drinking) can mark more rapid psycho-social development and can mean merely that a person is making an early transition to adult life (Jessor & Jessor 1975; Stacey & Davies 1972). Social psychologists have also connected
© Heili Pals September 2006
24
Chapter 2: Noncompliance and Life Consequences: A Literature Review nonconforming behavior to higher leadership ability and intellectual competence (Crutchfield 1955) and higher self-confidence in evaluative situations (Crowne & Liverant 1963). Negative Effects of Noncompliance in Youth Research. The sociological and youth development literatures have taken a uniformly negative view of the effects of rule-breaking in adolescence. It has been found that being socially nonconforming in adolescence is linked to dropping out of school and lower educational achievement (Alexander et al. 1997; Tanner et al. 1999), as well as criminal behavior (Sampson & Laub 1990), drug use (Hamil-Luker et al. 2004), and anti-social behavior (Gottfredson & Hirshi 1990) in adulthood. The sociological literature largely suggests that under the “normal” conditions of a stable society, conformists have higher educational and occupational attainment. The effect of adolescent nonconformity on educational attainment is direct, but its effect on occupational attainment is thought to be mainly indirect, operating through educational attainment. Alexander, Entwisle, and Horsey (1997) found that poor attitudes and undisciplined behavior among children in first grade have a strong impact on eventual school success, net of family background and measures of academic standing and ability. Tanner et al. (1999) argued that both common sense and previous research suggest that while skipping school ultimately hinders achievement in the labor market, it’s most evident direct negative effect is on educational attainment. They concluded that all measures of delinquency (including skipping school) have negative effects on educational attainment and occupational status. Skipping school also increased the likelihood of being unemployed as an adult. Chen and Kaplan (2003) connected deviance and early school failure to status
© Heili Pals September 2006
25
Chapter 2: Noncompliance and Life Consequences: A Literature Review attainment at midlife, through both a lower level of completed education and higher rates of deviant behaviors in early adulthood. Mechanisms. Several theories have tried to explain the negative effects of rebellious behavior in adolescence and the increased likelihood of antisocial and criminal behavior in adulthood for antisocial youth. In a general theory of crime, Gottfredson and Hirschi (1990) suggested that both rebellious behavior in adolescence and antisocial activities in adulthood are connected to a lack of self-control rather than the existence of societal control. Low self-control is supposed to explain an individual's propensity to commit crimes, just as high self-control explains an individual's likelihood of conforming to social norms and laws (Akers 1991). Cleckley (1976) and Wilson and Herrnstein (1985) supported the idea that some people are prone to criminal activity, which makes them more likely to commit crimes and engage in antisocial behavior. Simons et al. (1998, 2002) stressed the importance of weak social controls that allow adolescent antisocial behavior to lead to adult criminal activities. The importance of social ties has also been stressed (Laub et al. 1998; Laub & Sampson 1993; Sampson & Laub 1990). Laub, Nagin, and Sampson (1998) concluded that an early marriage resulting in cohesive bonds inhibits continuing criminal activities among juvenile offenders. Sampson et al. (1997) proposed a similar idea. They analyzed levels of crime at the neighborhood level and found that neighborhood social cohesion (measured by respondents’ evaluations of relationships, value sharing, and trust in the neighborhood) reduced levels of crime.
© Heili Pals September 2006
26
Chapter 2: Noncompliance and Life Consequences: A Literature Review Very few researchers have connected nonconforming behavior in adolescence with positive outcomes. In a discussion of student rebellion, Rafky (1979) recognized that disruptive acts may be intended to induce change in the school. Further, the desire to influence the school may not necessarily stem from the dissatisfaction that alienation implies, but rather from a positive desire to make school better or more responsive to student and societal concerns. Political efficacy refers to the general belief that one can affect the political process and other related institutions, such as schools. Students high in political efficacy may engage in rebellious or disruptive acts in order to change the schools. Rafky found that political efficacy is positively related to school rebellion and negatively associated with expressive alienation. That is, he argued that rebellion and expressive alienation are separate phenomena. Noncompliance to School Authorities’ Rules in Transitional Societies. The only research that has addressed some forms of noncompliance to the rules of authorities in transitional societies is Tuma and Titma (2003). They analyzed the effects of skipping school and other forms of noncompliance on educational transitions in regions of the Soviet Union. In accordance with the previous literature, they found that those who skipped school were less likely than those who did not skip school to take steps to get and complete higher education. This suggests that noncompliance to school authorities’ rules has expected negative effects on education similar to that in stable societies. The educational system is by its nature conformistfriendly. Tuma and Titma (2003) argued that teachers are more likely to reward ruleabiding behavior. Adolescent noncompliance is highly likely to be sanctioned by teachers, and these negative sanctions often yield lesser educational outcomes for
© Heili Pals September 2006
27
Chapter 2: Noncompliance and Life Consequences: A Literature Review those not complying with the rules of school authorities. However, Tuma and Titma (2003) did not study labor market outcomes. If labor market outcomes are more affected by the changes in society, the effects of noncompliance on labor market outcomes might differ from its effects on educational attainment. With few exceptions (Becker 1963; Phillips & Zuckerman 2001), the previous literature (whether it is sociological, social psychological, criminological, or adolescence literature) agrees on one point: in its broadest sense, noncompliance to authorities’ rules is negatively sanctioned and leads to negative outcomes. For example, a student not complying with the rules of school authorities is likely to be negatively sanctioned by teachers. Teachers’ sanctions often lead to lower grades and, therefore, to lower educational outcomes. Lower educational outcomes, in turn, predict lower labor market outcomes. EFFECTS OF NONCOMPLIANCE DEPEND ON THE SOCIAL CONTEXT The aim of this section is to discuss whether, and to what extent, the structural context influences the effects of noncompliance to authorities’ rules on life outcomes. The previous literature has not considered the social context in this sense. It has assumed that noncompliance to the rules of authorities has universal effects on life outcomes regardless of the social context in which someone operates. In contrast, I argue that the effects of noncompliance depend on the social context. My explanation comes from the action perspective: I describe how the level of predictability acts as a scope condition for the effects of noncompliance on life outcomes.
© Heili Pals September 2006
28
Chapter 2: Noncompliance and Life Consequences: A Literature Review Predictability and Changes in Society. Change in a society often brings about a period with a low level of predictability. The level of predictability depends on the clarity of the rules and norms of society. If the normative and legal structure is welldefined, then the level of predictability is high. Thus, in a society with a low level of predictability, both the normative and legal structures are not well-defined. Cohen (1966:19) described the level of predictability in terms of institutionalization. He admitted that in some situations there is no consensus about the rules of the system (low level of predictability). The level of institutionalization can be considered as a continuum ranging from the least institutionalized society to the most rigid society, in which everything is regulated. In a highly predictable society where norms and laws are well established, those who generally comply with the rules of authorities are in an advantageous position because they are accustomed to following rules, while those who do not comply with authorities’ rules are not. At the other end of the spectrum, in an unpredictable society with increased changes in norms and laws, the situation might be the opposite. A compliant youth who is accustomed to following authorities’ rules may lack the initiative to operate in an unregulated society. The unpredictable situation confuses the compliant youth and slows her down. A person with experience in not complying with the rules of authorities, however, is in an advantageous position because she is not expecting directions from regulations or authorities; therefore, she may find it easier to operate in an unpredictable society. Thus, in such an unpredictable society, it may be easier for a person with experience not complying with the rules of authorities to achieve higher labor market outcomes than a compliant
© Heili Pals September 2006
29
Chapter 2: Noncompliance and Life Consequences: A Literature Review youth. The disadvantages for compliant youths are probably even greater in the case of a rapid change from a predictable to a highly unpredictable society, because such a rapid change limits the time to adapt. Adaptation could work in two ways. First, those not complying with authorities’ rules are not bound by the rules of the old society and are therefore able to adapt to the rules of the new society more easily than compliant youths. Second, compliant youths may require more time to adapt to the rules of the new society. Therefore, the effects of noncompliance should change with the level of predictability and with the extent of changes in a society. Additionally, compliant youths are more likely to invest in old career paths; however, these old career paths are affected by the change in the society. On the other hand, noncompliant youths, who choose to look for new ways rather than investing in old career paths, get an early start with new types of jobs. Thus, under a high level of predictability, compliant youths should achieve higher labor market outcomes than those who do not comply with school authorities’ rules. However, in a society with a low level of predictability, the relationship between noncompliance and outcomes is likely to be the opposite: those who are compliant are likely to achieve lower outcomes than those noncompliant. This general idea can be applied to a variety of societies and life outcomes. To test the proposition that noncompliance has positive effects in societies with low level of predictability, in addition to conceptualizing noncompliance, I need to define: 1) Life course outcomes affected by noncompliance to authorities’ rules, and 2) A society that has rapidly changed toward less regulation.
© Heili Pals September 2006
30
Chapter 2: Noncompliance and Life Consequences: A Literature Review Outcomes Affected by Societal Changes. The life course outcomes suitable for this analysis are outcomes that: 1) have been recognized to be related to adolescent noncompliance, 2) clearly occur after adolescence (in order to establish the temporal order), and 3) are affected by the level of predictability in a society. Labor market outcomes provide a clear example matching all three criteria. Labor market outcomes occur chronologically after adolescence and have been found to be related to adolescent noncompliance to authorities’ rules (Tanner et al. 1999). Additionally, the level of predictability in society can be clearly related to labor market outcomes. The exact outcomes under study here are defined in the second half of Chapter 3. Excellent examples of societies that have rapidly changed toward lower levels of predictability are regions of the former Soviet Union after the collapse of the Soviet regime. The next chapter describes the changes in regions of the former Soviet Union, examines the characteristics that make regions of the former Soviet Union suitable for the analysis of noncompliance, and discusses how this change affected the relationship between noncompliance to school authorities’ rules and attainment in the adulthood.
© Heili Pals September 2006
31
CHAPTER 3: NONCOMPLIANCE
AND
LABOR MARKET OUTCOMES
IN
FORMER
SOVIET REGIONS SOCIETIES IN TRANSITION: THE FORMER SOVIET REGIONS Starting with the work of Elder (1974), the life course literature has paid considerable attention to societies in transition. Examples of such work include Elder’s (1974) analysis of life course changes during the Great Depression, Featherman and Sorensen’s (1983) analysis of changes in the school-to-work transition in Norway during its transformation from a rural to an industrialized urban society, and Titma and Tuma’s (2005) inquiry into changes in the life course during the transition in the former Soviet Union states. Transitional Societies as a Scope Condition. The life course literature has focused on societies in transition because the stages of a societal transition provide researchers with natural experiments that help them to assess the effects of structural changes on the life course. In a stable society, changes in societal structure occur too slowly to learn how changing structures affect the life course. Changing structures, however, create an opportunity to distinguish individual and structural explanations for social inequality (Schiff & Lewontin 1986). Therefore, changing structures help researchers to understand the intersection between political economy, stratification, labor market conditions, and social psychological factors (Wang et al. 1999). States that were formerly part of the Soviet Union experienced major and simultaneous changes in different domains during and after the collapse of the Soviet
© Heili Pals September 2006
32
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions Union. Extensive economic changes coincided with extensive changes in the legal system, attitudes, and culture. Former Soviet Regions. Several characteristics of this transformation make regions in the former Soviet Union suitable for studying the effects of adolescent noncompliance on labor market outcomes. These are: 1) fluctuations in the legal structure, 2) changes in socially accepted norms, and 3) changes in the opportunity structure. Significantly, these changes did not happen overnight. Rather, the transition involved a period of unsettled laws, norms, and opportunities that lasted several years. The societal changes in these three areas constitute the structural conditions necessary for the positive effects of adolescent noncompliance: a relatively unpredictable society that can be disadvantageous for compliant youths. Legal System. The fluctuations in the legal system that accompanied the transition from a command economy are widely documented (Clement & Murrell 2001; Gerber & Hout 1998; Murrell 2001, 1993; Sharlet 1993; Sterlin 1993; Tikhomirov 1996; Van Zon 1998). The changes affected various regulations covering economic freedoms such as property rights, tax laws, and contract laws. Modifications of the laws were frequent and not always straightforward. For example, some tax laws in Ukraine were even changed retrospectively (Van Zon 1998). New, often inconsistent, rulings from policy-makers occurred routinely (Sterlin 1993). Due to the extent and timing of these changes, the term “shock therapy” was used (Aslund 1992, 1993, 1994; Rostowski 1993). Under conditions of “shock therapy,” the command economy was dismantled before the institutions and legislation that could support a
© Heili Pals September 2006
33
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions new market economy were established (Gerber & Hout 1998; Murrell 1993; Clement & Murrell 2001). The legal confusion was especially pronounced due to the decentralization of legal reforms. Before the breakup of the Soviet Union, regions in the USSR were subject to a highly centralized system of government. Starting in 1991, individual republics began to create their own legislation, which often contradicted regulations at the federal level. The same happened at the level of krais and oblasts in Russia.4 For example, the Kurgan oblast unilaterally claimed the right to pre-empt federal legislation in its constitutional charter, whereas the Sverdlovsk oblast managed to get a seal of approval for this practice from the federal government through special provisions in bilateral treaties (Polishchuk 2001). The mushrooming of regional legislative acts made it difficult to monitor the process, let alone ensure the consistency and harmonization of regional legal regimes. Because of the lack of legislative acts, rushed and re-issued regulations, and differences at separate levels of the state, contradictions in state policies were common in all post-socialist countries. Such contradictions created a situation in which people had no idea what the new legislation would allow and forbid or when it would come
4
Krais and oblasts were administrative units in Russia with a federally appointed
governor and locally elected legislative body, two steps below the federal level (the higher step being the Soviet Republics). While administratively krais and oblasts were equal, only six large, thinly populated border regions were designated by the term krai.
© Heili Pals September 2006
34
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions into being. This created a society that was relatively unregulated legally, especially with regard to private property and entrepreneurial activities. Normative Change. In addition to changes in written laws, there was a transformation in the normative system of what was socially accepted. Unfortunately, most of the research on normative and value change during the transition lacks empirical evidence due to the absence of comparative data across time (Kyvelidis 2001). Based on the literature on normative change in Western industrialized societies, one would expect a slow change from materialistic to post-materialist values as GDP increased (Inglehart & Abramson 1994). Titma, Silver, and Anderson (1994) analyzed the work values of youth in Estonia in the 1980s and found that the majority of youth had internalized the values cultivated by the ideology of the socialist party-state, such as “jobs should be useful to society” and “work should be a major means of selfexpression.” In addition, material rewards from work were deemed much less important than in Western countries (Titma, Silver, & Anderson 1994: 15). Values became more individualistic with the passage of both historical time and time in the labor force. In contrast, Bowles and Weisskopf (1998) described socialist values as already post-materialist: collective rather than individual gains predominated. Because collective values were touted by political elites, they may have been expressed in opinion surveys (even anonymous ones) due to fear of punishment for expressing ‘reactionist’ views rather than because of true internalization of these values. A common view is that values have changed from collectivistic to more individualistic (Lauristin & Vihalemm 1994). Lauristin and Vihalemm (1997)
© Heili Pals September 2006
35
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions compared the extent of value change in Estonia and Sweden from 1991 to 1995 and found a considerably larger shift in values in rapidly changing Estonia than in relatively stable Sweden. However, the speed of change varied by subpopulations, such as generations (Lauristin & Vihalemm 1994) and minority and majority groups. Lauristin and Vihalemm (1997) found that Russians in Estonia had experienced a much sharper change in values than Estonians. They explained these differences by the disparate level of communist values between the nationality groups. That is, Russians had been more positive towards Soviet propaganda and therefore were less prepared for the new capitalist values. Attitudes about entrepreneurial activities seem to have become more favorable since the Soviet Union collapsed. Chernysh (1995) argued that with increasing economic hardship, a growing number of underprivileged youth became willing to resort to illegal forms of entrepreneurship. In Ukraine, however, the socio-cultural legitimacy of entrepreneurship has stayed at a fairly low level. Korzhov (1999) mentioned several reasons for this, including slow economic change and the spread of marginal entrepreneurship. Thus, deep normative changes do seem to have occurred in regions of the former Soviet Union. In general, norms and values have become more individualistic and more varied in these societies. Competition and economic hardship have created contradictory views and public conflicts about values. The lack of uniformity of norms across region, age, social group, and time made it difficult to operate in this environment of changing norms.
© Heili Pals September 2006
36
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions Opportunity Structure. Before the collapse of the Soviet Union, the labor market was shaped and managed institutionally in a completely different way than in a capitalist labor market. Institutional and legal arrangements, such as residence permits, tenure-related benefits, and social condemnation of active job-shopping, were designed to limit voluntary labor mobility (Gimpleson & Lippoldt 2001). Under the Soviet regime, tenure and loyalty were rewarded. With a transition from Soviet regime, the opportunity structure, defined by barriers of entry, vacancies, and job allocation, was poised to change. First, the state’s central allocation of jobs was discontinued toward the end of the 1980s. Until then, young adults who graduated from an institution of higher education were allocated to a job where they had to work for two to three years before they were allowed to initiate career moves. According to Clarke (1999), before 1992, almost two-thirds of those completing higher education and half of those completing technical education were assigned to their first jobs. This compulsory placement system had largely been dismantled by 1991 (Solnick 1993). New types of employment opportunities, including opportunities in the private sector and entrepreneurial activity, appeared, and some old ones (e.g., those connected to the Communist Party and to the Soviet regime) disappeared. Under socialism, state laws and
policies
forbade
people
from
engaging
in
entrepreneurial
activities.
Consequently, entrepreneurial activities occurred rarely and usually only on a small scale. In Russia, a major increase in the private sector started only after the adoption of the Basic Provision for the Privatization of State and Municipal Enterprises in the Russian Federation in 1992 (Chubais & Vishnevskaya 1993). However, by 1998,
© Heili Pals September 2006
37
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions privately owned and controlled firms accounted for between 30 and 40 percent of employment in Russia (Gimpleson & Lippoldt 2001). Some moves from the public to private sector were passive rather than active; that is, people moved from the public to private sector through the privatization of state-owned enterprises, often without changing their jobs. Additionally, the distribution of jobs in industries changed due to the weakening of state regulation. In Russia, the sharpest decline in industrial production occurred in 1992-1994, when output in most branches of industry fell by 45 percent (Clarke 1999). Subsequently, many industries were able to increase their outputs. Light industry, however, was almost extinguished in Russia by 1998 (Clarke 1999). Employment in the large and medium enterprises that were the main employers in the traditional Soviet economy fell sharply. The difference in employment was made up by the growth of small businesses. Labor turnover (i.e., net change in recorded employment at the enterprise level) was higher for blue-collar workers than for whitecollar occupations (Gimpleson & Lippoldt 2001). Gimpleson and Lippoldt (2001) argue that much of the turnover is linked to the reallocation of labor from large and medium enterprises to smaller firms. Youth and the Labor Market. How did these three conditions, fluctuations in legal acts, normative changes, and changes in the opportunity structure, affect young people entering the job market and their subsequent employment? As the entry to jobs changed from formal allocation to informal, young people had to use their own initiative to find jobs. A surge in start-ups to fill the gaps created by the decline in state companies meant more unpredictable labor conditions. Young, and therefore
© Heili Pals September 2006
38
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions probably more adaptable, people were over-represented in this unpredictable private sector (Gimpleson & Lippoldt 2001). I argue that all of the changes discussed above affected the labor market experience for youth during the transition period and resulted in unpredictability in labor market processes. There was no longer a widely shared sense within the social system of what was justly allowed and what was prohibited. Loyalty to one’s firm became less valued; behaviors that were formerly condemned became valued. Old secure jobs in the public sector were replaced with new, less secure opportunities in the private sector. Together these conditions changed youths’ experiences in the labor market during the transition and created suitable structural conditions in which noncompliance to authorities’ rules could have positive effects on labor market outcomes. Societal Changes and the Effects of Adolescent Noncompliance. I argue that if the structural conditions described above are not present (i.e., society is ruled by relatively clear legislation and normative boundaries), those who do not comply with the rules of school authorities fare worse in the labor market than those who do comply. I argue that under the structural conditions of the transitional former Soviet Union, which made labor market experiences far more unpredictable, the effects of adolescent noncompliance on labor market outcomes are positive. In addition, those who had the experience of not complying with authorities’ rules were often able to be first to react. They took advantage of new opportunities like private business before the new opportunities became institutionally structured. Being able to take on alternative opportunities operated as a buffer for those who are noncompliant individuals: they were better able than compliant youths to avoid or recover from the
© Heili Pals September 2006
39
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions collapse of the rigid state systems. Thus, the general argument that the effects of adolescent noncompliance depend on social context can be applied to these transitional societies changing from a command economy (see Figure 3.1).
18
20
22
24
26
28
30
Structural conditions facilitating a positive effect of noncompliance:
Noncompliance to school authorities’ rules
32
34
Age
Labor Market Outcomes: Second economy, Entrepreneurship, Economic sector, Occupational status
- Fluctuations in laws - Changes in norms - Changes in opportunities
Historical time 84
86
88
90
92
94
96
98
00
Collapse of the Soviet Union
Figure 3.1. The Effect of Adolescent Noncompliance on Labor Market Outcomes, As Dependent on Structural Conditions (Societal Changes)
GENERALIZABILITY If I find that the effects of adolescent noncompliance are positive in the unpredictable labor market conditions described above, the question arises whether the positive effects are particular to the transition from a command economy. Answering this question is beyond the scope of this study, but there is no reason to think that a similar mechanism would not work in other transitions with similar structural
© Heili Pals September 2006
40
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions conditions: confusion in the legal and normative systems, and change in the opportunity structure. There could be several reasons why noncompliance benefits people: 1) The main reason is the societal unpredictability: Those with the experience of not complying with the rules of authorities are better equipped to tackle the unpredictability of rapid social changes. 2) Another reason why adolescent noncompliance in the former Soviet Union is likely to contribute to positive labor market outcomes is because this noncompliance was against the rules of institutionalized authority and for individualism, even if inadvertently. This was the very value that was in many former Soviet regions eventually embraced by the new socio-economic system. This gave noncompliant adolescents an advantage: they were better suited to adapt to the new values. Furthermore, with the change from collectivist to individualist values, the value of noncompliance to institutionalized authorities’ rules changed from negative to positive. 3) Adolescent noncompliance to the rules of school authorities in the Soviet Union probably contributed to a greater sense of autonomy from authorities’ rules, and this sense of autonomy made it easier to operate during a period of societal fluctuations. Based on these mechanisms, which may operate simultaneously, the following conditions are needed for adolescent noncompliance to benefit a person: 1) a society must change and experience fluctuations in the normative and legal structures, and 2)
© Heili Pals September 2006
41
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions the noncompliance must be oriented against the values of the prior system and must be aligned with the values of the system replacing the old. NONCOMPLIANCE TO SCHOOL AUTHORITIES’ RULES AND NEW TYPES OF EMPLOYMENT I argue that the effects of adolescent noncompliance on labor market outcomes depend on the changes in the labor market. Here I discuss in more detail how the three structural conditions defining societal changes in the former Soviet regions influence the relationship between noncompliance to school authorities’ rules and adult labor market outcomes. I discuss the differences in societal changes over time and by the type of labor market outcome. In the next section, I take a more direct approach, comparing economic sectors and regions with different levels of societal changes. First, the absence or frequent changes of laws regulating the labor market make it unclear what is allowed and what is forbidden. Furthermore, it is difficult, if not impossible, for those operating in such an environment to predict the direction in which legislation will develop. The changing legislation makes it considerably more likely that those operating in such an environment will disobey laws at some point. Also, to operate successfully in this unpredictable environment, a person cannot rely solely on authorities, but must make independent decisions, even if these decisions contradict the wishes of authorities. Being accustomed to ignoring rules, those who did not comply with the rules of school authorities do not wait for guidelines from authorities. More than noncompliant adolescents, compliant youths expect to be directed by authorities and regulations, and thus may feel constrained if the regulating structure is fluctuating and
© Heili Pals September 2006
42
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions unclear. However, those with the experience of not complying with the rules of authorities are expected to be more independent from legislature, rules, and authorities, which makes it easier for them to operate in an unpredictable legal environment. In a clearly regulated, predictable job market, noncompliant youths do not have the advantage that they have in a poorly regulated, unpredictable market. The second important characteristic of societal changes is the change in social norms. Having the experience of going against commonly acknowledged norms, those who did not comply with school authorities’ rules find it easier to operate in an environment where norms are changing. In addition, being less restricted in general by the normative structure of society than compliant youths, those with the experience of not complying with the rules of authorities might find it easier to adapt to the values of the new society. Finally, changes in the opportunity structure are expected to alter the relationship between adolescent noncompliance and labor market outcomes. First, youths were accustomed to a central allocation of jobs that guaranteed them a position after graduation. The disappearance of central allocation placed the stress of finding a job on each individual. Entering the job market became a much more proactive and unregulated process. Those with the experience of not complying with the rules of authorities are likely to be more comfortable than compliant youths with such an unregulated process. Which labor market outcomes are affected by these structural conditions? The changes in regulations, norms, and opportunity structure are likely to be much greater in new types of employment opportunities. New employment opportunities are more
© Heili Pals September 2006
43
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions likely to have an underdeveloped legal framework; they are less governed by authority or tradition. The norms and values attached to these positions are still in formation. In terms of both process and requirements, entrance into new types of employment opportunities is less regulated than entrance into established employment opportunities. Therefore, in terms of all three dimensions, laws, norms, and the opportunity structure, those who did not comply with the rules of school authorities should be able to enter the less predictable environment of new employment opportunities more easily than those who did comply. Compliant youths, on the other hand, are more likely to invest in the old career paths. H1: In a rapidly changing society, adolescent noncompliance to school authorities’ rules increases the likelihood of working in new types of employment in adulthood. In the transitional regions of the former Soviet Union, several new employment
opportunities
emerged
that
fit
the
description
given
above.
Entrepreneurial activities and work in the private sector are examples of the new types of employment opportunities. Second Economy. The earliest entrepreneurial activity, participation in the second economy, emerged before the collapse of the Soviet Union. Participation in the second economy refers to participation in entrepreneurial activities before the collapse of the Soviet Union, when private business was illegal. All three structural conditions were present in the case of participation in the second economy. First, private business was forbidden. Even after the first laws allowing some types of private businesses
© Heili Pals September 2006
44
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions were introduced, rapid and frequent changes by the legislature made it difficult to obey the laws. Secondly, the overall attitude toward participation in the second economy was unpredictable and changing. On one hand, the popular attitude was negative because private business entailed illegal activity. On the other hand, participation in the second economy was often associated with high financial returns, which created more positive attitudes among many people. Thirdly, even before the collapse of the Soviet Union and the legalization of private businesses, there were numerous open niches for entrepreneurs. These included the transportation of goods from one republic to another, collecting scrap metal, and importing agricultural products to other republics. Those who obeyed the rules in school probably had a harder time operating in such an unpredictable environment and were therefore less likely to be engaged in the second economy. In addition, since compliant adolescents are accustomed to following rules, they are less independent in making their own decisions, which makes it harder for them to operate their own business in an environment with a weak or changing legal structure. H1a: In a rapidly changing society, noncompliance to school authorities’ rules increases the likelihood of participating in the second economy in adulthood. Titma and Tuma (2005) found that participation in the second economy before the collapse of the Soviet Union increased income and the likelihood for upward mobility during the transition from a command economy. They conceptualized participation in the second economy as a proxy for self-efficacy (Bandura 1992, 1995)
© Heili Pals September 2006
45
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions and argued that youth participating in the second economy were action-oriented and actively strove to improve their material situation. Furthermore, action-oriented youth searched out new opportunities and developed new careers instead of holding onto their current jobs while waiting for the economy to get better. Those who started entrepreneurial activities early in life developed business networks, were able to accumulate some capital, and learned the new capitalist ways of thinking about business, making decisions, and formulating strategies. Thus, entrepreneurial activities early in life tended to have positive effects on occupational attainment later in life. If hypothesis 1a is supported, it suggests that adolescent noncompliance has indirect effects on labor market outcomes later in life through participation in the second economy: the first mover effect. To establish whether adolescent noncompliance has direct effects on adult labor market outcomes, I examine the effects of noncompliance to school authorities’ rules on entrepreneurial activities and occupational status in adulthood. Entrepreneurial activities and high occupational status (focusing mainly on being a manager) have been used by Titma and Trapido (2002) to denote the “winners” of the post-Soviet transition in Estonia and Latvia. Entrepreneurship after the Collapse of the Soviet Union. After the collapse of the Soviet Union, private property was legalized and private business was allowed. This opened up entry to entrepreneurial activities for a wider segment of the population. Participation in the second economy during the Soviet regime was usually a secondary activity, done in addition to the main job, and by only a small percentage
© Heili Pals September 2006
46
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions of people. After the legalization of private business, it was possible for people to engage in entrepreneurial activities as their main job. The effects of adolescent noncompliance on participation in the second economy and on later entrepreneurial activities may differ due to different structural conditions. Soviet laws forbade participation in the second economy. Therefore, like noncompliance to school authorities’ rules, participation in the second economy entailed ignoring socially and formally accepted rules. However, it is also likely that adolescent noncompliance is advantageous as it indicates the long-term propensity to engage in entrepreneurial activities after private business is legalized. The structural conditions needed are also present after legalizing private business. Laws governing entrepreneurial activities changed frequently, the normative structure was unsettled, and the vacancies in the labor market were even greater due to the collapse of state industries. H1b: In a rapidly changing society, adolescent noncompliance to school authorities’ rules increases the likelihood of engaging in entrepreneurial activities in adulthood. Private Sector. Being one’s own employer, however, was not the only new type of employment. A whole new sector, the private sector, appeared in addition to the continuing (and shrinking) state sector. The emerging private sector included two different types of companies: start-ups (newly established companies) and firms that changed their ownership from state to private. The structural conditions needed for positive effects of adolescent noncompliance on adult outcomes (changes in the legal framework, attitudes, and the opportunity structure) were more prominent in both of
© Heili Pals September 2006
47
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions these newly-emerged private sector firms than in the labor market in general. Compared to the shrinking state sector, the private sector had an abundance of structural vacancies. In addition, compared to established state firms, start-ups had a less institutionalized organizational culture and more flexible rules. The change in ownership also created uncertainty about the future in privatized firms, although possibly to a lesser extent. In general, the legal framework covering the private sector was in constant development. Norms and values about the private sector were not yet stabilized. H1c: In a rapidly changing society, adolescent noncompliance to school authorities’ rules increases the likelihood of working either as an entrepreneur or as an employee in the private sector in adulthood. NONCOMPLIANCE TO SCHOOL AUTHORITIES’ RULES AND OCCUPATIONAL STATUS The question arises whether noncompliance to school authorities’ rules in transitional societies has positive effects only on new types of employment or on occupational status in general as well. The fluctuations in the legal and normative framework and the restructuring of the opportunity structure apply to the labor market as a whole in a transitional society, although possibly to a lesser extent than in new types of employment opportunities. The opportunity structure of the labor market changed dramatically after the collapse of the Soviet Union. In addition to the shift in occupational structure discussed earlier, changes in labor market entry also affected the opportunity structure. Those with experience of not complying with authorities’ rules are less likely to rely
© Heili Pals September 2006
48
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions on the old job placement system and more likely to look for innovative routes, especially personal ties, to gain a better position. In describing an unstructured labor market in Samara, Russia, Yakubovich (2001) noted that if a job is poorly paid and subject to high turnover, employers do not always advertise the position and simply wait for job candidates to show up “off-the-street.” This is a clear example of an unstructured and unpredictable labor market. Changes in legislation influenced employees by creating a more diverse pool of types of companies. Normatively, the evaluation of types of occupations changed. Some lost prestige and others gained. Therefore, all three structural conditions needed for positive effects of adolescent noncompliance on labor market outcomes were also present in the labor market more generally, albeit to a lesser extent than for the new types of employment. H1d: In a rapidly changing society, adolescent noncompliance to school authorities’ rules increases occupational status in adulthood. To summarize, I hypothesize that both before and after the collapse of the Soviet Union adolescent noncompliance has positive effects on engaging in new types of employment such as working in the private sector and entrepreneurial activities. In addition, I hypothesize that in transitional societies, adolescent noncompliance has positive long-term effects on occupational status in general. STATUS DIFFERENCES IN THE EFFECTS OF NONCOMPLIANCE One characteristic that researchers have considered when analyzing nonconformity and deviance is an actor’s social status. A series of early studies in
© Heili Pals September 2006
49
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions social psychology found a relationship between status and nonconformity to a group and argued that the occurrence of nonconformity is higher at either extreme of status (Blau 1960, 1963; Dittes & Kelley 1956; Harvey & Consalvi 1960; Homans 1961; Menzel 1960). Two different explanations have emerged. First, high-status actors are more likely to deviate from group norms because they feel confident about their position in the group (Hollander 1960). Low-status actors, on the other hand, may deviate from group norms because they are excluded regardless of their actions. Middle-status actors, however, are likely to suffer for nonconforming behaviors, and therefore are less likely to engage in nonconformists actions. In a test of Hollander’s theory (1960), Ridgeway (1981) found that nonconformity often increases the influence of an actor whose status in a group decision-making situation is similar to his or her partners, but can be a liability for low status group members. She used gender as a measure of status: a female confederate has lower status in an otherwise male group, and a male confederate has equal status in an otherwise male group. Differences in rates of nonconformity by the status of an actor are also considered in Merton’s (1968) strain theory describing anomie. Merton suggested that structural inequality makes some people more prone to deviance because some subgroups do not have good access to socially endorsed means of reaching the goals and values accepted by society. Thus, people who do not have (or perceive they do not have) the means to achieve socially accepted goals may still accept these goals as desirable. However, due to scarcity of legitimate means, they might innovate or design their own means rather than using socially acceptable means (innovative adaptation). The likelihood of innovative adaptation is largely defined by structural constraints that
© Heili Pals September 2006
50
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions prevent access to legitimate means. Based on this logic, low-status people may be more likely to use innovative adaptation because their access to legitimate goals is limited. Phillips and Zuckerman (2001) take the findings of social psychologists and apply them to a market environment. They discuss the interplay between status and nonconformity among entrepreneurs and find that middle-status actors are far more likely to be conformists than high- or low-status actors.
They conceptualize
nonconformity as a market analyst’s recommendation to sell (a rare occasion that is almost always followed by sanctions from firms) and status as someone’s ranking by colleagues. Krohn et al. (1980) analyze the relationship between social class and rates of deviance in a school setting. They did not find consistent support for the idea that working-class schools have higher rates of deviance. They concluded that social class is not an important predictor of adolescent delinquency. However, they conceptualized delinquency differently than I conceptualized noncompliance here. They focused on self-reported vandalism, theft, and the usage of social and hard drugs. Social status was measured by father’s occupation using the U.S. Census occupational scale (filling in mother’s occupation if the father did not live with the adolescent). The notion that people at either extreme of the status hierarchy are more likely to use noncompliance as a method than people of middle status might also apply to the relationship between noncompliance to school authorities’ rules and labor market outcomes. Thus, it is reasonable to expect interaction effects between status and adolescent noncompliance. Some researchers suggest that the interaction should be
© Heili Pals September 2006
51
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions linear; that is, the effects of noncompliance on labor market outcomes should be stronger for low-status people and weaker for high-status people. Other research suggests that the relationship is U-shaped: the effects of adolescent noncompliance on labor market outcomes are highest among low- and high-status people, but are weaker among middle-status people. H2: In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for high- and low-status actors than for middle-status actors. Status among Adolescents in the Former Soviet Regions. While studies have considered the interplay between status and nonconformity from a variety of perspectives, most of these studies have not considered adolescent noncompliance. Thus, these past conceptualizations of status might not be the best indicators of status among adolescents. Different studies of adolescents have used a variety of ways to conceptualize status in adolescence. Four of them, namely, gender, grades, type of secondary school, and socioeconomic status, seem to fit well with this study’s goal of analyzing the effects of adolescent noncompliance on labor market outcomes. Gender is still considered a status measure, especially when analyzing labor market success (McLeod & Owens 2004). Where noncompliance is measured in a school setting status could be measured by grades. In this case, high-status students are those who generally earn higher grades, while low-status students generally earn lower grades. Middle status, the status level emphasized by Phillips and Zuckerman (2001), would be measured by average grades. A particular measure of status among secondary school students in Soviet regions is the type of secondary school. Since
© Heili Pals September 2006
52
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions tracking was used heavily and the type of secondary school determined the next steps in the adolescent’s life, status might be measured by the type of school. Vocational schools represent a low-status school; those graduating from vocational school were expected to start working immediately, usually in a blue-collar job. In contrast, graduates of academic secondary schools were expected to continue to university and thus represent high status. Socioeconomic status provides another potential way of conceptualizing status in adolescence (Bradley & Corwyn 2002). For adolescents, socioeconomic status would be translated as parents’ socioeconomic status. Students from families with high socioeconomic status could afford to wear imported jeans, obtain Western music records, and go on vacation in more distant and nicer places. Thus, parents’ status translated to a higher status in a classroom setting. On the other hand, students from a low socioeconomic background did not have the means to obtain imported clothing or small odds and ends, and they vacationed in nearby summerhouses. Thus, when translating the first general hypothesis about the interaction between status and adolescent noncompliance, the following would be expected: H2a: In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for men than for women. H2b: In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for those with the lowest or highest grades than for those with average grades.
© Heili Pals September 2006
53
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions H2c: In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for those who graduated from either vocational or academic secondary schools. H2d: In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for those with parents of low or high socioeconomic status than for those with parents of average socioeconomic status. STRUCTURAL EXPLANATION: NONCOMPLIANCE TO SCHOOL AUTHORITIES’ RULES The previous section considered the effects of adolescent noncompliance from an individual perspective. However, I cannot ignore the structural mechanisms behind the positive effects of adolescent noncompliance on labor market outcomes in transitional societies. In particular, it is likely that the relationship between adolescent noncompliance to school authorities’ rules and adult outcomes is not only affected by the presence of the three structural conditions discussed here, but also by their strength. Extensive societal changes create higher level of unpredictability than relatively less extensive societal changes. Therefore, economic structures that change more extensively should create more suitable conditions for positive effects of adolescent noncompliance. H3: In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for adult labor market outcomes in economic structures with more extensive changes than in structures with less extensive changes.
© Heili Pals September 2006
54
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions At a more specific level, three separate tests of this hypothesis are possible: 1) a comparison of the effects of adolescent noncompliance on occupational status before and after the collapse of the Soviet Union, 2) a comparison of the effects of adolescent noncompliance in different geographic and political regions with varying extents of societal change, and 3) a comparison of the effects of adolescent noncompliance in economic sectors experiencing changes to different extents. Occupational Status before the Collapse of the Soviet Union. In contrast to early entrepreneurial activities (participation in the second economy), I do not expect adolescent noncompliance to have positive effects on the school-to-work transition in general if the move occurred before the collapse of the Soviet Union. The structural conditions needed (fluctuations in laws, norms, and the opportunity structure) were not present in the labor market overall. That is, the effects of adolescent noncompliance on the first job are likely to be negative or absent. Why? In general, the first job depends much more heavily on completed education than on anything else. In the old placement system, employers were looking at educational credentials rather than evaluating the person as a whole. Tuma and Titma (2003) found that those who skipped school in Soviet societies were less likely to take steps to get higher education and to complete it. It is likely that a similar dampening effect carried over to the first job since hiring was heavily dependent on educational credentials. If the transition from school to work occurred after the collapse of the Soviet Union, the same factors that facilitate the positive effects of adolescent noncompliance on labor market outcomes in an unstructured society may apply.
© Heili Pals September 2006
55
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions H3a: In a rapidly changing society, adolescent noncompliance to school authorities’ rules decreases the status of the first job if it is attained before the rapid social changes. Differences in the Transition in Regions of the Former Soviet Union. In the regions of the former Soviet Union, the main period of social, legal, and economic changes occurred over a period of approximately 10 years. These changes began before the collapse of the Soviet Union in 1989 and continued until the end of the millennium. Some regions of the former Soviet Union made almost a full transition from a command to a capitalist economy during this time; others changed but did not yet come close to the level of a capitalist society. For regions that made a full transition during this decade, the changes were relatively rapid, likely leading to confusion in the population about these changes. If the changes were relatively minor and slower, then people had time to adapt to them, which made it easier to operate in the changing society. Therefore, operating in a region with more extensive or more rapid changes should be more challenging for compliant youths, who are relying on guidance from authority, and less so for those who did not comply with the rules of school authorities and feel comfortable without the guidance from authority. H3b: In the former Soviet Union, adolescent noncompliance to school authorities’ rules is more advantageous in regions experiencing more extensive and more rapid changes as compared to regions with a less extensive societal transition. In order to test the above hypothesis, the extent of changes in laws, social norms, and opportunity structure across regions of the former Soviet Union under
© Heili Pals September 2006
56
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions study needs to be established. Titma and Murakas (2004) conducted a similar type of regional analysis by comparing income inequalities in the Baltic states and Russia, Ukraine, and Belarus. This comparison is justified by the differences in the speed and timing of the changes: Estonia and Latvia experienced shock therapy with rapid changes, while Russia, Ukraine, and Belarus introduced their market reforms considerably later. The Estonian transition from a command to a capitalist economy is also described by Titma and Silver (1996). Differences in the Extent of Changes by Sector. The extent of structural change differs not only by region, but also by economic sector. The clearest distinction in terms of the extent of change can be made by separating the state and private sectors. While the state sector was usually relatively stagnant, in terms of both modernization and labor mobility, the emerging private sector was in many ways opposed to the state sector. In general, the emerging private sector in the former Soviet areas had a higher rate of vacancies and a more fluctuating legal and normative framework than the state sector. The emerging private sector created new positions and had to manage in an unsettled legal environment because the laws and regulations covering private business did not yet exist or were in the process of creation. The private sector emerged through two different processes: privatization of state firms, and the creation of small businesses. While the latter definitely had unpredictable structures by virtue of newness, privatized firms experienced a great deal of change as well. The shift in ownership created a dimension of unpredictability for employees, who often did not know the plans of new owners and whether their position would be retained in the newly privatized firm. Therefore, the effects of
© Heili Pals September 2006
57
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions adolescent noncompliance on adult outcomes should differ between the private and state sectors during the transition from a command economy. H3c: In the former Soviet Union, adolescent noncompliance to school authorities’ rules tends to increase status more in the private sector more than in the state sector. There was also a certain level uncertainty and unpredictability in the state sector. With worsening economic conditions, even those in the state sector had to continue working while their salary was sometimes delayed weeks or months. However, the unpredictability in the private sector was even greater. Compared to the state sector there was an unknown future, unknown rules, and a higher potential for failure in the private sector. Table 3.1 summarizes the above hypotheses.
© Heili Pals September 2006
58
Chapter 3: Noncompliance and Labor Market Outcomes in Former Soviet Regions Table 3.1. Summary of Hypotheses. Ch 5: New types of employment In a rapidly changing society, adolescent noncompliance to school authorities’ rules increases the likelihood of … … participating in the second economy
H1a
… engaging in entrepreneurial activities
H1b
… working in the private sector
H1c
Ch 5: Occupational status In a rapidly changing society, adolescent noncompliance to school authorities’ rules… … increases the status of the job in adulthood
H1d
Ch 5: Status & adolescent noncompliance In a rapidly changing society, adolescent noncompliance to school authorities’ rules is more advantageous for … … men than for women
H2a
… those with the lowest or highest GPA
H2b
… those in vocational or academic secondary schools
H2c
… those with parents with the lowest or highest SES
H2d
Ch 6: Structural explanation In a rapidly changing society, adolescent noncompliance to school authorities’ rules … … decreases the status of the first job if it is attained before rapid social changes
H3a
… is more advantageous in rapidly changing regions
H3b
… is more advantageous in the private sector than in the state sector
H3c
© Heili Pals September 2006
59
CHAPTER 4: DATA AND MEASURES DATA To test hypotheses about the effects of adolescent noncompliance, I use data from the “Paths of a Generation” (PG) study, initiated and directed by Mikk Titma. PG is a longitudinal study of a sample of one generation of youth in selected regions of the Soviet Union who were about to graduate from secondary school in 1983-1984. The study was designed to follow the life course of this cohort (Titma & Koklyagina 1989). Three follow-up surveys were conducted: in 1988-1989, in 1992-1994, and in 1997-19995. At the time of the last survey, respondents were typically 32-33 years old. The PG data includes information on many aspects of the life course, including educational attainment, occupational history, family history, attitudes, perceptions, and values.6 For a more detailed description of the PG survey, see Titma and Tuma (1995) and Tuma et al. (1995). I use data from five regions: Belarus, Estonia, Latvia, the Kurgan region of Russia, and the Kharkiv region of Ukraine. In total, the five regions have eight and half thousand respondents, with regional sample sizes varying from 1,194 in Kurgan to 2,167 in Latvia.7 5
In most regions, the fourth wave interviews were conducted in 1997-1998; only a
small fraction of the sample was interviewed in 1999. 6
The summary of all variables used is listed in Appendix I.
7
A total of 22 regions are included in at least one wave of the study. However, some
regions were included for only one or two waves of data collection, making it
© Heili Pals September 2006
60
Chapter 4: Data and Measures Sample. The sample for the first wave of the study was designed to be representative of the population of 1983-1984 secondary school graduates. A clustered random design was used for data collection. The clustering variables were: 1) the type of secondary education institution: general, vocational, and specialized secondary schools, 2) the type of settlement and the territorial distribution of the population within the regions, and 3) the language of instruction in schools. Finally, one class was chosen from each selected school and everyone in this class was interviewed. The sample does not include a rather small fraction of Soviet youth who did not attend secondary school or who left before graduating (dropouts). By restricting study to adolescents who were still in school at the end of their secondary education, I concentrate on mainstream or “ordinary youth” and how their noncompliance to school authorities’ rules affected their attainment. Noncompliance to authorities’ rules may have different effects for adolescents who did not complete secondary education and had thus already rejected the institutionalized education system. It would have been helpful if the data had included people of variety of different ages. Observing different age groups going through the societal transformation would have allowed a better assessment of changes in the life course.
impossible to use those regions in this study. In addition, I have omitted regions with considerably different cultures and economies (Tajikistan, Kazakhstan). Mikk Titma has been the principle investigator on the "Paths of a Generation" project. Numerous individuals in the research sites played key roles in making the field work and interviewing a success.
© Heili Pals September 2006
61
Chapter 4: Data and Measures The age range in the data is about 10 years: respondents are 29-39 years old in 19971998. An additional analysis contrasting the effects of noncompliance by age groups within this range reveals no significant differences in the effects of noncompliance by age. Thus, instead of comparing across generations, I contrast economic sectors and regions with different levels of transition to estimate changes in the usual mechanisms of attainment. ADOLESCENT NONCOMPLIANCE TO SCHOOL AUTHORITIES’ RULES To measure noncompliance to school authorities’ rules, I use four variables collected in 1983-1984, when respondents were near the end of secondary school (see Table 4.1). Table 4.1 reports the frequencies of indicators of adolescent noncompliance. Approximately 30 percent of all youth reported only compliant activities. Another 34 percent had engaged in at least one noncompliance activity. Almost 25 percent had engaged in two, 10 percent in three, and 2 percent in all four activities. Of the four activities, the one most frequently reported was not fulfilling school rules; 50 percent of respondents admitted not following school rules. Next, 38 percent of the youth admitted skipping school at least sometimes. Conflicts were reported less often; twenty percent of the youth had conflicts with teachers or the school administration, but only 13 percent reported conflicts with their parents. Those engaged in any single noncompliance activity were more likely to have been involved in other forms of noncompliance (χ2 tests, p<.001). Table 4.2 shows the variation in the distribution on noncompliance in various regions.
© Heili Pals September 2006
62
Chapter 4: Data and Measures Table 4.1. Measures of Noncompliance to School and Family Authorities’ Rules. Valid Percent
Frequency
V49. Have you ever skipped school? No, never Rarely Yes, sometimes Yes, often
18.5 43.2 33.1 5.2
1,453 3,396 2,605 407
V242. I have had conflicts with teachers and the school administration No Yes
79.9 20.1
6,249 1,573
V56. I fulfill school rules and regulations without objection Agree strongly Agree Disagree Disagree Strongly
16.6 34.3 38.9 10.1
1,302 2,685 3,048 792
V306. I often had conflicts with my parents Definitely not Not really Yes, to some extent Yes, definitely
54.6 32.8 9.0 3.6
4,266 2,563 707 281
Number of nonconformist activities engaged 0 1 2 3 4
30.1 33.6 24.4 9.6 2.2
2,325 2,598 1,885 743 171
Source: Paths of a Generation, Wave 1. Regions: Belarus, Estonia, Latvia, Kurgan, Kharkiv. Valid N varies from 7,817 to 7,861.
© Heili Pals September 2006
63
Chapter 4: Data and Measures Table 4.2. Adolescent Noncompliance by Region (% difference from the overall percentage). Estonia
Latvia
Belarus
Kharkiv
Kurgan
Overall
V49. Have you ever skipped school? No, never Rarely Yes, sometimes Yes, often
-5.7 -2.2 7.0 0.9
-0.7 5.7 -4.2 -0.7
8.8 -2.7 -5.2 -0.9
1.6 -0.9 -0.2 -0.6
-0.4 -1.2 0.4 1.2
18.5 43.2 33.1 5.2
1.1 -1.1
2.0 -2.0
0.9 -0.9
-4.1 4.1
-1.3 1.3
79.9 20.1
-4.5 -4.1 5.9 2.9
1.3 4.1 -3.2 -2.1
3 1.5 -2.3 -2.1
1.5 -0.4 -0.7 -0.3
1.2 -0.4 -1.8 1.1
16.6 34.3 38.9 10.1
-1.3 1.7 0.0 -0.4
-5.2 4.6 1.0 -0.3
-0.3 0.4 0.4 -0.6
6.4 -7.1 -0.4 1.1
3.3 -2.5 -1.4 0.6
54.6 32.8 9.0 3.6
2,167
1,511
V242. I have had conflicts with teachers and the school administration No Yes V56. I fulfill school rules and regulations without objection Agree strongly Agree Disagree Disagree Strongly V306. I often had conflicts with my parents Definitely not Not really Yes, to some extent Yes, definitely Total N
2,141
1,443
1,194
Source: Paths of a Generation, Wave 1. Total valid N varies from 7,817 to 7,861.
© Heili Pals September 2006
64
Chapter 4: Data and Measures Adolescent Noncompliance in Relation to Other Variables. Previous literature has identified several patterns related to adolescent noncompliance. Nonconformity has been found to be related to gender (Rafky 1979), disengagement from school (Stinchcombe 1964), and autonomy (Lightfoot 1997). Similar relationships can be found in the PG data using the indicators of noncompliance listed in the previous section. This gives me confidence that the measures of adolescent noncompliance used here are capturing behaviors similar to those discussed in the previous literature. In particular, boys were more likely than girls to have engaged in all four measures of adolescent noncompliance (χ2 tests, p<.001). Compared to girls, twice as many boys skipped school often (7 percent compared to 3.5 percent). Eleven percent of boys and 24 percent of girls had never skipped school. About 30 percent of boys and only 14 percent of girls had had conflicts with teachers or school administration. Disliking school, trying to manage with minimal effort at school, and placing little importance on studies are indicators of school disengagement. Each of these indicators of school disengagement is related to at least some of the adolescent noncompliance measures. There was a clear negative relationship between liking school and skipping school (χ2 test, p<.001) and between liking school and following school rules (χ2 test, p<.001). Having conflicts with teachers or parents was not significantly connected to liking school, but the association is in the same direction. Not surprisingly, those who skipped school, did not fulfill school rules, or had conflicts with parents tended to report that they tried to manage with minimal effort at school. A similar pattern was less clear for those who had conflicts with teachers; this
© Heili Pals September 2006
65
Chapter 4: Data and Measures might be because students may need to feel integrated into school in order to have conflicts. Those engaged in adolescent noncompliance tended to put less importance on their studies and reported lower grade point averages than compliant youth (ANOVA, p<.001). Finally, noncompliant students consistently had a higher locus of control8 than compliant students (ANOVA, p<.001). The expected relationships with other variables and a high association between noncompliance measures (those engaged in one type of noncompliance were more likely to be engaged in other types of noncompliance) give confidence that these indicators do measure an underlying common trait. I use a variety of data reliability and data reduction techniques to analyze further whether the four indicators are measures of a single underlying adolescent noncompliance concept. Index Creation with Categorical Indicator Variables. Often ordinal or dichotomous variables are used in techniques designed for continuous variables under the assumption that the discrete variables represent an underlying continuous structure. However, there are several problems with this approach (Olsson 1979; Bollen & Barb 1981; Johnson & Creech 1983; Babakus et al. 1987; Dolan 1994; DiStefano 2002; Kolenikov & Angeles 2004).
8
Locus of control is measured by respondents’ perception of who or what influenced
them in making an important life decision: their choice of the type of secondary school to attend (see Pals and Tuma 2004 for a more detailed description of the measure).
© Heili Pals September 2006
66
Chapter 4: Data and Measures The most commonly used method for index creation is factor analysis, which assumes interval data. Kim and Mueller (1978: 74-75) note that ordinal variables may be used if the ordinal categories do not seriously distort the underlying metric scaling or if the underlying metric correlations between the dichotomous variables are moderate (.70) or lower. However, these criteria are nearly impossible to test empirically because the underlying metric scaling is unmeasurable. Similar problems arise with principal component analysis. Kolenikov and Angeles (2004) demonstrate that even when the underlying variables are perfectly correlated, their discretized manifestation may show much lower correlations unless the categorization thresholds match exactly. The use of dichotomous variables may lead to even more substantial distortions of the results because information about the natural ordering of categories is lost. To overcome the shortcomings of using traditional index creation methodologies assuming continuous indicator variables, I turn to methods for index creation using ordinal or dichotomous variables. I use Item Response Theory (IRT) to construct an index of adolescent noncompliance.9
9
I also explored other methods for creating the noncompliance index using categorical
variables. I created the index using principal component analysis based on polychoric correlations (Uebersax 2000; Kolenikov & Angeles 2004), categorical principal component analysis technique incorporated in SPSS (Meulman & Heiser 1999), and the Mokken scale procedure performed by a variety of programs (Stata, MSPWIN) (proposed by Hemker, Sijtsma, and Molenaar 1995). The resulting indices were very
© Heili Pals September 2006
67
Chapter 4: Data and Measures Item Response Theory and Graded Response Model. Item Response Theory (IRT) is commonly used in the creation of educational and psychological scales or tests. IRT assumes: 1) that the answer pattern for a set of items can be predicted by a latent trait, and 2) that a monotonically increasing item characteristic curve (ICC) describes the relationship between the respondent’s answers and the trait guiding them (Hambleton, Swaminathan, & Rogers 1991; Ostini & Nering 2006). The ICC indicates how the probability of a certain response to an item increases with corresponding increases in the underlying trait. As the trait of adolescent noncompliance increases, the likelihood of reporting regular engagement in individual noncompliance measures should increase as well. I use the graded response model proposed by Samejima (1969). The graded response model assumes, in addition to the usual assumptions of IRT, that the available categories of the indicator variables can be ordered. This model attempts to obtain more information from respondents’ responses than simply whether they give correct or incorrect answers. An advantage of IRT models is that they recognize that some variables in the scale can be “easy” indicators (weak noncompliance, for example), while others can be more “difficult” indicators (such as strong noncompliance). The graded response
similar, with correlations between indexes created by different methods ranging from .97 to .99. I chose the results from Item Response Theory because it is the most developed of all the alternatives and the one most widely used in different social science fields (see an example of usage in McFarland and Thomas 2006).
© Heili Pals September 2006
68
Chapter 4: Data and Measures model recognizes the different levels of difficulty for items, while component analysis and reliability analysis do not. Both component analysis and reliability analysis assume that items have the same relative frequency distribution (i.e., a respondent is equally likely to select them), while IRT models do not make this assumption. I use Multilog, a program created for estimating IRT models, to create a noncompliance scale using the graded response model. Multilog 7 was created by David Thissen, Wen-Hung Chen, and Darrell Bock (Du Toit 2003). Adolescent Noncompliance Index. The four available noncompliance indicators reflect two different spheres of adolescent life: three items deal with noncompliance to the rules of school authorities (a public entity) and one item deals with noncompliance to the rules of family authorities (private domain). The different index creation methods agree on one thing: noncompliance in the private domain (home) differs from noncompliance in the public domain (school). Therefore, the variable measuring conflict with parents does not have as high loadings in the overall noncompliance index as the school noncompliance variables do. The index reliability of the graded response model increases from .48 to .59 when the variable measuring conflict with parents is dropped. As all data reduction models agree on this issue, it is reasonable to restrict the noncompliance index to the school-related indicators. There could be several reasons why noncompliance within the family behaves differently than noncompliance to rules of school authorities. Most importantly, conflicts with parents might not always be because of the traits of an adolescent. Quarrelling parents can increase the number of conflicts that a child has with his or her parents, independent of the child’s own level of noncompliance. In addition, conflicts
© Heili Pals September 2006
69
Chapter 4: Data and Measures with parents might not always capture the disagreement with the rules of family. Furthermore, from the point of view of accepted norms, norms in schools are more consistent across schools than norms at home. Therefore, measuring noncompliance using only school-related noncompliance indicators should yield a better and more consistent measure of noncompliance to rules of authorities. The graded response model calculates item characteristic curves (ICC), the probability of a certain response across the adolescent noncompliance scale. Figure 4.1 shows the ICC for each of the three school-related noncompliance indicators. The item characteristic curves for adolescent noncompliance demonstrate that those with a very low level of noncompliance (noncompliance scale less than -2): 1) have a very high probability of never skipping school (probability of .9 or higher), 2) are likely to fulfill school rules without objections (probability of .85 and higher for selecting disagree or disagree strongly), and 3) have a minimal probability of having conflicts with teachers or the school administration. Moving to the right along the adolescent noncompliance scale increases the chances that someone has skipped school, does not fulfill school rules without objection, and has had conflicts with teachers or the school administration. Figure 4.2 shows the information magnitude of the adolescent noncompliance scale. For the most part, the scale is reliable at the level of r=.60. The scale is less reliable at a low level of noncompliance and is more precise at the middle and higher levels of noncompliance.
© Heili Pals September 2006
70
Chapter 4: Data and Measures
Skips school:
1
Never
Probability
.8 .6
Rarely
.4 Sometimes .2 Often
0
Fulfills school rules:
1
Yes!!
Probability
.8 .6
Yes
.4 No .2 No!!
0 1
Probability
.8
Conflicts
.6
with .4
teachers
.2 0 -3 -2.5 -2 -1.5 -1 -.5
0
.5
1 1.5 2 2.5 3
Adolescent Noncompliance
Figure 4.1. Item Characteristic Curves a) Skipping school; b) Fulfills school rules; c) Conflicts with teachers Source: Paths of a Generation, Wave 1
© Heili Pals September 2006
71
Chapter 4: Data and Measures
3
Total Information
2.5
2
1.5
1
.5
0 -3
-2.5
-2
-1.5
-1
-.5
0
.5
1
1.5
2
2.5
3
Adolescent Noncompliance Scale Figure 4.2. Total Information of the Adolescent Noncompliance Scale Source: Paths of a Generation, Wave 1
To make the scale for the adolescent noncompliance index more intuitive, I used a linear transformation to change the scale to vary between 0 and 1 (Figure 4.3). The mean of the new index is .40, and its standard deviation is .19. The adolescent noncompliance scale has a reasonable amount of variation, with more people concentrated at the lower end of the noncompliance scale. However, there are also people with a high level of noncompliance to school authorities’ rules.
© Heili Pals September 2006
72
Chapter 4: Data and Measures
20 18
Percent of Sample
16 14 12 10 8 6 4 2 0 0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Adolescent Noncompliance Scale Figure 4.3. Distribution of the Adolescent Noncompliance Scale Solid curve = normal density with the same mean and std deviation Source: Paths of a Generation, Wave 1
ADULT OUTCOME VARIABLES I evaluate the effects of adolescent noncompliance on several adult outcomes, all of which were measured at Wave 4 of the PG study, when the respondents were typically 32-33 years old. More than a decade of time had elapsed between the measurement of noncompliance to school authorities’ rules and the measurement of adult outcomes. The large time gap between the measurements makes it possible to eliminate the possibility of reverse causation.
© Heili Pals September 2006
73
Chapter 4: Data and Measures New Types of Employment. New types of employment are defined as entrepreneurial activities (hypotheses 1a and 1b) and as work in the private sector (hypothesis 1c). Entrepreneurial activities are measured in the following way: 1) participation in the second economy, 2) ownership of a firm, 3) ever being selfemployed, and 4) engagement in entrepreneurial activities in general.10 Participation in the second economy was measured retrospectively at Wave 4 as engagement in any of the following business activities before 1992: making or growing things to sell; trading; offering services to others for pay; and working in the respondent’s own business. About one-tenth of the sample reported engagement in the second economy before 1992. The most common entrepreneurial activity before 1992 is making or growing things to sell without a license (slightly more than 4 percent of the sample admitted doing this). Giving services without a license follows, with slightly less than 4 percent. Licensed entrepreneurial activity was very rare; only about half a percent reported making or growing things to sell or giving services with a license. Typically, collecting data about an illegal activity is biased due to underreporting. In this case, however, the measurement of participation in the second economy occurred in 1997-1998, long after the collapse of the old regime, when reporting participation in the second economy could not have put the respondent in danger. Furthermore, by that time, social stigma about the second economy had more or less disappeared; in fact, participation was often even viewed favorably. In addition, the fact that licensed entrepreneurial activities are reported less frequently further adds confidence that one
10
The measures for 2) to 4) were introduced in Pals and Tuma (2004).
© Heili Pals September 2006
74
Chapter 4: Data and Measures does not need to worry too much about underreporting when using the data on participation in the second economy. Ownership of a firm, a proxy for entrepreneurial activities after the collapse of the Soviet Union, is measured at the time of the Wave 4 interview. Nearly 10 percent of respondents reported owning a business or firm. I next measure whether respondents reported having been self-employed at any time from January 1992 to 1997-1998. This (0-1) indicator uses retrospective work histories collected in the Wave 4 interview. Approximately 6 percent of respondents reported having been self-employed at some point during the period from the collapse of the Soviet regime in 1992 to 1997-1998. The fourth indicator is based on respondents' reports of their main economic activity in 1997-1998. In particular, it measures whether respondents were employed, and whether they were engaged in entrepreneurial activities, such as making or growing things to sell, trading, offering services to others for pay, and working in their own or their family’s business. From responses to these questions, I create a polytomous variable with four categories: 1) 64 percent of respondents were employed and did not engage in any entrepreneurial activity (“employees”); 2) 15 percent were not employed for whatever reason (e.g., student, maternity leave, housewife, unemployed) and did not engage in any entrepreneurial activity (“economically inactive”); 3) 18 percent were “working entrepreneurs,” people whose main activity was either entrepreneurship or self-employment, or who were both employees and also engaged in some entrepreneurial activity; and 4) 4 percent were not employed but
© Heili Pals September 2006
75
Chapter 4: Data and Measures were engaged in some entrepreneurial activity as a secondary activity (“nonworking entrepreneur”) (defined in Pals and Tuma 2004). Job sector (hypothesis 1c) is measured in 1997-1998 at Wave 4. I differentiate between the state sector (state financed, state firm, or kolkhoz) and the private sector (private firm, private person, or stock company).11 The level of uncertainty is likely to be somewhat higher in start-up firms than in privatized companies. PG data, however, does not allow for such a distinction. Occupational Strata. Occupational strata of both the current and first full jobs were both measured in Wave 4 (hypotheses 1d and 1e). Occupation was measured using the 1988 ISCO (International Standard Code of Occupations) scale; I reduced it to seven occupational strata. About 23 percent of respondents did not have a job, 20 percent were workers, 5 percent worked in agriculture, and 11 percent worked in sales or service or were clerks in 1997-1998. In addition, about 13 percent of respondents were semiprofessionals, nearly 20 percent were professionals, and almost 9 percent were managers. The same coding was applied to the first full-time job that occurred before 1992 (the data come from the retrospective work histories collected at the Wave 4 interview). About 10 percent of respondents never had a job before 1992. For about 30 percent of respondents, their first job was as a skilled or unskilled worker, 8 percent worked in agriculture, 12 percent were in sales or service or worked as a clerk, 16 percent worked as semiprofessionals, and 20 percent worked as professionals. Fewer
11
Kolkhoz was a cooperative agricultural enterprise operated on state-owned land.
© Heili Pals September 2006
76
Chapter 4: Data and Measures people worked as managers than in 1997-1998. Only about 4 percent of respondents were managers in their first job before 1992. It might seem surprising that the number of people without a job is larger at the later measurement of occupational status. However, this should not be surprising, because occupational status at 1997-1998 is really a one-time measure, whereas occupational status of first job is recorded if a person had any full time job at any time before 1992. Most people in the PG sample did start working fulltime before the Soviet Union collapsed. CONTROL VARIABLES I employ several control variables commonly used in the attainment literature: gender, region, urban-rural location, parents’ and respondent’s years of education, and grade point average (GPA) in the last academic year of secondary school. I also control whether a parent had held a decision-making position in the Soviet era, which is a rough proxy for being in the Soviet nomenclature (Titma et al. 2003)12. In addition, I include a set of social psychological variables. Social psychological characteristics have increasingly been included in attainment research (Elder & Crosnoe 2002; Marshall 2000; Elder 1980; Clausen 1991a,b; Looker & Pineo 1983; Otto & Haller 1979). Social psychological variables are hypothesized to be 12
The results presented in the tables do not include the indicator of parents being in a
decision-making position because the measure was not asked in the Kurgan region. However, in all analyses, I have estimated an additional model including the other four regions and the measure of parents being in a decision-making position. The results regarding adolescent noncompliance are unchanged.
© Heili Pals September 2006
77
Chapter 4: Data and Measures especially important during the transition from a socialist system (Titma & Tuma 2005; Pals & Tuma 2004; Słomczyński & Mach 1996). I also include measures of self-esteem (Rosenberg 1965; Gecas 1982) and locus of control (Rotter 1966) at the end of secondary school. Both measurements are described in detail in Pals and Tuma 2004.13 ANALYSIS METHODS All of the analyses in this research are accompanied by full exploratory univariate and bivariate analyses, starting with the description of adolescent noncompliance already presented here. To assess the effects of adolescent noncompliance on adult outcomes (hypotheses 1a to 1e), different types of regression models are used depending on the outcome variable. I employ binary logistic regression models to examine participation in the second economy, firm ownership, and ever being self-employed. Multinomial logistic regression models are used to analyze working in different economic sectors, occupational strata, and participation in entrepreneurial activities as a main job. In addition, I use ordered logistic regression models to analyze the intensity of involvement in entrepreneurial activities. Independent variables are included in sets: 1) common control variables used for attainment models, 2) social psychological variables, 3) participation in the second 13
I also perform additional analyses controlling for whether or not a respondent
considered herself to be an activist in school and whether she thought she was popular at school (Wave 1). These results are not included in the tables presented here; however, the results regarding the effects of adolescent noncompliance are unchanged.
© Heili Pals September 2006
78
Chapter 4: Data and Measures economy (except when participation in the second economy is the outcome), and 4) adolescent noncompliance score14. To assess the structural explanation by the extent of changes in regions, I estimate separate regression models similar to those used for hypotheses 1a to 1e in region groups that experienced similar extent of changes (the Baltic states versus the Belarus, Kharkiv, and Kurgan regions). To estimate the structural effects in the state and private sectors, I modify the measure of occupational stratum to differentiate between professionals and managers in the private and state sectors.
14
Due to the large number of models estimated, not all nested models are presented in
the tables; however, in all cases, the increase in model fit due to the inclusion of adolescent noncompliance to the model is presented.
© Heili Pals September 2006
79
CHAPTER 5: EFFECTS OF NONCOMPLIANCE ON LABOR MARKET OUTCOMES This chapter is divided into six different sections. The first two focus on testing the hypotheses about the effects of adolescent noncompliance on new employment opportunities. These are followed by an analysis of the effects of noncompliance on occupational status. I also consider the linearity of the effects of adolescent noncompliance and explore the effects of noncompliance on the chances of not working. The chapter concludes with a discussion of how the effects of adolescent noncompliance depend on status as a student. ENTREPRENEURSHIP BEFORE AND AFTER THE COLLAPSE OF THE SOVIET UNION To test my hypotheses about the effects of adolescent noncompliance on new employment opportunities, I use a variety of outcome measures (hypotheses 1a, 1b, and 1c). Based on the argument that the effects of noncompliance depend on social context, adolescent noncompliance is expected to increase the chances of working in new types of employment opportunities during the transition from a command economy. Examples of such employment are entrepreneurship before and after the collapse of the Soviet regime, owning a firm, being self-employed, and working in the private sector. The results are presented in two parts. First, I discuss the results of the analysis in terms of model fit and the statistical significance of coefficients. Second, I calculate scenarios for a “typical” person which: 1) vary only in the level of adolescent noncompliance, and 2) hold all other independent variables constant. Predicted
© Heili Pals September 2006
80
Chapter 5: Effects of Noncompliance on Labor Market Outcomes probabilities calculated in this way help to illustrate the magnitude of the effects of adolescent noncompliance. Although bivariate analysis consistently finds that those engaged in entrepreneurial activities before or after the collapse of the Soviet Union had higher levels of adolescent noncompliance, multivariate analysis is needed in order to analyze the relationship in detail. First, I discuss the results of the analysis of the effects of adolescent noncompliance on engaging in entrepreneurial activities either before or after the collapse of the Soviet Union. The discussion of participation in the second economy (entrepreneurial activities before 1992) is followed by a discussion of various indicators of entrepreneurial activity after the collapse of the Soviet Union.
© Heili Pals September 2006
81
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.1. ML Estimates of Binary Logistic Regression Model of Participation in the Second Economy Before 1992. (Source: PG, Waves 1 and 4; N=6,305). Model I
II .74†††
Adolescent Noncompliance Control Variables Female
-.79†††
-.75†††
Education (years)
-.06††
-.06††
Grade point average
.03
.08
Parents’ education (years)
.01
.00
Rural
.21†
.24†
City
.14
.12
Estonia
.56†††
.55†††
Latvia
.40†††
.41†††
Belarus
-.38†††
-.37†††
Kurgan region of Russia
-.36††
-.37††
Self-esteem
.29
.35
Locus of control
.49
.42
-2.07
-2.09
Place of residence
Region
Intercept L.R. χ2 vs. null model (d.f.) L.R. χ2 vs. model with control variables (d.f.)
189.27††† (12)
199.50††† (13) 10.23††† (1)
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, selfesteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
82
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Second Economy. The binary logistic regression model of the likelihood of participating in the second economy controls for gender, region, locality, both respondent’s and parents’ educational level, GPA, self-esteem, and locus of control. Adolescent noncompliance increases the fit of the model (p<.01) and has a significant positive effect (p<.01) on participation in the second economy (see Table 5.1)15. Thus, adolescent noncompliance increases the likelihood of participating in the second economy before the collapse of the Soviet Union. Firm Ownership and Self-employment. Columns 1 and 2 of Table 5.2 present the results of binary logistic regression models of owning a firm in 1997-1998 and of ever being self-employed after 1992. Both firm ownership and self-employment represent new types of job opportunities, career paths that were poorly-regulated following the collapse of the Soviet Union. Therefore, as stated in hypothesis 1b, I expected adolescent noncompliance to have positive effects on the likelihood of being self-employed and owning a firm.
15
I estimated the same model including two other variables measured at the end of
secondary school: 1) how popular the student perceived herself to be in the class, and 2) whether the student considered herself an activist. I controlled for these variables to eliminate the possibility that adolescent noncompliance might measure popularity or activism in school. However, including these variables in the model did not change the effect of adolescent noncompliance, indicating that adolescent noncompliance is not a proxy for popularity. The analysis reported here does not include these two additional measures because they do not change the effect of adolescent noncompliance.
© Heili Pals September 2006
83
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.2. ML Estimates of Binary Logistic Regression Models of Owning a Firm in 19971998 and Ever Being Self-employed Since 1992. (Source: PG, Waves 1 and 4).
Adolescent Noncompliance
Owns a Firm
Ever Been SelfEmployed Since 1992
.81†††
.62†
-.33†††
-.69†††
Control Variables Female Education (years) Grade point average Parent's education (years)
.12††† -.03
.00 -.24†
.09†††
.02
-.46†††
-.17
Place of residence Rural City
.02
.28†
Region Estonia
.88†††
-.95†††
Latvia
.46†††
.51†††
Belarus
-.41†††
Kurgan region of Russia
-.10
-.03 .49†††
Self-esteem
.79††
Locus of control
.28
1.17††
Participation in second economy
.79†††
1.13†††
Intercept L.R. χ2 vs. null model (d.f.=14) L.R. χ2 vs. model with control variables (d.f.=1) Valid N
.33
-2.81
-2.88
348.46††† 10.43††† 6,247
231.98††† 4.06†† 6,219
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, selfesteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
84
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Indeed, including the adolescent noncompliance index increases the fit of the model significantly for both self-employment (p<.10) and firm ownership (p<.01). Furthermore, the coefficient for adolescent noncompliance is significant and positive for both outcomes. However, the effect is greater for firm ownership than for selfemployment. One standard deviation increase in adolescent noncompliance increases the odds of owning a firm in 1998 by 17 percent and the odds of ever being selfemployed between 1992 and 1998 by 13 percent. To get a better sense of the magnitude of the effects of adolescent noncompliance, I calculated scenarios for a typical person. Figure 5.1 shows how participation in the second economy and the likelihood of owning a firm is affected by varying adolescent noncompliance for a male from a city in the Kharkiv region with 12 years of education, an average GPA, average self-esteem, average locus of control, and parents who have 10 years of education. Figure 5.2 shows a similar prediction for self-employment after 1992. The slopes of the curves reveal the magnitude of the effects of adolescent noncompliance; the shaded area denotes the 95 percent confidence intervals. The steepest slope is for participation in the second economy; a Kharkiv male who was extremely compliant in adolescence has a .08 probability of participating in the second economy before 1992, while a similar male who was extremely noncompliant in adolescence has a probability of .15. A Kharkiv male with the characteristics enumerated above who was extremely compliant in adolescence is predicted to have a probability of .03 of owning a firm in 1998, whereas a similar male who was extremely noncompliant in adolescence is predicted to have a .06 probability of owning a firm. Similarly, as adolescent noncompliance changes from extremely low
© Heili Pals September 2006
85
Chapter 5: Effects of Noncompliance on Labor Market Outcomes to extremely high the probability of ever being self-employed increases from .06 to almost .11. Of course, one must be careful when interpreting predicted probabilities using the extremes of the independent variable. From Figure 4.3, we know that the 10th percentile of the adolescent noncompliance scale is .17. Thus, the most compliant 10 percent of the sample have, on average, a probability of .08 or lower of participating in the second economy. The people who are among the top 10 percent in terms of adolescent noncompliance have a probability of .12 or higher of participating in the second economy as adults.
© Heili Pals September 2006
86
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
Second Economy Owns a firm in 1997-1999
.16 Predicted Probability
.14 .12 .1 .08 .06 .04 .02 0 0
.2
.4
.6
.8
1
Adolescent Noncompliance Figure 5.1. Predicted Probability of Participation in the Second Economy and Owning a Firm in 1997-1998, by Adolescent Noncompliance Source: Paths of a Generation, Waves 1 & 4. Predicted probabilities from the binary regression models presented in Tables 5.1 and 5.2. Note: Shaded area marks 95% confidence intervals. Predicted probabilities for a hypothetical male born in a big city in the Kharkiv region of Ukraine with 12 years of education who did not participate in second economy and whose most-educated parent has 10 years of education. This hypothetical person has an average GPA, average selfesteem, and an average locus of control.
© Heili Pals September 2006
87
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
.16
Predicted Probability
.14 .12 .1 .08 .06 .04 .02 0 0
.2
.4
.6
.8
1
Adolescent Noncompliance
Figure 5.2. Predicted Probability of Self-Employment after 1992, by Adolescent Noncompliance Source: Paths of a Generation, Waves 1 & 4. Predicted probabilities from the binary regression models presented in Table 5.2. Note: Shaded area marks 95% confidence intervals. Predicted probabilities for a hypothetical male born in a big city in the Kharkiv region of Ukraine with 12 years of education who did not participate in second economy and whose most-educated parent has 10 years of education. This hypothetical person has an average GPA, average selfesteem, and an average locus of control.
© Heili Pals September 2006
88
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Economic Activity in 1997-1998. Next, I use multinomial logistic regression models to estimate the effects of adolescent noncompliance on economic activity in 1997-1998. (See Appendix II about the assumption of the independence of irrelevant alternatives [IIA] for the multinomial logistic regression model.) The dependent variable has 4 categories: 1) being economically inactive, 2) working and being an entrepreneur, 3) not working, but being engaged in entrepreneurial activity, and the reference category, 4) being an employee (see Table 5.3). Again, including adolescent noncompliance in the model increases model fit significantly (p<.05). Compared to being an employee, adolescent noncompliance has a significant positive effect on working and being an entrepreneur. Thus, those who were noncompliant as adolescents are predicted to be more likely to work and be entrepreneurs than to work just as employees. The level of adolescent noncompliance of economically inactive people and nonworking entrepreneurs is similar to that of employees. A one standard deviation increase in adolescent noncompliance increases the odds of being a working entrepreneur rather than an employee by 11 percent. The predicted probabilities of being economically inactive or being engaged in entrepreneurial activities without a main job are almost constant across the range of adolescent noncompliance (see Figure 5.3). The predicted probability of being an employee decreases from .69 to .60 as the level of adolescent noncompliance for a typical male increases from the lowest to the highest value. However, the predicted probability for being a working entrepreneur increases from .14 to .22 for the typical male as adolescent noncompliance increases from its lowest to highest value.
© Heili Pals September 2006
89
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.3. ML Estimates of Multinomial Logistic Regression Model of Economic Activity in 1997-1998, Effects Relative to “Employees” (Source: PG, Waves 1 and 4; N=6,305). “Economically “Working “Nonworking inactive” entrepreneur” entrepreneur” Adolescent Noncompliance Control Variables Female
.18
.58††
1.06†††
-.74†††
Education (years)
-.15†††
Grade point average
-.12
.42†† -.14†††
.02 -.07
-.26†
.04††
.02
.09
.14
.16
.07
.08
Estonia
-.18††
-.17††
Latvia
-.04
Parent's education (years)
.03††
.53
Place of residence Rural City
-.05 †
Region
Belarus Kurgan region of Russia
-.38
†††
.24††
.17†† -.26
††
.27††
-.07 -.29† .20
Self-esteem
-.10
.43
-.79†
Locus of control
-.09
.52
.33
Participation in second economy Intercept L.R. χ2 vs. null model (d.f.=42) L.R. χ2 vs. model with control variables (d.f.=3)
†
-.77†††
†††
-.41
3.17
-2.14
-1.43
3.31††† -3.64
1771.11††† 8.45††
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, selfesteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
90
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
Predicted Probability
.5
.4
.3
.2
.1
0 0
.2
.4
.6
.8
1
Adolescent Noncompliance Not an Employee
Economically inactive
Working entrepreneur
Nonworking entrepreneur
Figure 5.3. Predicted Probability of Economic Activity, by Adolescent Noncompliance Source: Paths of a Generation, Waves 1 & 4. Predicted probabilities from the binary regression models presented in Table 5.3. Note: Shaded area marks 95% confidence intervals. Predicted probabilities for a hypothetical male born in a big city in the Kharkiv region of Ukraine with 12 years of education who did not participate in second economy and whose most-educated parent has 10 years of education. This hypothetical person has an average GPA, average self-esteem, and an average locus of control. The slope for “working entrepreneurs” is significantly different (p < .01) from the slope for employees.
© Heili Pals September 2006
91
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Thus, adolescent noncompliance before the collapse of the Soviet Union does indeed increase the chances of being engaged in entrepreneurial activities, both immediately before and after the collapse of the Soviet Union. Engaging in entrepreneurship, a new career opportunity that was very poorly regulated by legislation and norms before and after the collapse of the Soviet Union, meant operating in a very unpredictable atmosphere. The positive effect of adolescent noncompliance on participation in the second economy seems to be somewhat greater than its effects on the various measures of entrepreneurship after the collapse of the Soviet regime. This is consistent with the argument that the effect of adolescent noncompliance depends on the social context because the model predicts stronger positive effects for adolescent noncompliance on outcomes that are least covered by established rules and regulations. Regulation of participation in the second economy was very unclear and inconsistent. What was allowed and what was prohibited by legislation was even less clear immediately before the collapse of the Soviet Union than after the collapse of the Soviet regime, when the first regulations covering private business were introduced. Similarly, norms about the second economy were not clear. The overall attitude towards small business activities was contradictory, with some people opposing them but many others supporting or being ignorant about such activities. The distinction between legal and illegal activity was further complicated by the uncertain pay. The pay was frequently not monetary; rather services or products were often exchanged for other services, homemade or homegrown products, or alcohol.
© Heili Pals September 2006
92
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Of the various entrepreneurial activities after the collapse of the Soviet regime, adolescent noncompliance has the weakest effect on ever being self-employed since 1992. This may be because self-employment involves a rather small-scale entrepreneurial activity that does not involve as much unpredictability as larger-scale entrepreneurial activities often do. Variety of Entrepreneurial Activities. I have used three different indicators of entrepreneurial activities after the collapse of the Soviet Union: self-employment since 1992, owning a firm, and claiming to be engaged in entrepreneurial activities in 19971998 (the same outcomes were used in Pals and Tuma 2004). These indicators reflect slightly different types of entrepreneurships. To better understand the effects of adolescent noncompliance on entrepreneurial activities after the collapse of the Soviet Union, I created a composite index of the number of different types of entrepreneurial activities someone has participated in. About 1.5 percent of the sample participated in all three entrepreneurial activities. More than 7 percent participated in two, 14 percent participated in one, and 77 percent were not engaged in any type of entrepreneurial activity.
© Heili Pals September 2006
93
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.4. ML Estimates of Ordered Logistic Regression Model of the Number of Types of Entrepreneurial Activities Participated in (Source: PG, Waves 1 and 4). Number of Types of Entrepreneurial Activities Adolescent Noncompliance
.55†††
Control Variables Female Education (years) Grade point average Parent's education (years)
-.62††† .06††† -.05 .05†††
Place of residence Rural
-.09
City
.08
Region Estonia
.19†††
Latvia
.29†††
Belarus
-.29†††
Kurgan region of Russia
.13
Self-esteem
.46†
Locus of control
.65†
Participation in second economy
2.04†††
Cut-point 1
1.29
Cut-point 2
2.60
Cut-point 3
5.49
L.R. χ2 vs. null model (d.f.=14) L.R. χ2 vs. model with control variables (d.f.=1) Valid N
943.68 ††† 9.75 ††† 6,163
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the mosteducated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
94
Chapter 5: Effects of Noncompliance on Labor Market Outcomes An ordinal logistic regression model of the number of different types of entrepreneurial activities in which a person has participated indicates, as expected, that the greater the level of adolescent noncompliance, the greater the likelihood of participating in a higher variety of entrepreneurial activities (see Table 5.4). Inclusion of the noncompliance index significantly improves the model fit (p<.01). Adolescent noncompliance increases engagement in a greater variety of entrepreneurial activities. Predicted probabilities calculated from the ordered logistic model illustrate the effect of adolescent noncompliance on the number of types of entrepreneurial activities (see Figure 5.4). A change from an extreme adolescent compliance to an extreme adolescent noncompliance increases the probability of engaging in any entrepreneurial activities from .18 to .26. A similar shift in adolescent noncompliance increases the probability of engaging in one entrepreneurial activity from .12 to .18. Thus, adolescent noncompliance increases the likelihood of participating in entrepreneurial activities in general. As compared to those who were compliant as students, those who were noncompliant tend to participate in more types of entrepreneurial activities and are more involved in private business.
© Heili Pals September 2006
95
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
Predicted Probability
.3 .25 .2 .15 .1 .05 0 0
.2
.4
.6
.8
1
Adolescent Noncompliance Entrepreneurial Activities:
One to three
One
Two
Three
Figure 5.4. Predicted Probability of the Number of Entrepreneurial Activities, by Adolescent Noncompliance Source: Paths of a Generation, Waves 1 & 4. Predicted probabilities from the binary regression models presented in Table 5.4. Note: Shaded area marks 95% confidence intervals. Predicted probabilities for a hypothetical male born in a big city in the Kharkiv region of Ukraine with 12 years of education who did not participate in second economy and whose most-educated parent has 10 years of education. This hypothetical person has an average GPA, average self-esteem, and an average locus of control.
© Heili Pals September 2006
96
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
PRIVATE VS. STATE SECTOR Like entrepreneurship, working in the private sector meant taking on a new career opportunity in the former Soviet Union. It differs from being an entrepreneur in the sense that one can be either an employee or employer in the private sector, whereas entrepreneurship means that someone is an employer of other people or his or her own employer, but not an employee. How do being an employee, being an entrepreneur, and being self-employed differ from one another? Even if the level of unpredictability in society is high, employees have lower levels of responsibility than entrepreneurs. Operating in a situation where the future is hard to predict is not as problematic for employees because the responsibility for actions lies more on the shoulders of the employer. Entrepreneurs, on the other hand, are directly accountable for their businesses. Working in the private sector includes both employee and employer positions. By definition, entrepreneurs work in the private sector and not in the state sector. Earlier results already established that adolescent noncompliance has a positive effect on entrepreneurship in the former Soviet regions. Therefore, I separate people working in the private sector into two groups: 1) entrepreneurs (and therefore employers), and 2) non-entrepreneurs (therefore people likely to be employees). Based on the results discussed above, I expect that adolescent noncompliance has a positive effect on the likelihood of working in the private sector as an entrepreneur as compared to the state sector. The separation of entrepreneurs and non-entrepreneurs in the private sector allows one to estimate whether the positive effect of adolescent noncompliance is also
© Heili Pals September 2006
97
Chapter 5: Effects of Noncompliance on Labor Market Outcomes present for non-entrepreneurs in the private sector. The dependent variable has three other categories: state sector, stock companies, and no job. The state sector includes state-financed and state-owned companies and kolkhozes. The results for the multinomial logistic regression model show that adolescent noncompliance has positive effects on being both an entrepreneur and a nonentrepreneur in the private sector (Table 5.5). Thus, those who were noncompliant as adolescents are more likely to work in the private sector rather than in the state sector, even if they work as employees. This model also reveals a different effect of adolescent noncompliance. Namely, those who were noncompliant as adolescents are also more likely not to have a job in 1998 as compared to working in the state sector. That is, adolescent noncompliance seems to have both hindering and fostering effects on labor market outcomes: high adolescent noncompliance increases the chances of both working in the private sector and not working at all. These diverse effects of adolescent noncompliance are explored in more detail later in this chapter.
© Heili Pals September 2006
98
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.5. ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, Effects Relative to the State Sector (Source: PG, Waves 1 and 4; N=6,256).
Adolescent Noncompliance
No job
Stock company
.47††
.20
Control Variables Female Education (years) Grade point average Parent's education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control In second economy Intercept
-.88
Private Sector Employee Entrepreneur .82†††
1.15†††
.38 ††† -.19††† -.20†† .04††
-.54††† -.13††† -.04 .02
-.35††† -.13††† -.00 .03††
-1.02††† -.07†† -.07 .05††
-.19†† .24††
-.20†† .33†††
-.28†† .56†††
-.18 .48†††
.17†† .10 -.36††† -.12 -.09 -.29 .40††
1.52††† .26†† -.96††† -.12 -.06 -.94†† .22†
-.10 1.19††† -.67††† -.32†† .37 .38 -.87†††
-.31†† .82††† -.44††† .31†† .42 -.07 1.51†††
L.R. χ2 vs. null model (d.f.=56) L.R. χ2 vs. model with only control variables (d.f.=4)
-1.04
-1.58
-1.62
1781.47††† 25.00†††
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the mosteducated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
99
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Next, I compare work in the private sector with work in other sectors. Based on the multinomial regression results presented in Table 5.5, I calculate coefficients for all possible comparisons for being an entrepreneur and an employee in the private sector. These two positions are compared to the odds of not working, working in a stock company, working in the state sector, and working in the private sector (see the first column of Table 5.6).
Table 5.6. Estimated Effect of Adolescent Noncompliance on the Odds of Working in the Private Sector (Based on the Multinomial Logistic Regression Analysis of Economic Sector in 1997-1998 in Table 5.5. Source: PG, Waves 1 and 4; N=6,235).
Estimated Effect of Adolescent Noncompliance
Percent Change in Odds for 1 Std Dev Increase in Adolescent Noncompliance
An employee in the private sector vs. … … No job … Stock company … An entrepreneur in the private sector … State sector
.35 .62†† -.33 .82†††
7.1 12.8 -6.2 17.4
An entrepreneur in the private sector vs. … … No job … Stock company … An employee in the private sector … State sector
.68†† .95†† .33 1.15†††
14.2 20.4 6.6 25.2
Odds comparing
†
p<.05,
††
p<.01,
†††
p<.001 (1-tailed test)
Note: Controlled for gender, type of secondary school, GPA, education, parents’ education, place of residence, region, adolescent self-esteem and locus of control, and participation in the second economy.
© Heili Pals September 2006
100
Chapter 5: Effects of Noncompliance on Labor Market Outcomes The results are fairly similar for both entrepreneurs and employees in the private sector. Adolescent noncompliance increases the likelihood of working either as an employee or as an entrepreneur in the private sector compared to working in a stock company or working in the state sector. Adolescent noncompliance also has a significant positive effect on working in the private sector as an entrepreneur as compared to not working at all. Adolescent noncompliance seems to have the largest positive impact on the likelihood of working in the private sector as an entrepreneur compared to working in the state sector. The second column of Table 5.6 demonstrates that a one standard deviation change in adolescent noncompliance increases the odds of being an entrepreneur in the private sector compared to working in the state sector by 25 percent. The same change in adolescent noncompliance increases the odds of working in the private sector as an employee compared to working in the state sector by 18 percent. OCCUPATIONAL STATUS The results presented in the previous two sections clearly show that adolescent noncompliance is positively related to engaging in entrepreneurial activities and working in the private sector after the collapse of the Soviet Union. However, adolescent noncompliance does not increase the likelihood of working in the state sector. I hypothesized, however, that due to the high frequency of changes in society, there should be positive effects for adolescent noncompliance at a more general level rather than just for work in new types of employment. I argue that those who were noncompliant as adolescents managed better in the job market than those who were
© Heili Pals September 2006
101
Chapter 5: Effects of Noncompliance on Labor Market Outcomes compliant as adolescents, and therefore in general tended to achieve higher occupational attainment. To test this hypothesis, I use occupational status in 1997-1998 as a dependent variable (see Table 5.7). The results for various occupational statuses are compared to workers. For example, all coefficients for female are positive and significant, meaning that females are more likely than men to be in any category other than worker. The effects of the classical attainment variables have the expected patterns. The chances of being in an occupational category other than worker rise as educational level increases. A higher grade point average increases the chance of being a professional or manager rather than a worker. There is also the expected effect for parental educational level: people with better-educated parents have a greater likelihood of being in higher status occupations, such as professional or managerial positions. Increases in parental education also increase the chances of not working; some of those people might have returned to school to gain higher education for themselves.
© Heili Pals September 2006
102
Table 5.7. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998, Effects Relative to Workers (Source: PG, Waves 1 & 4, Valid N=6,321). No Job Agricultural Clerk, Sales, SemiProfessional Manager Worker Service professional Adolescent Noncompliance .37 -.23 .43 -.13 .29 .58† Control Variables Female Education (years) Grade point average Parents’ education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control Participation in second economy Intercept L.R. χ2 vs. null model (d.f.=84) L.R. χ2 vs. model with control variables (d.f.=6) †
1.76††† .22††† -.05 .06†††
.43†† .08† -.02 -.04
3.04††† .14††† .14 .03
1.66††† .38††† -.02 .01
-.01 .10
.94††† -.58††
-.06 .18
.03 -.07
-.01 -.02 -.29††† -.02 .28 -.06 .30†
.34††† .63††† -.91††† .08 -.15 .68 .62††
.17 .25† -.48††† -.01 .38 .23 .23
-.49
-2.02
-2.24
.17 .06 .00 .03 .76†† .94† .23 -.94
1.73††† .88††† .28†† .09†††
.28† .65††† .33†† .08†††
.07 .04
.03 .12
.00 .23† -.18 .15 .55 -.72 .02
.52††† .35††† -.54††† .00 1.11†† -.43 .34†
-1.26
-1.07
†††
3958.68 10.40
103
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Furthermore, the effect of adolescent noncompliance approaches statistical significance for the “no job” category in 1997-1998 (p=.052, one-tailed). Thus, adolescent noncompliance may increase the likelihood of being without a job. In addition, adolescent noncompliance increases the chances of becoming a manager as compared to becoming a worker (p<.05). Thus, there does seem to be a positive effect of adolescent noncompliance on occupational status in 1997-1998. Adolescent noncompliance does not increase the chances of becoming a professional (a heavily education-driven outcome). While becoming a manager is also education-driven (both years of education and grade point average have a positive effect on the chance of being a manager), adolescent noncompliance in this unpredictable society also has a significant positive effect. Controlling for Occupation of First Job. To consider the effects of adolescent noncompliance further, I analyze occupational status in 1997-1998 as a dependent variable, controlling for occupational status of the first job before 1992 (Table 5.8). I include the same control variables as above, as well as dummy variables for not having a job before 1992, working in agriculture, working in sales or service or working as a clerk, semiprofessional, professional, or manager in the first job before 1992 (the omitted category for first job is worker). This model essentially measures career change after the first job before 1992.
© Heili Pals September 2006
104
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.8. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998 Controlling for Occupational Status of First Job before 1992, Relative to Workers (PG, Waves 1 & 4; Valid N=6,321).
Adolescent noncompliance
No Job
Agricultural Worker
Clerk, Sales, Service
Semiprofessional
Professional
.24
-.30
.35
-.37
.23
.56†
.30† .06 -.03 .00
2.14††† .06† .15 .02
.88††† .24††† -.13 -.01
1.00††† .57††† .12 .07†††
-.28† .43††† .29† .07††
.58††† -.31
-.11 .15
.07 .03
.05 .09
.01 .14
-.02 .39†† -.60††† .20 -.01 .59 .67†††
-.03 .24† -.43††† .22† .27 .61 .24
.01 -.04 .03 .09 .49 .61 .29†
-.17 .12 -.09 .24† .46 -.32 .04
.30†† .29†† -.41†† .03 .82† -.28 .38†
1.47††† -1.21††† .19 -.21 .72†† -.19
-2.26††† -3.49††† -2.49††† -2.60††† -2.37††† -1.93†††
.88†† -.84††† 1.80††† .30 .93†† .66
.52 -1.09††† -.60† 2.02††† .24 -.60
1.47†† -.71† 1.17††† .83† 2.82††† .95†
1.24†† -.57† 1.31††† .81† 2.02††† 2.80†††
.12
.48
Control Variables Female 1.24††† Education (years) .11††† Grade point average -.12 Parent's education (years) .05†† Place of residence Rural -.06 City .10 Region Estonia -.26†† Latvia -.07 Belarus -.35††† Kurgan .18† Self-esteem .17 Locus of control .08 Second economy .31† First Job Before 1992 No job Agricultural worker Sales , service, clerk Semiprofessional Professional Manager Intercept
L.R. χ2 vs. null model (d.f.=102) L.R. χ2 vs. model with control variables (d.f.=6)
-1.89
-.82
-1.47
Manager
-1.37
6603.44††† 10.81†
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, selfesteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
105
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Controlling for years of education and first job, grades in secondary school have a significant effect on becoming a manager or a professional by 1997-1998. Similarly, parental education, holding all other variables constant and controlling for early attainment, has a significant positive effect on the likelihood of becoming either a professional or a manager. As expected, the dummy variables measuring occupation before 1992 have significant effects for many categories of occupational status in 1997-1998. This means that the occupation of the first job is a strong predictor of people’s career achievement later on in life. For most occupational categories, with the exception of agricultural worker, there is a strong tendency to continue in the occupation of the first job. However, the effects of adolescent noncompliance are similar to the effects in the model without controls for the first job presented in Table 5.7. Adolescent noncompliance, even when controlling for the first job, increases the chances of becoming a manager by 1997-1998. Adolescent noncompliance does not increase the chances of being without a job, nor does its coefficient approach statistical significance (p=0.28) when first job is controlled. Thus, the negative effect of adolescent noncompliance on the occupation in 1997-1998 is not present when we control for the first job. Rather, after the collapse of the Soviet Union, the most noteworthy effect of adolescent noncompliance is positive: it increases the chances of becoming a manager.
© Heili Pals September 2006
106
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
ADOLESCENT NONCOMPLIANCE AND NOT HAVING A JOB The previous analysis indicates both positive and negative effects of adolescent noncompliance on occupational status and economic sector. Those who were noncompliant as adolescents had a greater likelihood of being managers in 1997-1998, but at the same time, they had a weak tendency toward not working in 1997-1998 (p=.052). A similar result was described in the second section of this chapter, when economic sector was used as a dependent variable (Table 5.5). Adolescent noncompliance increases the chances of working in the private sector, but also the likelihood of not working (p<.001). The latter result seems to support the general notion in the literature that adolescent noncompliance has negative effects on life outcomes. The effect of adolescent noncompliance may, however, be U-shaped. That is, it may increase the likelihood of two very different statuses: being a manager and being out of a job. If so, both the usual theory of the effects of adolescent noncompliance and my argument connecting adolescent noncompliance with changes in society find some support. Deconstructing the No Job Category. There are various reasons why someone might report not having a main job. Therefore, one cannot conclude that the effects of adolescent noncompliance are negative as well as positive from the analysis of the combined category of no job alone. “Not having a job” includes women on maternity leave, people who are in school, those who do not have a full-time job but are engaged in entrepreneurial activities, and those who are not working for health reasons, as well as those who do not have a job but are looking for a job.
© Heili Pals September 2006
107
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Being without a job and looking for one is the closest equivalent to the official U.S. definition of unemployment. The other categories, such as maternity leave, being a housewife, being an entrepreneur, or just being out of a job and not looking are instances where there is a certain willingness to be without a job. Those who do not have a job but are looking for one are clearly people who would like to have a job, but do not for some reason. The effects of adolescent noncompliance might be rather different for these different groups. Based on the classical view, adolescent noncompliance would be expected to increase the likelihood of being without a job but looking for one rather than being without a job voluntarily. The argument that the effects of noncompliance vary by the societal context suggests that adolescent noncompliance should have a positive effect on being without a main job but still be engaged in entrepreneurial activities. It does not suggest that adolescent noncompliance should increase the likelihood of being without a job but looking for one. To understand the relationship between adolescent noncompliance and unemployment in more detail, I divided the “no job” category in 1997-1998 into 3 relatively homogenous groups: 1) Unemployed: those who do not have a job, but are looking for a job; 2) Those who do not have a main job, but claim to be either self-employed or engaged in entrepreneurial activities; and 3) People who are out of labor force including those on maternity leave, housewives, students, those not working due to health reasons, and those who are not working and not looking for a job.
© Heili Pals September 2006
108
Chapter 5: Effects of Noncompliance on Labor Market Outcomes About a fifth of the sample (1,529 people) did not have a job in 1997-1998 (see Table 5.9). However, only a third of those (416) were unemployed. Another 6 percent (94) claimed not to have a job but to be involved in entrepreneurial activities. About two-thirds of those who did not have a job were out of labor force for various personal reasons (e.g., poor health, being a student, housewife, or maternity leave).
Table 5.9. Break-down of No Job Category.
No job in wave 4 and …
Frequency Percentage
… Looking for a job … Entrepreneur … Out of labor force Total
416 94 1,019
6.2 1.4 15.1
1,529
22.6
Percentage from those without a job 27.2 6.1 66.6 100
Source: Paths of a Generation, Wave 4.
Regression Results. I estimated a multinomial logistic regression model with the same control variables, but I used the more detailed categories of occupation and economic sector as dependent variables. The other categories for occupation and economic sector remain the same as in Tables 5.5 and 5.7, respectively. Table 5.10 presents the results for the “no job” category, broken down into the three subgroups for economic sector in 1997-1998.
© Heili Pals September 2006
109
Table 5.10. ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, Effects Relative to the State Sector (Source: PG, Waves 1 & 4; Valid N = 6,256). Private Sector No Job Stock Company OLF1 Employee Entrepreneur Unemployed Entrepreneur Adolescent noncompliance .21 .82††† 1.16††† .36 1.29† .43† Control Variables Female Education (years) Grade point average Parents’ education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control Participation in second economy Intercept L.R. χ2 vs. null model (d.f.=84) L.R. χ2 vs. model with control variables (d.f.=6) †
-.54††† -.13††† -.04 .02
-.35††† -.13††† .00 .03†
-1.02††† -.07†† -.07 .05††
-.34†† -.25††† -.14 .03
-.73†† -.09 -.44† .09†
.84††† -.17††† -.21† .04††
-.21† .33†††
-.28†† .55†††
-.18 .48†††
-.24† .14
-.43 .25
-.15 .28††
1.52††† .26†† -.96††† -.12 -.06 -.93†† .24† -1.04
-.10 1.19††† -.68††† -.32† .37 .39 -.88††† -1.58
-.30† .82††† -.45††† .31†† .43 -.05 1.52††† -1.61
-.07 .33†† -.84††† .13 .02 .03 .30 -1.81
1.02††† -.52† .14 -.50 .21 .21 1.44††† -3.44
.18† .07 -.25††† -.17† -.16 -.46 .28† -1.62
2036.25††† 27.19†††
110
p < .05, †† p < .01, ††† p < .001 (1-tailed test); 1 Out of labor force Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 5: Effects of Noncompliance on Labor Market Outcomes An interesting result emerges from this analysis. Adolescent noncompliance does not have a significant effect on the likelihood of traditional unemployment. It does increase the chances of being out of labor force (e.g., a student, on maternity leave, being a housewife, not looking for a job, and not working for health reasons). Furthermore, adolescent noncompliance does have a significant positive impact on not having a main job and being engaged in entrepreneurial activities. There is a similar result for occupation in 1997-1998 (see Table 5.11). Adolescent noncompliance increases the chances of not having a main job and being engaged in entrepreneurial activities as compared to the chances of being a worker. In contrast to the results for economic sector, adolescent noncompliance does not increase the chances of being out of the labor force as compared to the chances of being a worker. There is no significant effect for adolescent noncompliance on being unemployed. Thus, the positive effect of adolescent noncompliance on the all-inclusive category of “no job” seems to be mainly due to respondents being involved in entrepreneurial activities as a side business. This result shows that adolescent noncompliance in the former Soviet regions did not have negative effects on labor force outcomes; that is, it did not increase the likelihood of being unemployed in the traditional sense.
© Heili Pals September 2006
111
Table 5.11. ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998, Relative to Workers (PG, Waves 1 & 4; Valid N=6,321). AgriClerk, No Job SemiProfessional Manager cultural Sales, professional Unemployed Entrepreneur OLF1 Worker Service Adolescent noncompliance -.23 .43 -.12 .30 .60† .19 1.05† .36 Control Variables Female .42†† 3.03††† 1.67††† 1.75††† .29† 1.01††† .70†† 2.26††† Education (years) .08† .14††† .39††† .88††† .65††† .15††† .36††† .25††† †† †† Grade point average -.02 .14 -.02 .27 .33 .01 -.36 -.06 Parents' education (years) -.04 .03 .01 .09††† .08††† .05† .11†† .07††† Place of residence Rural .94††† -.05 .04 .08 .03 -.04 -.26 .05 †† City -.58 .19 -.06 .05 .13 -.02 .07 .17 Region Estonia .34†† .17 .17† .02 .53††† -.30† .87††† .02 ††† †† †† ††† † † Latvia .63 .25 .06 .23 .35 .20 -.67 -.03 Belarus -.91††† -.46††† .02 -.16 -.53††† -.74††† .33 -.17† Kurgan region of Russia .08 -.02 .02 .14 -.01 .25† -.51 -.09 †† † †† Self-esteem -.15 .38 .76 .55 1.11 .30 .65 .21 Locus of control .67 .20 .91† -.76† -.44 .20 .46 -.26 Participation in second economy .62††† .22 .23 .02 .35† .14 1.31††† .20 Intercept -2.02 -2.24 -.95 -1.28 -1.07 -1.37*** -3.06 -1.28 4225.88††† L.R. χ2 vs. null model (d.f.=112) 2 12.12 L.R. χ vs. model with control variables (d.f.=8) †
112
p < .05, †† p < .01, ††† p < .001 (1-tailed test) ; 1 Out of labor force Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
NONLINEARITY OF THE EFFECTS OF ADOLESCENT NONCOMPLIANCE The graphs of predicted probabilities presented in the first sections of this chapter (Figures 5.1 through 5.4) show the effects of adolescent noncompliance on labor market outcomes under the assumption that a unit change in noncompliance increases (or decreases) the log odds of each particular outcome to the same extent. However, it is conceivable that a certain threshold exists: adolescent noncompliance below this threshold could have positive effects on labor market outcomes whereas adolescent noncompliance above the threshold might not.16 If different levels of adolescent noncompliance have effects in opposite directions, then the logistic regression model incorrectly estimates the relationship between the adolescent noncompliance and the dependent variables. To test whether the effects of adolescent noncompliance vary with the level of noncompliance, I performed a variety of tests: 1) polynomial logistic regressions, 2) a Box-Tidwell transformation test, 3) linear splines, and 4) categorization of the noncompliance index.
16
It should be kept in mind that the adolescent noncompliance index used here
includes only non-criminal behaviors. More extreme forms of deviance are not included.
© Heili Pals September 2006
113
Chapter 5: Effects of Noncompliance on Labor Market Outcomes The results for the polynomial regression models17 and the Box-Tidwell tests18 are very similar. Neither test suggests that the relationship between adolescent noncompliance and the labor market outcomes examined here is non-linear. The only exception is self-employment. Both the square of adolescent noncompliance (in polynomial regression) and the interaction term (in the Box-Tidwell test; see footnotes 17 and 18) have a significant effect in the regression model on self-employment. It implies that the effect of adolescent noncompliance is different at higher levels of noncompliance. In particular, a high level of noncompliance inhibits rather than increases the chances of becoming self-employed. Third, I transformed the noncompliance index into three linear splines with knots at .35 and .56 of the scale, which runs from 0 to 1. The coefficient for the first spline (the original noncompliance variable) gives the baseline effect of adolescent noncompliance. Most of the spline variables do not have significant effects on the labor market outcomes. However, the second knot has a significant and positive effect for manager in the model of occupational status in 1997-1998. This means that the effect of adolescent noncompliance is greater at a medium level of noncompliance. The third knot is
17
The polynomial regression model includes both the original independent variable
and its square (second-order polynomial or quadratic regression). This method is often highly criticized due to the multicollinearity problems it creates. 18
The Box-Tidwell transformation adds an interaction term that is the cross-product of
adolescent noncompliance and its natural logarithm [X ln(X)]. If this interaction term is significant, then there is nonlinearity.
© Heili Pals September 2006
114
Chapter 5: Effects of Noncompliance on Labor Market Outcomes significant and negative for self-employment. This indicates, similar to the results of the polynomial regression and Box-Tidwell tests, that there is no evidence that extreme adolescent noncompliance increases the chances of becoming self-employed. All three methods do not directly allow for contrasting the effects of low and high levels of adolescent noncompliance with the middle level of noncompliance. For this comparison, I created two dichotomous variables from the original adolescent noncompliance index: low noncompliance (39 percent, everyone with noncompliance less than .35) and high noncompliance (19 percent, everyone with noncompliance higher than .56). These two categories were compared to the middle level of adolescent noncompliance (41 percent, everyone with noncompliance between .35 and .56). This test evaluates differences in the adolescent noncompliance effect at the low and high ends of the noncompliance scale. It allows for different effects when comparing those who were relatively compliant as adolescents and those who were noncompliant against those with middle levels of noncompliance.
© Heili Pals September 2006
115
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.12. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Binary Logistic Regression Models of Owning a Firm in 1997-1998 and Ever Being Selfemployed since 1992. (Source: PG, Waves 1 and 4).
Firm Owner
Selfemployed
-.24†† .10
-.35††† .05
-.19 -.06
-.76††† -.06†† .07 .00
-.34††† .12††† -.04 .09†††
-.71††† .00 -.25† .02
.23† .12
-.46††† .03
-.18 .30†
.55††† .42††† -.37††† -.37††† .34 .43
.89††† .46††† -.41††† -.10 .79†† .29 .79†††
-.95††† .50††† -.02 .49††† .25 1.22†† 1.16†††
Second Economy Low adolescent noncompliance High adolescent noncompliance Control Variables Female Education (years) Grade point average Parent's education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control In second economy Intercept
-2.01 †††
2
-2.35 †††
198.86 L.R. χ vs. null model (d.f.=15) 2 9.59†† L.R. χ vs. model with control variables (d.f.=2)
387.53 13.51††
Valid N
6,247
6,305
-2.79 228.80††† 2.01 6,163
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to middle level of adolescent noncompliance, general secondary school, and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, and locus of control are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5. Degrees of freedom for second economy are 14 and 2, respectively.
© Heili Pals September 2006
116
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.13. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Economic Activity in 1997-1998, Effects Relative to “Employees” (Source: PG, Waves 1 and 4; N=6,305). “Economically “Working “Nonworking Inactive” Entrepreneur” Entrepreneur” Low adolescent noncompliance High adolescent noncompliance Control Variables Female Education (years) Grade point average Parent's education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control In second economy Intercept 2
-.07 .09
-.21†† .09
.15 .49†††
1.07††† -.15††† -.11 .03†
-.75††† .02 -.07 .04††
.42†† -.14††† -.26† .02
-.05 .16† -.18†† -.04 -.38††† .24†† -.09 -.10 -.41 -2.13
L.R. χ vs. null model (d.f.=45) L.R. χ2 vs. model with control variables (d.f.=6)
.09 .07
.13 .08
-.17† .17† -.25†† .27†† .43† .52 3.17†††
-.77††† -.07 -.30† .19 -.81† .32 3.32†††
-1.36
-3.81 †††
1779.51 16.85††
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to middle level of adolescent noncompliance, general secondary school, and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, and locus of control are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
117
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Table 5.14. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, Effects Relative to State Sector (Source: PG, Waves 1 and 4; N=6,235).
Low adolescent noncompliance High adolescent noncompliance Control Variables Female Education (years) Grade point average Parent's education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control In second economy Intercept 2
Private Sector
No job
Stock Company
-.11 .18†
-.02 .03
-.32††† .04
-.21† .31††
.38††† -.19††† -.20†† .04††
-.54††† -.13††† -.04 .03†
-.35††† -.13††† -.03 .04†
-1.03††† -.07†† -.08 .05††
-.20† .24††
-.21† .34†††
-.27† .56†††
-.19 .49†††
.18†† .11 -.37††† -.13 -.10 -.28 .39†††
1.52††† .27††† -.96††† -.13 -.10 -.90†† .22
-.09 1.21††† -.68††† -.36†† .35 .40 -.88†††
-.30†† .83††† -.47††† .31†† .40 -.05 1.50†††
-.87
L.R. χ vs. null model (d.f.=60) L.R. χ2 vs. model with control variables (d.f.=8)
-1.04
Employee Entrepreneur
-1.47
-1.57 †††
1780.81 26.87†††
†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to middle level of adolescent noncompliance, general secondary school, and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, and locus of control are centered at the sample mean of the variable. Parents’ education is measured as years of education of the mosteducated parent. GPA ranges from 3.5 or less to 5.
© Heili Pals September 2006
118
Table 5.15. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Occupational Status of First Job before 1992, Relative to Workers (PG, Waves 1 & 4, Valid N=6,303). No Job Low adolescent noncompliance High adolescent noncompliance Control Variables Female Education (years) Grade point average Parents' education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control Participation in second economy Intercept L.R. χ2 vs. null model (d.f.=90) L.R. χ2 vs. model with control variables (d.f.=12) †
-.26† .19
Agricultural Clerk, Sales, Worker Service -.01 -.21† .03 -.07
1.98††† .43††† .16 .09†††
.62††† .07† .09 -.06††
3.50††† .24††† .00 .03
.21 .21
.86††† -.63†††
.09 .16
1.00††† .41††† .55††† -1.05††† .54 -.23 -.01 -2.18
.82††† .62††† -1.06††† -.06 -.16 .48 .08 -1.99
.49††† .12 -.33††† -.54††† .50 -.92† .05 -2.88
Semiprofessional -.05 .14
Professional
2.17††† .48††† .18† .03†
2.13††† 1.05††† .49††† .07†††
.00 -.16
.06 -.08
.14 -.03
.40††† .41††† -.31†† -.09 .47 -1.05†† -.12 -2.07
.96††† .46††† -.67††† -.02 1.44††† -.14 -.10 -2.49
.45††† .22†† -.07 -.16 .88††† .68† -.09 -1.35 5186.45††† 14.63
-.13 .08
Manager -.03 .28 .71††† .74††† .05 .06†
119
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to middle level of adolescent noncompliance, general secondary school, and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, and locus of control are centered at the sample mean. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Table 5.16. Testing the Nonlinearity of Adolescent Noncompliance: ML Estimates of Multinomial Logistic Regression Model of Occupational Status in 1997-1998, Effects Relative to Workers (Source: PG, Waves 1 & 4, Valid N = 6,300). No Job Low adolescent noncompliance High adolescent noncompliance Control Variables Female Education (years) Grade point average Parents’ education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control Participation in second economy Intercept 2
L.R. χ vs. null model (d.f.=90) L.R. χ2 vs. model with control variables (d.f.=12) †
Clerk, Sales, Service -.21† -.05
Semiprofessional -.06 -.13
Professional
-.19† .06
Agricultural Worker .03 -.06
-.19† -.02
-.44††† -.08
1.78††† .22††† -.05 .06†††
.44††† .08† -.03 -.04
3.03††† .14††† .13 .03
1.68††† .38††† -.02 .01
1.75††† .88††† .28† .09†††
.29† .65††† .33†† .08†††
.00 .09
.94††† -.58††
-.06 .19
.04 -.08
.33†† .63††† -.91††† .08 -.12 .66 .63†††
.17† .25†† -.48††† -.01 .40 .25 .25
-.01 -.01 -.29††† -.03 .29 -.05 .31† -.42
-2.02
-2.14
.17† .07 .01 .02 .78† .97† .25 -.90 3954.11††† 20.38†
Manager
.07 .03
.04 .12
.00 .23†† -.18† .15 .58† -.72† .04
.52††† .36††† -.54††† .00 1.15††† -.44 .34†
-1.18
-.88
120
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to middle level of adolescent noncompliance, general secondary school, and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, and locus of control are centered at the sample mean. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 5: Effects of Noncompliance on Labor Market Outcomes
Categorizing
Levels
of
Noncompliance.
Categorizing
adolescent
noncompliance levels reveals the most interesting results (Tables 5.12 through 5.16). With the exception of self-employment and occupation at the first job, the estimated coefficient for the lowest level of adolescent noncompliance is significant and negative. This means that those who were compliant as adolescents have a lower likelihood of participating in the second economy before 1992 and a lower likelihood of owning a firm, being an entrepreneur, working in the private sector, and being a manager in 1997-1998 than those in the middle level of adolescent noncompliance. This result clearly illustrates the disadvantages that result from adolescent compliance in a society with frequent changes. It tells that individuals who are accustomed to following rules and suggestions from authorities, are less successful in operating in an environment where norms and laws are constantly changing. DEPENDENCE OF EFFECTS OF ADOLESCENT NONCOMPLIANCE ON STATUS The analysis above has shown that in general, adolescent noncompliance in regions of the former Soviet Union does increase the chances of becoming involved in entrepreneurial activities, becoming a manager, and working in the private sector. The question arises whether the effects of adolescent noncompliance are similar for different subgroups. The hypotheses raised in Chapter 3 state that the effects of adolescent noncompliance vary with the status of the student. Namely, high and low status students are likely to benefit more from adolescent noncompliance than middle status students do. Status among the secondary school students in the former Soviet regions can be operationalized in terms of gender, grades, type of secondary school, and parental education.
© Heili Pals September 2006
121
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Adolescent noncompliance is related to all four status measures for secondary school students in the five former Soviet Union areas studied here. Females have lower levels of adolescent noncompliance than male students, with a difference of approximately .08 in the mean noncompliance between the two genders (p<.001, t-test comparing the means of noncompliance for males and females). Students whose parents have more education tend to have higher levels of adolescent noncompliance. The mean of adolescent noncompliance varies from .36 for those whose parents had 4 years of education (258 respondents) to .42 for those whose parents had 15 or more years of education (1621 respondents). In contrast, those with higher grades are more likely to be compliant. The mean of noncompliance for students with a GPA of 3 is .47; while students with a GPA of 5 have a mean noncompliance level of .31. Vocational school students have slightly higher levels of noncompliance than average. Similarly, those in specialized or academic secondary schools have slightly higher levels of noncompliance. Those in general secondary schools tend to have lower levels of adolescent noncompliance. To further test whether the effects of adolescent noncompliance on labor market outcomes differ by the status of the student, I conduct multivariate analyses with interaction terms between different measures of student status and adolescent noncompliance. Thus, the simplest regression, using gender as the status measure, includes a dichotomous indicator of gender, the continuous measure of adolescent noncompliance, and an interaction term between gender and noncompliance. I estimated a series of regression models, both with the regular controls used elsewhere in this research as well as without the controls, on participation in the second
© Heili Pals September 2006
122
Chapter 5: Effects of Noncompliance on Labor Market Outcomes economy, entrepreneurial activities, working in the private sector, and occupational status both before 1992 and in 1997-1998. In most instances, the interaction term between gender and adolescent noncompliance was not significant, indicating that noncompliance has similar effects on labor market outcomes for females and males in the former Soviet regions19. I recoded GPA into two dichotomous variables, contrasting higher and lower GPAs with average grades. The middle GPA group is the most numerous: 70 percent of the sample has GPAs from 3.75 to 4.75. Roughly 15 percent of the sample has a low GPA (less than 3.75) and a similar percentage has high GPAs (4.75 or higher). The model includes two interaction terms: the interaction between the low GPA group and adolescent noncompliance and the interaction between the high GPA group and adolescent noncompliance. Again, the results show no significant differences in the effects of noncompliance by student status. Interactions between parents’ educational level and adolescent noncompliance show some differences between the effects of noncompliance by student status. Adolescent noncompliance increases the likelihood of engaging in the second 19
The interaction term between gender and adolescent noncompliance was significant
only for models on occupational status in 1997-1998. Noncompliance has a stronger positive effect for females than males on the likelihood of being an agricultural worker or being in a sales or service job or a clerk compared to being a worker. However, the effect of noncompliance does not differ by gender for higher prestige jobs such as managers and professionals.
© Heili Pals September 2006
123
Chapter 5: Effects of Noncompliance on Labor Market Outcomes economy more for those whose parents had only a basic education (less than 8 years) as compared to those whose parents had 8 or 9 years of education. At the same time, for low status students, adolescent noncompliance has a negative effect on chances of being an employee in the private sector as compared to working in the state sector. Thus, with regard to participation in the second economy, adolescent noncompliance is more beneficial for the lowest student status group, while for the private sector, adolescent noncompliance is less beneficial for the lowest student status group when status is defined by parental education. Finally, I measure student status based on the type of secondary school. I compared vocational, specialized, and academic secondary schools to general secondary schools. Interaction terms between the type of school and adolescent noncompliance were not significant, with the exception of economic sector and occupational strata. Adolescent noncompliance has a significant negative effect on vocational school students’ chances of being an employee in the private sector. Similarly, adolescent noncompliance has a significant negative effect for academic secondary school students on the chances of being an entrepreneur in the private sector. Finally, adolescent noncompliance has a significant positive effect for those who went to vocational school on the chances of becoming a manager and a semiprofessional. Status as a Student and Adolescent Noncompliance. To summarize, I used four different operationalizations of status as a student: 1) gender, 2) GPA in secondary school, 3) parents’ level of education, and 4) type of secondary school. The previous literature on noncompliance alludes to two possible effects of noncompliance: 1) its
© Heili Pals September 2006
124
Chapter 5: Effects of Noncompliance on Labor Market Outcomes effect increases steadily with increases in the status of the actor (i.e., it is highest among those in the high-status group and lowest among those in the low-status group); 2) its effect is U-shaped (i.e., the effects of noncompliance are greatest at both extremes of status and least among the middle-status actors). The results of these analyses were inconclusive. Most frequently, the effects of adolescent noncompliance do not vary by the different measures of student status. In particular, when status is measured by gender or GPA, there is no apparent interaction between the effects of adolescent noncompliance and status as a student. However, when using parental education as a measure of student status, I observe both types of patterns. There seems to be a U-shaped relationship between parental education and adolescent noncompliance when using the likelihood of participating in the second economy as a dependent variable, and a relatively linear increase in the effect of adolescent noncompliance as the status as a student increases when using the likelihood of working in the private sector as a the dependent variable. A somewhat Ushaped relationship is found when looking at the chance of working in the private sector and defining status by the type of secondary school. However, here both lowstatus students (those in vocational school) and high-status students (those in an academic secondary school) seem to benefit less from adolescent noncompliance than middle-status students. Thus, the few significant effects do not clearly support either view: neither a U-shaped relationship, nor linearly increasing effects are consistently observed. For most outcomes and most operationalizations of student status, I fail to find an interaction effect between adolescent noncompliance and student status. A possible explanation for the lack of an interaction effect for adolescent noncompliance
© Heili Pals September 2006
125
Chapter 5: Effects of Noncompliance on Labor Market Outcomes and student status on success in the labor market of the former Soviet Union is related to Merton’s strain theory. In the former Soviet Union, the goals and values of the new society were developed and spread relatively rapidly. However, did society provide the means to achieve these goals? The legislature was slow to act, the redistribution of capital was a slow process, and people did not have the right education needed for the goals of this new society. Thus, there was a real lack of legitimate means to reach the new goals. Therefore, according to Merton’s strain theory, it is likely that people resorted to means that were not widely accepted by society in order to achieve the new goals. Merton’s strain theory suggests that status differences could be the source of the structural constraints. However, in the former Soviet regions during transitional period with the first most rapid changes, most people lacked the legitimate means regardless of their past status. An alternative explanation for the lack of interaction between student status and adolescent noncompliance could lie in the operationalization of status. There are different ways to operationalize status, and it could be that the options provided by Paths of a Generation data are not the most suitable for the analysis of adolescent noncompliance and status. In addition, student status might not be the most relevant status to capture the true relationship between status and adolescent noncompliance. CONCLUSION: LABOR MARKET OUTCOMES The
results
presented
in
this
chapter
demonstrate
that
adolescent
noncompliance has positive effects on different labor market outcomes in the unpredictable conditions of transitional former Soviet regions. The first two sections
© Heili Pals September 2006
126
Chapter 5: Effects of Noncompliance on Labor Market Outcomes clearly show that adolescent noncompliance increases the chances of participating in new types of employment opportunities that appeared due to the changes in the economic and political regime. Adolescent noncompliance has positive significant effects on the likelihood of participating in entrepreneurial activities before and after the collapse of the Soviet regime and the chances of working in the private sector compared to working in the state sector. As new opportunities, these employment options were not fully covered by legislation and the norms about such work were unclear, which is consistent with the argument that the effect of noncompliance depends on social context. Furthermore, a multivariate model of occupational status after the collapse of the Soviet Union revealed that adolescent noncompliance increases the chances of being a manager, even when controlling for first job. In addition, both controlling for first job and dividing the “no job” category in 1997-1998 into subcategories indicate that adolescent noncompliance does not have negative effects on labor market outcomes in the rapid transition after the collapse of the Soviet Union: adolescent noncompliance does not increase the likelihood of being unemployed in the traditional sense of unemployment. An additional analysis about the nonlinearity of adolescent noncompliance revealed that at least part of the advantage of adolescent noncompliance in the labor market is due to the disadvantages of being compliant in adolescence. Regressions with categorized levels of adolescent noncompliance show that a low level of adolescent noncompliance (i.e., compliance) has negative impacts on several of the labor market outcomes discussed here.
© Heili Pals September 2006
127
Chapter 5: Effects of Noncompliance on Labor Market Outcomes Finally, the hypotheses about the different effects of adolescent noncompliance for people from different statuses as a student did not find consistent support. Most of the interaction terms between different measures of student status and adolescent noncompliance had insignificant effects. In the few instances where there was a significant effect, the results are not consistent across outcomes and different measures of student status.
© Heili Pals September 2006
128
CHAPTER 6: STRUCTURAL DIFFERENCES IN THE EFFECTS OF ADOLESCENT NONCOMPLIANCE The analysis in Chapter 5 did not compare the effects of adolescent noncompliance across societies with different degrees of social change. Rather, it was assumed that the effects of noncompliance would be positive only in societies with frequent changes and that these effects would not be found in societies that are not changing rapidly. This chapter offers three tests for the varying effects of adolescent noncompliance on labor market outcomes in structures with different levels of unpredictability: 1) I estimate the effect of adolescent noncompliance on occupational status of the first job prior to 1992, before the major changes occurred in the former Soviet regions; 2) I compare the effects of adolescent noncompliance on occupational status in the private and state sectors; and 3) I compare the effects of adolescent noncompliance on labor market outcomes in different regions of the former Soviet Union. OVER-TIME DIFFERENCES: OCCUPATIONAL STATUS OF THE FIRST JOB The Paths of a Generation data set allows me to analyze occupational status at more than one point in time. I use the status of the first job before 1992 to compare the effect of adolescent noncompliance during a relatively stable period (before the collapse of the Soviet Union) to its effect during a relatively unstable period (19971998). Based on the hypotheses proposed earlier, adolescent noncompliance should have either no effect or a negative effect on occupational status before the collapse of
© Heili Pals September 2006
129
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance the Soviet Union. Of course, unique mechanisms of attainment that are not present in long-term occupational attainment might occur for the first job. Most respondents (90 percent) had a job at some point before the collapse of the Soviet Union. About 30 percent of respondents began as workers, 8 percent worked in agriculture, and 12 percent were in sales or service or worked as a clerk. A considerable number of people began at a higher point on the occupational ladder: 16 percent worked as semiprofessionals, 20 percent worked as professionals, and about 4 percent were managers in their first job before 1992. Table 6.1 presents the results of multinomial logistic regression models of occupational status of the first job before 1992. The effects of the control variables are fairly similar to their effects on occupational status in 1997-1998 (see Table 5.7). However, before 1992, adolescent noncompliance increases the likelihood of not working compared to the likelihood of working as a worker. Adolescent noncompliance does not influence the chances of being a manager, professional, or semiprofessional in the first job. Thus, other than the effect on not working, there is no clear effect of adolescent noncompliance on occupational status before the collapse of the Soviet Union. However, the category of being without a job is a broad category, including not only unemployed, but also students, those at home with their children, and other forms of being without a job, as mentioned in Chapter 5. Therefore, the meaning of the negative effect of adolescent noncompliance on not having a job before 1992 is not completely clear.
© Heili Pals September 2006
130
Table 6.1 ML Estimates of Multinomial Logistic Regression Model of Occupational Status of the First Job before 1992, Relative to Workers (Source: PG, Waves 1 & 4; Valid N=6,324). No Job Agricultural Clerk, Sales, SemiProfessional Manager Worker Service professional -.03 .43 .37 .26 .06 Adolescent noncompliance .82†† Control Variables Female Education (years) Grade point average Parents' education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan region of Russia Self-esteem Locus of control Participation in second economy Intercept L.R. χ2 vs. null model (d.f.=84)) L.R. χ2 vs. model with control variables (d.f.=6) †
1.98††† .43††† .16 .08†††
.61††† .07 .08 -.06††
3.51††† .24††† .01 .03
.22 .21
.86††† -.63†††
.09 .16
1.00††† .40††† .56††† -1.05††† .55 -.23 .01 -2.25
.82††† .62††† -1.06††† .06 -.17 -.49 .09 -1.99
.49††† .12 -.33††† -.55††† .50 -.93† .04 -2.98
2.17††† .48††† .18† .03†
2.12††† 1.05††† .48††† .07†††
.01 -.16
.05 -.08
.13 -.02
.41††† .41††† -.31††† -.09 .45 -1.04†† -.12 -2.10
.97††† .45††† -.68††† -.02 1.40†† -.10 -.09 -2.43
.44††† .22††† -.06 -.16 .88†† .67† -.09 -1.34 5182.1††† 10.2
.68††† .74††† .01 .06†
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
131
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance Prior to the collapse of the Soviet regime in 1992, changes in legislation, norms, and the opportunity structure occurred relatively slowly. Therefore, analyzing occupational status before the collapse of the Soviet Union allows me to evaluate the effects of adolescent noncompliance in a society with a relatively high level of predictability and a low level of change. In contrast, the models of occupational status in 1997-1998 (Tables 5.7 and 5.8) tested the effects of noncompliance in a society with a high level of unpredictability. The results indeed reveal differences in the effects of noncompliance on occupational status. Adolescent noncompliance increased the chances of being a manager after the collapse of the Soviet Union but not before; of course, respondents were also older in 1997-1998 than before the Soviet Union collapsed, and older people are in general more likely to be managers. Thus, adolescent noncompliance might have some positive effects on occupational status in a highly unpredictable society, whereas it might not directly influence occupational status in a stable society. This is not a perfect test because the sample analyzed here consists of only one generation. Analyzing attainment at different times is affected by both changes in society and changes in the life course. Thus, change in individual life course might be influencing the different effects of noncompliance before 1992 and in 1997-1998. Furthermore, the mechanisms behind entry into the first job might be different from the attainment mechanisms for later occupational success. DIFFERENCES IN THE EFFECTS OF ADOLESCENT NONCOMPLIANCE BY ECONOMIC SECTOR Contrasting attainment in the private and state sectors of the former Soviet Union allows me to compare attainment in structures that vary in their
© Heili Pals September 2006
132
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance unpredictability. After the collapse of the Soviet Union, the private sector was less regulated and changed more rapidly than the state sector. Much of the state sector retained the practices and organizational structures of the Soviet era. For this reason, I expect that adolescent noncompliance has either no effect or only weak positive effects on attainment in the state sector, but has strong positive effects on attainment in the private sector (hypothesis 3c). To test this hypothesis, I modify the multinomial logistic regression model of occupational strata in 1997-1998. I divide the occupational categories into two groups: the occupational stratum in the private sector (stock companies, self-employment, and working in private firms) and the corresponding stratum in the state sector (state companies, state-financed firms, and kolkhozes). The effects of noncompliance are the same across public and private sectors for lower-status occupational outcomes, such as being an agricultural worker or a clerk. However, the effects of noncompliance on both professionals and managers differ in the two sectors. Therefore, the final model distinguishes between the sectors for professionals and managers, but does not differ for other occupational strata. The final dependent variable has 9 categories: those not working, workers, agricultural workers, sales/service/clerks, semiprofessionals, professionals in the state sector, professionals in the private sector, managers in the state sector, and managers in the private sector. Being a worker is the reference category.
© Heili Pals September 2006
133
Table 6.2 ML Estimates of Multinomial Logistic Regression Model of Adolescent Noncompliance on Occupational Status by Economic Sector, Relative to Workers (Source: PG, Waves 1 & 4; N=6,230). Professional Manager Agricultural Clerk, Sales, SemiNo job Worker Service professional Private State Private State Adolescent noncompliance Control Variables Female Education (years) Grade point average Parent's education (years) Place of residence Rural City Region Estonia Latvia Belarus Kurgan Self-esteem Locus of control Second economy Intercept 2
.35
-.25
-.13
1.11†††
-.19
.79†
.06
.06 .59††† .32† .10†††
.77††† .81††† .21 .07†
1.75††† .24††† -.04 .06†††
.44†† .09† -.05 -.05†
3.04††† .14††† .12 .03
1.70††† .39††† -.03 .00
1.44††† .73††† .11 .13†††
1.96††† 1.00††† .37†† .06††
.00 .07
.93††† -.57††
-.08 .15
.05 -.09
-.02 .30†
.20 -.14
-.17 .27†
.49† -.21
-.07 -.02 -.29††† .06 .30 -.04 .28†
.32†† .64††† -.91††† .09 -.10 .68 .62†††
.14 .22† -.47††† .03 .40 .41 .21
.14 .04 .02 .04 .82†† 1.13†† .24
.30† .69††† -.34† -.07 1.08†† -1.07† .16
-.14 -.05 -.04 .29† .26 -.38 -.10
.77††† .60††† -1.00††† .07 1.10†† -.73 .46†
.03 -.09 .04 .11 1.31† .30 -.06
-.39
-1.99
L.R. χ vs. null model (d.f.=112) L.R. χ2 vs. model with only control variables (d.f.=8) †
.42
-2.20
-.95
-2.12 4174.3††† 25.8†††
-2.04
-1.46
-2.57
134
p<.05, †† p<.01, ††† p<.001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance The results in Table 6.2 show the difference in the effects of adolescent noncompliance in two sectors for higher occupational strata. The shaded columns represent occupations in the private sector. The effects of adolescent noncompliance differ for professionals in the private sector and for professionals in the state sector. Adolescent noncompliance does not affect the likelihood of being a professional in the state sector, but it does increase the likelihood of being a professional in the private sector. Similarly, adolescent noncompliance does not affect the likelihood of being a manager in the state sector, but it does increase the likelihood of being a manager in the private sector as compared to the likelihood of being a worker. Those who were noncompliant in adolescence seem to be better equipped than those who were compliant to succeed in the private sector after the collapse of the Soviet Union. Both of the significant coefficients indicate that the positive effects of adolescent noncompliance are more pronounced in rapidly changing structures like those found in the private sector. The effect is not present in the state sector of the economy, which did not change as rapidly as the private sector. To understand better the results and the interpretation of the regression model, I calculate scenarios for a typical person. I vary the level of adolescent noncompliance and calculate predictions for the probability of being professionals and managers in each sector (Figure 6.1).
© Heili Pals September 2006
135
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance
Professionals .12
Predicted Probability
.1 .08 .06 .04 .02 0
Managers
Predicted Probability
.12 .1 .08 .06 .04 .02 0 0
.2
.4
.6
.8
1
Adolescent Noncompliance State sector
Private sector
Figure 6.1. Predicted Probability of Occupational Attainment by Economic Sector in 1997-1998 and Adolescent Noncompliance Source: PG, Waves 1 & 4. Predicted probabilities from the multinomial regression models presented in Table 6.2. Note: Shaded area marks 95% confidence intervals. Predicted probabilities for a hypothetical male born in a big city in the Kharkiv region of Ukraine with 12 years of education who did not participate in second economy and whose most-educated parent has 10 years of education. This hypothetical person has an average GPA, average self-esteem, and an average locus of control.
© Heili Pals September 2006
136
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance The solid lines represent the effects of adolescent noncompliance on the chances of becoming a professional or manager in the private sector. The dashed lines represent the corresponding effects in the state sector. For both professionals and managers, the flat dashed lines indicate that adolescent noncompliance has essentially no effects in the state sector. In fact, its effect on professionals in the state sector is even slightly negative, although not significant. In contrast, the probability of being a manager or professional in the private sector increases as adolescent noncompliance increases. The probability that a typical male who was compliant in adolescence is a professional in the private sector is about .02, while a typical male who was noncompliant as an adolescent has a probability of nearly .06. The probability of being a manager in the private sector increases from .06 to .10 with the change from extreme compliance to extreme noncompliance. There are still some problems with this analysis. The model presented here does not consider people who move from one sector to another. It measures only the end state: occupational status in a given industry in 1997-1998. To overcome this limitation, I control the economic sector of the first job. The results do not change even when the economic sector of the first job before 1992 is controlled. Thus, all of these results support the hypothesis that the effects of adolescent noncompliance differ across structures with different levels of change and unpredictability. Adolescent noncompliance has a positive and significant effect on occupational status in the private sector, which experienced a high degree of change, but does not have a similar effect on occupational status in the state sector, which has changed much less in comparison and continues to be highly regulated.
© Heili Pals September 2006
137
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance
DIFFERENCES IN EFFECTS OF ADOLESCENT NONCOMPLIANCE ACROSS REGIONS The previous analysis provided two tests of the effects of adolescent noncompliance in settings with different levels of change: an over-time comparison (comparing the status of the first job and the status of the job in 1997-1998) and a comparison of the effects on occupational status in the public and private sectors. Both tests have limitations. As discussed above, the over-time comparison confounds the effects of changes in the structure of society with changes over the life course. The cross-sector comparison allows a more direct analysis of the effects of adolescent noncompliance in different types of settings, but it examines only one labor market outcome: occupational status. I cannot analyze differences in the effects of adolescent noncompliance on entrepreneurial outcomes because, by definition, entrepreneurship or private business means working in the private sector. A comparison of the effects of adolescent noncompliance across regions of the former Soviet Union allows me to examine a wider variety of labor market outcomes. Different regions changed at different speeds after the collapse of the Soviet Union and therefore had different levels of societal changes. The Baltic states, Latvia and Estonia, changed the most rapidly. Therefore, I expect the positive effects of adolescent noncompliance on labor market outcomes to be greater in these regions. Ukraine, Belarus, and the Kurgan region of Russia changed at a considerably slower speed. Therefore, the effects of adolescent noncompliance on labor market outcomes in these regions are likely to be smaller and may not be positive. To test this hypothesis, I estimate multivariate models of occupational status, different measures
© Heili Pals September 2006
138
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance of entrepreneurship, and working in the private sector separately in the Baltic states and the rest of the regions. A similar regional comparison looking at income inequalities was conducted by Titma and Murakas (2004). They compared income inequalities in two groups of regions: 1) Estonia and Latvia, where the changes were very rapid (these regions adopted so-called “shock therapy”), and 2) Russia, Ukraine, and Belarus, where market reforms were introduced considerably more slowly. Belarus, and the Kharkiv and Kurgan regions have changed relatively slowly following the dissolution of the Soviet Union. For example, the current economic system in Belarus more or less replicates the system existing in Soviet times.
Table 6.3 Summary of Regional Differences in the Effects of Noncompliance. Baltic Regions:
Other Regions:
Estonia, Latvia
Belarus, Kharkiv, Kurgan
Second economy
+
NS
Firm owner
+
NS
NS
NS
Entrepreneur
+
NS
Private sector
+
+
NS
+
Self-employment
Manager in 1997-1998
+ significant at least at p<.05 (one-tailed test); NS – not significant Note: Results from binary and multinomial logistic regressions presented in Tables 6.4 to 6.8. Models control for gender, education, GPA, parental education, place of residence, selfesteem, locus of control, and participation in second economy. The significance of adolescent noncompliance is shown. Source: PG, Waves 1 & 4.
© Heili Pals September 2006
139
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance
Table 6.3 summarizes the results of binary and multinomial logistic regressions on different labor market outcomes in these two types of regions. A plus sign indicates a significant positive effect of adolescent noncompliance on the labor market outcomes; NS means that no significant effect was found. The Baltic states consistently show significant positive effects of adolescent noncompliance on most entrepreneurial activities. The coefficient fails to be positive and significant only in the case of self-employment and occupational status. The more slowly changing regions (Belarus, Kharkiv, and Kurgan), however, show different results. The coefficient for adolescent noncompliance is not significant for any measures of entrepreneurship, either before or after the collapse of the Soviet Union. Noncompliance does have significant positive effects on working in the private sector and on chances of being a manager in 1997-1998 in the slowly changing regions. The lack of the positive effect of adolescent noncompliance is expected because the changes in these regions were relatively slow and thus the level of uncertainty was lower than in the Baltic states. More detailed results are given in Tables 6.4 to 6.8. Table 6.4 presents the full results of the binary logistic regressions on participation in second economy, owning a firm, and being self-employed. The effect of adolescent noncompliance is highly significant and positively related to the chances of participating in the second economy and being a firm owner in the Baltic states. No such effects can be found in the pooled results for Belarus, Kharkiv, and Kurgan.
© Heili Pals September 2006
140
Table 6.4. ML Estimates of Binary Logistic Regression Models of Entrepreneurial Activities, by Region Group (Source: PG, Waves 1 and 4). Baltic Regions: Estonia, Latvia
Adolescent noncompliance
Second Economy
Firm Owner
.97†††
.94†††
Other Regions: Belarus, Kurgan, Kharkiv Selfemployed .61
Second Economy .49
Firm Owner .72
Selfemployed .15
Control Variables Female
-.82†††
Education (years)
-.09†† †
-.31†† .12†††
-.74†††
-.65†††
-.32† .09†
-.66†††
.02
-.03
-.24
-.16
.01
.01
.09††
.01
-.39†
Grade point average
.26
.05
Parent's education (years)
.00
.08†††
Place of residence Rural
.21
-.28†
-.06
.29
-.87††
-.49†
.16
.00
.12
.02
.28
.32
City
†
†
†
-.15
-.02
Self-esteem
.08
.60
1.15
.73
.34
.09
Locus of control
.54
.33
1.08
.16
.37
1.35†
.66††
1.39†††
.83†††
Second economy Intercept 2
L.R. χ vs. null model (d.f.=10) L.R. χ2 vs. model with control variables (d.f.=1) Valid N
-1.58 †††
97.7 11.0††† 3,099
-1.86 †††
128.9 10.1††† 3,080
.93††† -2.74 †††
70.0 2.0 3,027
-2.39 †††
38.5 1.68 3,206
-3.01 †††
72.5 2.36 3,167
-2.74 98.1††† .11 3,206
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
†
141
Table 6.5 ML Estimates of Multinomial Logistic Regression Model of Economic Activity in 1997-1998, by Region Group, Effects Relative to “Employees” (Source: PG, Waves 1 and 4). Baltic Regions: Other Regions: Estonia, Latvia Belarus, Kurgan, Kharkiv
Adolescent noncompliance Control Variables Female Education (years) Grade point average
“Working Entrepreneur”
“Nonworking Entrepreneur”
“Inactive”
.27
.85††
.39
.16
1.13†††
-.96†††
.42†
.90†††
-.19
†††
.02 †
††
“Working “Nonworking Entrepreneur” Entrepreneur” .27
.23 -.17†††
-.19†
-.13
-.27
-.09
-.12
-.03
-.17
.04
.05
-.02
.02
.03
Place of residence Rural
.10
.15
.42
-.27†
-.02
.09
.54
-.03
.07
†††
-.56†††
Parent's education (years)
City
-.29
†
.17
.07
.38
-.25
††
-.26
††
.06†
Self-esteem
-.32
.87
.05
-.06
.11
-1.13†
Locus of control
-.08
.39
.44
-.23
.67
.15
Second economy
-.42
2.94†††
2.88†††
-.45
3.79†††
3.99†††
Intercept L.R. χ2 vs. null model (d.f.=30) L.R. χ2 vs. model with control variables (d.f.=3) Valid N †
“Inactive”
-2.33
-1.22 945.3††† 8.6† 3,099
-4.01
-1.83
-1.65
-2.93
755.5††† 1.9 3,206
142
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance Table 6.5 presents the results of the multinomial logistic regression of entrepreneurial activity in 1997-1998. Again, I estimate separate models for the Baltic states and the other three regions. The inclusion of the adolescent noncompliance index improves the model fit only in the Baltic states. Similarly, adolescent noncompliance has a significant and positive effect on the chance of being an entrepreneur in the Baltic states, but not in Belarus, Kurgan, and Kharkiv. Multinomial logistic regression models of economic sector in Table 6.6 reveal a significant positive effect of adolescent noncompliance on working in the private sector in the Belarus, Kharkiv, and Kurgan regions. Thus, the positive effect of adolescent noncompliance is apparent in the fast-changing private sector, even in relatively slow-changing regions. In the Baltic states, adolescent noncompliance increases the likelihood of working in the private sector and the likelihood of working in a stock company as compared to the chances of working in the state sector. In addition, adolescent noncompliance increases the likelihood of being out of the labor force in the Baltic states. These results seem to indicate that in the Baltic states, the effect of adolescent noncompliance on the chances of working in the state sector is quite different from its effects on other sectors. A similar multinomial logistic regression models using other reference categories show that, indeed, the effect of noncompliance on the chances of working in the state sector is significantly negative when compared against any other economic sector.
© Heili Pals September 2006
143
Table 6.6 ML Estimates of Multinomial Logistic Regression Model of Economic Sector in 1997-1998, by Region Group, Effects Relative to State Sector (Source: PG, Waves 1 and 4). Baltic Regions: Other Regions: Estonia, Latvia Belarus, Kurgan, Kharkiv
Adolescent noncompliance Control Variables Female Education (years) Grade point average Parent's education (years) Place of residence Rural
Stock Company
1.05†††
1.14†††
1.11†††
-.77††† -.19†††
.19 -.29††† .02 †
.29†† †
Private Sector
.28
-.41
.76††
-.83†††
.40†††
-.26†
-.55†††
-.15†††
-.13†††
-.12†††
-.09†††
-.02
-.23††
-.23
-.07
††
.05
-.01
-.08
-.11
-.34††
-.36††
-.02
††
Self-esteem
-.20
-1.08†††
L.R. χ2 vs. null model (d.f.=30) L.R. χ2 vs. model with control variables (d.f.=3) Valid N
Stock Company
.02
.29
Intercept
No Job
.04
.16
Second economy
Private Sector
.04
City Locus of control
†
No Job
-.43 †
-.95
†
†
.17 .88†† -.23 †††
.67
.70†††
-.13
.16
.47
-.37
-.78
.42 .67†††
.32
.33
.66
.48
-.31
-.58
.11
.08
-.87
-1.99
410.7††† 23.7††† 3,063
-.22
.17
††
†††
.06††
-1.32
280.3††† 10.0†† 3,172
144
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance
Baltic States 100% 5.6
11.1
18.4
80%
18.9
Percentage
13.6 60%
12.4
Manager Professional Semiprofessional
10.0
Sales, service, clerk
12.0
8.8
6.4
40%
Agricultural worker 21.7
18.4
Worker
20% 22.0
20.7
No job
0%
Job in 1991
Job in 1997-1998
Belarus, Kharkiv, and Kurgan 100%
3.7
6.4
18.4
19.9
80%
Percentage
15.0 60%
13.0
10.6
10.2
4.3
4.0
28.3
21.5
40%
Manager Professional Semiprofessional Sales, service, clerk Agricultural worker Worker
20% 19.8
25.1
No job
0%
Job in 1991
Job in 1997-1998
Figure 6.2. Distribution of Occupational Status in 1991 and 1997-1998.
© Heili Pals September 2006
145
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance Before analyzing the results for occupational strata in these two groups of regions, a closer look at the distribution of occupational strata is needed. Figure 6.2 shows the comparative distribution of occupational strata at a job in 1991 and at the job in 1997-1998 for both types of regions. This comparison shows the extent of the shift in the overall structure of the occupations over this time period for the generation considered here. Surprisingly, the differences between the regions are relatively similar. In both regions, about one-fifth of the sample did not have a job at both time points. The proportion of jobless people was somewhat higher in Belarus, Kharkiv, and Kurgan in 1997-1998 than in the Baltic states. The proportion of workers decreased in both types of region. The other major change occurred at the top of the hierarchy. In both types of regions, the proportion of managers increased considerably. In 1997-1998, there were twice as many managers as in 1991. This is probably because people are more likely to be chosen as managers as they get older.
© Heili Pals September 2006
146
Table 6.7 ML Estimates of Multinomial Logistic Regression Model of Occupational Status in the Baltic States, Relative to Workers (Source: PG, Waves 1 & 4; Valid N = 3,094). No Job Agricultural Clerk, Sales, SemiProfessional Manager Worker Service professional Adolescent noncompliance
.38
-.56
.06
1.67†††
.26
3.01†††
-.30
.28
.62
1.49†††
1.61†††
-.13
†††
†††
Control Variables Female
†††
††
†
Education (years)
.16
.13
.10
.31
.86
.60††
Grade point average
.06
-.25
.09
.04
.26
.35†
Parent's education (years)
.05†
-.04
.01
-.03
.11†††
.09††
-.08
.16
.22
-.01
.29
.16
.30
.22
Place of residence Rural City
.18
.73††† -.50
†
Self-esteem
.64
-.10
1.03
1.08
1.75
1.50††
Locus of control
.02
1.09
-.37
.98
-.85
-.60
Participation in second economy Intercept L.R. χ2 vs. null model (d.f.= 60) L.R. χ2 vs. model with control variables (d.f.= 6) †
.20
†
†
††
†††
.07
.39
.18
.12
-.08
.15
-.57
-1.34
-2.14
-.89
-1.35
-.51
1960.0††† 9.0
147
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Table 6.8 ML Estimates of Multinomial Logistic Regression Model of Occupational Status in Belarus, Kharkiv, and Kurgan, Relative to Workers (Source: PG, Waves 1 & 4; Valid N = 3,206). No Job Adolescent noncompliance
Agricultural Clerk, Sales, SemiProfessional Worker Service professional
Manager
.48
.51
.87†
.12
.29
.78†
1.76†††
.49†
2.94†††
1.91†††
1.81†††
.75†††
.16†††
.43†††
.89†††
.71†††
Control Variables Female Education (years) Grade point average Parent's education (years)
.27††† -.05 .08†††
-.05 .26
.24
-.06
.22
.37†
-.04
.06†
.04
.06†
.07†
Place of residence Rural
-.27†
1.33†††
City
-.10
Self-esteem Locus of control Participation in second economy Intercept L.R. χ2 vs. null model (d.f.= 60) L.R. χ2 vs. model with control variables (d.f.= 6) †
-.06
-.02
-.10
.21
-.59
.03
-.23
-.23
.14
-.04
.42
-.15
.58
-.11
.91†
-.16
-.13
.93
1.00†
-.61
†††
†††
.66
1.03
.35
.41
.27
-.22
-2.80
-2.18
-1.02
-1.12
-.36 .63† -1.65
1908.6††† 7.9
148
p < .05, †† p < .01, ††† p < .001 (1-tailed test) Note: Effects relative to general secondary school and town. Region is effect-coded with the Kharkiv region of Ukraine as the omitted category. Education, GPA, parents’ education, self-esteem, locus of control, and adolescent noncompliance are centered at the sample mean of the variable. Parents’ education is measured as years of education of the most-educated parent. GPA ranges from 3.5 or less to 5.
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance Table 6.7 presents the results of the multinomial logistic regression model of occupational strata in 1997-1998 in the Baltic states. Noncompliance does not have any significant effects on occupational status in 1997-1998, although its effect does approach significance for the likelihood of being a manager (p<.10, one-tailed test). Adolescent noncompliance does increase the chances of being a clerk, working in sales or service, or being a manager in the other three slowly-changing regions (Table 6.8). One possible explanation is suggested by Figure 6.2. The Belarus, Kharkiv, and Kurgan regions have experienced a slightly bigger shift in the occupational structure. There are more people not working in these regions than in the Baltic states. In addition, the proportion of managers is rather small in these three regions. This indicates that this is a relatively exclusive stratum to be in for this cohort. In conclusion, the results of the regional comparison give some support to the hypothesis that the positive effects of adolescent noncompliance are only present under rapidly changing societal conditions. The positive effects are more consistently present in the Baltic regions that experienced rapid changes in regulations and norms. The positive effects of noncompliance are either not present at all or rarely present in Ukraine, Belarus, and the Kurgan region of Russia. Changes in these regions have been slower; the level of regulation is comparatively much higher in these regions than in the Baltic states. CONCLUSION: STRUCTURAL DIFFERENCES The results in this chapter have supported my argument that the effect of adolescent noncompliance depends on social context. I conducted three different types
© Heili Pals September 2006
149
Chapter 6: Structural Differences in Effects of Adolescent Noncompliance of tests by comparing the effects of adolescent noncompliance in three ways: 1) over time on occupational status before and after the collapse of the Soviet Union, 2) in the state and private sectors, and 3) across two region groups of the former Soviet area. All three tests yielded a similar conclusion: the positive effects of adolescent noncompliance on labor market outcomes are only present in unpredictable societal conditions. The positive effect of adolescent noncompliance on occupational status is present after the collapse of the Soviet regime and not before. Similarly, the positive effect of adolescent noncompliance can be observed in the more unpredictable private sector and not in the relatively stable state sector. Finally, the argument that the effects of adolescent noncompliance depend on social context also holds across regions: the results indicate that adolescent noncompliance has more positive effects on labor market outcomes in the Baltic states and smaller effects in relatively slowly changing regions such as Belarus and the Kharkiv and Kurgan regions.
© Heili Pals September 2006
150
CHAPTER 7: CONCLUSION AND DISCUSSIONS In this concluding chapter, I review the theoretical contributions of this research and give a summary of empirical results. In addition, I discuss the limitations and strengths of the current study and provide directions for further research. CONCLUSION AND DISCUSSION The goal of this research was to gain a better understanding of the effects of adolescent noncompliance to school authorities’ rules on labor market outcomes in different social structures. This study has investigated adolescent noncompliance using regions of the former Soviet Union as examples of societies in which structural conditions have allowed adolescent noncompliance to have positive effects on labor market outcomes. This research draws on classical sociological theories, such as Merton’s (1968) strain theory, social psychological work on nonconformity, and more specific work in adolescence research, deviance research, and studies of transitional societies. Most of the research to date takes a rather negative view of adolescent noncompliance and related concepts, claiming that such behaviors have generally adverse effects on life outcomes (Sampson & Laub 1990; Alexander et al. 1997; Tanner et al. 1999; Hamil-Luker et al. 2004; Gottfredson & Hirshi 1990; Chen & Kaplan 2003; Akers 1991; Cleckley 1976; Wilson & Hernstein 1985). Some other strands of research recognize that noncompliance could also have positive impacts (Becker 1963; Phillips & Zuckerman 2001). One of the more frequent findings is that nonconformity, a related concept to noncompliance, might be connected to greater leadership skills and greater self-confidence (Lightfoot 1997; Crutchfield 1955; Crowne & Liverant 1963).
© Heili Pals September 2006
151
Chapter 7: Conclusion and Discussions This research goes a step further than claiming either negative or positive effects of adolescent noncompliance on labor market outcomes. I have developed a framework for understanding the interplay between adolescent noncompliance and unpredictability in societies. I argue that societal changes and the unpredictability of the social context largely determine the sign of the effects of noncompliance. Three characteristics describing the level of unpredictability are discussed: 1) the level of legal clarity, 2) the level of normative clarity, and 3) the supply of open niches. A frequently changing society is delineated by disarray in legal boundaries, normative confusion, and an abundance of open niches. I argue and find that if these three conditions are satisfied, those who were noncompliant as adolescents are better able to manage confusion in the legal and normative worlds than those who were compliant as adolescents. Thus, the effects of adolescent noncompliance on labor market outcomes under these conditions are positive. These structural conditions are likely to occur during a time of societal transformation. In particular, I chose regions of the former Soviet Union at the time of their transition from a command economy as examples of societies where the extent of societal changes is relatively high. Several mechanisms might be behind the positive effects of adolescent noncompliance on labor market outcomes in unpredictable societies. Those who have experience with not complying with the rules of school authorities have essentially already gained experience with going against the rules of an institutionalized structure. With the transition from a command economy, the value of opposition to the rules of authorities changed. In contrast to Soviet times, opposition to the rules of the
© Heili Pals September 2006
152
Chapter 7: Conclusion and Discussions institutionalized structure became valued; individuality became more valued. Those who did not comply with the rules of the institutionalized structure of Soviet secondary schools were more likely to exhibit both traits. In addition, those who did not comply with the rules of school authorities were more comfortable during the major societal changes, when it was unclear what was allowed and what was not. Those who used to comply with school rules, on the other hand, were less comfortable in an unpredictable society. Trying to follow authorities’ rules point-by-point during rapid changes was difficult, as the rules were changing constantly. The argument developed here is important in several aspects. The first and most obvious implication is the need to consider the variation of effects depending on social context. I found positive effects of adolescent noncompliance in rapidlychanging social structures, but a lack of the effect in relatively stable structures. Thus, one needs to carefully consider the social context when analyzing life course outcomes. Secondly, this study of adolescent noncompliance sheds light on the importance of differentiating between different kinds of noncompliance. It is important to distinguish noncompliance to the rules of school authorities from more serious rule-breaking. Criminal deviance is not likely to have the same effects on labor market outcomes as found here for adolescent noncompliance, even in rapidlychanging societies. Failing to differentiate between criminal deviance and noncompliance to school rules might yield incorrect conclusions, as the effects of criminal deviance might be opposite the effects of adolescent noncompliance in
© Heili Pals September 2006
153
Chapter 7: Conclusion and Discussions rapidly-changing societies. I have attempted to give an overview of the different definitions and terminology used with regard to noncompliance and related concepts in the previous literature. The previous literature has employed a large variety of conceptualizations, often using different terms for substantively similar concepts. Therefore, I have taken a careful look at the existing terminology and given a detailed conceptualization of adolescent noncompliance. Finally, the argument developed here and the results of this research confirm once again that the mechanisms that work in stable societies do not necessarily apply in transitional societies. Sometimes the social changes themselves affect the nature of the relationships between different phenomena, and distinctive explanations are needed to describe the mechanisms that apply in rapidly-changing societies. SUMMARY OF RESULTS With the example of the former Soviet Union regions, I have demonstrated that, indeed, in structures that are rapidly changing and therefore have less established normative and legal structures, adolescent noncompliance increased the chances of being engaged in entrepreneurial activities, working in the private sector, and becoming a manager. A summary of the results of the hypothesis tests is presented in Table 7.1.
© Heili Pals September 2006
154
Chapter 7: Conclusion and Discussions Table 7.1 Summary of Hypotheses. Hypotheses
#
Results
Tables
… participating in the second economy
H1a
Supported
5.1
… engaging in entrepreneurial activities
H1b
Supported
5.2-5.4
… working in the private sector
H1c
Supported
5.5-5.6
Ch5: New types of employment In a rapidly changing society, adolescent noncompliance increases the likelihood of …
Ch 5: Occupational status In a rapidly changing society, adolescent noncompliance … … increases the status of the job in adulthood
H1d Partly supported 5.7-5.8
Ch 5: Status & adolescent noncompliance In a rapidly changing society, adolescent noncompliance is more advantageous for … … men than for women
H2a
Not supported
… those with the lowest or highest GPA
H2b
Not supported
… those in vocational or academic secondary schools
H2c
Not supported
… those with parents with the lowest or highest SES
H2d Partly supported
Ch 6: Structural explanation In a rapidly changing society, adolescent noncompliance … … decreases the status of the first job if it is attained before rapid social changes
H3a Partly supported
6.1
… is more advantageous in rapidly changing regions
H3b
Supported
6.3-6.8
… is more advantageous in the private sector than in the state sector
H3c
Supported
6.2
© Heili Pals September 2006
155
Chapter 7: Conclusion and Discussions In Chapter 5, I tested whether adolescent noncompliance increased the likelihood of participating in new types of employment opportunities. I used a variety of different empirical outcomes to measure new types of employment. I found that adolescent noncompliance increases the likelihood of participating in the second economy before the collapse of the Soviet Union. The second economy was risky to engage in for a number of reasons. It was illegal. It was a new type of activity, requiring a considerable amount of individual initiative and acceptance of capitalist thinking. Participation in the second economy was not in line with the rules at the time: those engaging in the second economy clearly went against the directives of Soviet institutionalized authorities. This explains why those who had the experience with not complying with school authorities’ rules were more likely to engage in the second economy in the Soviet era, where those who were compliant as youths were held back by their habit of following school authorities’ rules. Furthermore, I demonstrated that adolescent noncompliance is also an advantage when engaging in private business after the collapse of the Soviet Union. Those who did not comply with the rules of school authorities were more likely to own a firm, be engaged in entrepreneurial activities, and be self-employed. All three are indicators of new career paths involving entrepreneurship. I used economic sector as the dependent variable to more directly compare old and new career paths. Those who did not comply with the rules of school authorities were more likely to work as employees in the private sector (a new career path) than as employees in the state sector (old career path). Thus, those who were noncompliant as adolescents were more likely to invest in new career paths and therefore seemed to
© Heili Pals September 2006
156
Chapter 7: Conclusion and Discussions be less affected by the collapse of the old labor market structure. Those who were compliant as youths, who were uncomfortable not following the rules of authorities, were still investing in old career paths when the society was already changing. Thus, this gave two advantages for those who were noncompliant in adolescence: 1) the first mover advantage, and 2) the avoidance of the collapsing system. Those who were noncompliant as adolescents were more likely to engage in early entrepreneurial activity even before the collapse of the Soviet Union. Such engagement gave them a better opportunity to take advantage of changes that followed: to engage in entrepreneurial activities once they were legally allowed. At the same time, by exploiting alternative career routes, those who were noncompliant in adolescence became less dependent on the old career paths that were shrinking after the collapse of the Soviet regime. One example of an alternative career route in Estonia was the relatively common practice of selling cucumbers at the markets of what is now St. Petersburg. Essentially, these early entrepreneurs realized that the big city was in short supply of quality vegetables. They either grew the cucumbers themselves or bought them from local farmers and made money by making the trip to Russia to sell them. It was an illegal activity at the time as selling for personal profit was illegal. Thus, people who resorted to this kind of activity had to be accustomed to going against the established rules of authorities. However, through this early entrepreneurship, they gained not only knowledge and experience, but also some capital to engage in other, more profitable entrepreneurial activities. Those who were compliant as adolescents, however, were less likely to participate in early entrepreneurial activity. Rather, they
© Heili Pals September 2006
157
Chapter 7: Conclusion and Discussions invested in old, well-established labor market routes that were negatively affected by the collapse of the Soviet regime. Thus, they got a later start in working in the new types of employment opportunities, and their careers were more likely to be affected by the collapse of the Soviet system. This pattern is also supported by additional analysis of the nonlinearity of the effects of adolescent noncompliance. The nonlinearity tests clearly showed the disadvantages of adolescent compliance with authorities’ rules during societal changes. Those who were compliant as adolescents, compared to youths with an average level of compliance, were less likely to engage in the second economy, own a firm, and engage in private business in other ways. They were also less likely to become professionals or managers during the rapid societal changes. Thus, the overall positive effects of noncompliance are largely due to the disadvantages of compliance when authorities’ rules are frequently changing. Titma and Trapido (2002) defined the “winners” of the post-Soviet transition in Estonia and Latvia as those in managerial positions or those engaged in entrepreneurial activities. The analysis here has shown that noncompliance to the rules of school authorities during the Soviet regime increases the chances of engaging in entrepreneurial activities and being managers. Thus, those who were noncompliant as adolescents managed the uncertain and unpredictable transition from a command economy after the collapse of the Soviet Union more adeptly. The argument developed here contends that the effects of adolescent noncompliance on labor market outcomes vary by the amount of fluctuations in the legal, normative, and institutional framework of society. I conducted several tests to
© Heili Pals September 2006
158
Chapter 7: Conclusion and Discussions show this. First, I compared the effects of adolescent noncompliance on occupational status before and after the collapse of the Soviet Union. Before the major societal fluctuations (i.e., before 1992), noncompliance did not have an impact on occupational status. However, during the time of societal changes after the collapse of the Soviet Union, adolescent noncompliance did increase the chances of becoming a manager. A second test of the argument that the effects of noncompliance depend on social context distinguished between occupations in the state and private sectors. Again, adolescent noncompliance did not have an effect on the chances of being a manager or a professional in the state sector. However, it did increase the chances of being a manager or a professional in the private sector. Finally, I compared the effects of adolescent noncompliance in different types of regions. Adolescent noncompliance had more consistently positive effects on labor market outcomes in the rapidlychanging Baltic states than in the comparatively slower-changing Belarus, Kurgan, and Kharkiv regions. All three tests supported the general argument of this research: that the effects of noncompliance on labor market outcomes are different in structures with different levels of unpredictability. STRENGTHS OF THE STUDY This study has several strengths that deserve to be noted. The first and foremost strength of this study is the longitudinal nature of the data, Paths of a Generation. Adolescent noncompliance is measured in this study about fifteen years before many of the outcome measures. It is impressive that the positive effects of adolescent noncompliance can be established across such a time gap. Furthermore,
© Heili Pals September 2006
159
Chapter 7: Conclusion and Discussions such a time difference between the measurement of adolescent noncompliance and labor market outcomes ensures that the causality of the effect can work in only one direction. The possibility that the level of noncompliance is affected by labor market outcomes is eliminated due to the time difference between the measurements of adolescent noncompliance and labor market outcomes. Many studies have analyzed the effects of concepts related to noncompliance on life outcomes; however, they have done so with data for a single point in time, so they cannot establish the direction of the effect. Measuring noncompliance in adolescence solves this problem. My study is further strengthened by the comparisons across different economic sectors and different regions of the former Soviet Union, as well as across time. These comparisons permit me to test my hypothesis about the variation of the effects of noncompliance depending on the social context. I am able to show that the effects of noncompliance are different in contexts with different levels of unpredictability. Third, the results here are more convincing due to the wide range of outcome measures that I am able to use. I measure entrepreneurial activities using a variety of indicators. Besides entrepreneurial activities, I distinguish between working in the state and private sectors, and use occupational status as a dependent variable. Finding support for the general argument that adolescent noncompliance has positive effects in societies with a high level of unpredictability using a variety of outcomes gives added support to the hypothesis tests and increases confidence in the results. Another strength of this study is the way that adolescent noncompliance is measured. Instead of using one narrow indicator of noncompliance, I have measured adolescent noncompliance using three different indicators: skipping school, conflicts
© Heili Pals September 2006
160
Chapter 7: Conclusion and Discussions with teachers, and disobeying school rules. I combine these three into an index of noncompliance, thus creating a more general notion of noncompliance than any one indicator could measure. FURTHER RESEARCH AND APPLICATIONS The data used and the length of the study has placed some restrictions that further analyses should address. One major restriction is that I am able to analyze only one generation, the generation in secondary school in 1983-1984, which entered the labor force sometimes around the collapse of the Soviet regime and was typically around 32 years old in 1998. This generation might have been a special generation due to their age when the Soviet Union collapsed. These people experienced two different major changes at the same time: the individual transition from adolescence to adulthood and the societal transition from the command economy. At the same time, this is also a strength of this study; very few studies have been able to assess the life course for such a generation. Despite the special position of this generation in the midst of change, the results presented here are probably not particular to a single generation, but are generalizable across generations in these types of structural conditions. Future studies might investigate if the results found here are more likely for younger people than for older people whose careers are already set. Secondly, the Paths of a Generation sample includes only respondents who were approaching the end of their secondary education at the time of the first interview in 1983-1984. Youths who had dropped out during the early stages of
© Heili Pals September 2006
161
Chapter 7: Conclusion and Discussions secondary school or even sooner are not included in the sample. This is a rather small section of the population (approximately 10-15 percent); however, it is also probably an unusual group. However, enough variation remains in the amount of adolescent noncompliance among those who stayed in school to analyze the relationship between adolescent noncompliance and labor market outcomes. A third restriction of this study is the limited ability to compare the effects of adolescent noncompliance over time. The only over-time comparison I could make was a comparison of the effects of noncompliance on occupational status of the first job before 1992 versus occupational status in 1997-1998. Ideally, a similar over-time comparison should be done with other labor market outcomes, but the available data do not permit this. A final restriction of this study is that I can only assess the effects of adolescent noncompliance on labor market outcomes in one type of society and one type of transition. I use the case of the former Soviet Union regions during and after the collapse of the Soviet regime. However, similar effects of adolescent noncompliance might be found in other structures with high unpredictability. I expect that the positive effects of noncompliance are also present in new firms, new industries, and other transitional societies. A classic example could be a comparison of a start-up company to a wellestablished large corporation. A start-up company is likely to lack a well-established authority structure. The hierarchy and the rules attached to the hierarchy within the company are not fully established. In contrast, in a large, older corporation, the hierarchy structure is well-established: everybody knows how information should be
© Heili Pals September 2006
162
Chapter 7: Conclusion and Discussions moving along the lines of hierarchy, and each position is well-defined. In comparison, positions are often not clearly defined in a start-up: each employee has to perform a variety of tasks. Authority structure and norms within the company are usually much more definitely set in an older company than in a start-up. Thus, by comparing an average start-up company and a well-established corporation, one should be able to observe that the start-up company has a high level of unpredictability, while the well-established company is highly predictable in terms of institutionalized structure. Based on the hypotheses presented here, we would expect to see positive effects of noncompliance on labor market outcomes in the startup environment, but should not see them in the well-established corporation. Similar effects might be apparent in other settings as well. A new industry might be a setting with a high level of unpredictability because the opportunity structure is shifting; the norms and regulations and the authority structure within the new industry or economic area are not set. Alternatively, similar conditions could potentially be found in other societies in transition that have experienced major legal and economic change. Further research is needed to confirm a more general application of the theoretical ideas presented here; however, there is no reason to believe that this is something particular to the former Soviet regions. Rather, the findings here are likely to be applicable in any societal structure with high a level of unpredictability. In the former Soviet regions, it remains to be seen whether the advantage of adolescent noncompliance will continue as these societies move toward a new, more stable economic and social system. The same mechanisms may cease to
© Heili Pals September 2006
163
Chapter 7: Conclusion and Discussions work in a society where social changes have already occurred and new mechanisms have become institutionalized.
© Heili Pals September 2006
164
APPENDIX 1. ENGLISH TRANSLATION OF THE QUESTIONS. Variables
Answer Choices
Timing
never; rarely; sometimes; often
Wave 1
Adolescent Noncompliance Have you ever skipped school?
I fulfill school rules and regulations disagree strongly; disagree; Wave 1 without objection agree; agree strongly I have had conflicts with teachers and yes; no the school administration
Wave 1
I have often had conflicts with my no; not really; yes, parents sometimes; yes
Wave 1
Adult Outcomes Participation in the second economy yes; no In 1991 or before, were you involved in regularly making or growing things to sell; regularly selling things bought for trading; working in a private business that belongs fully or partly to you, or in providing services for private persons?
1991 & before (measured in Wave 4)
Firm ownership Do you own a business or firm?
Wave 4
yes; no
Ever self-employed since 1992 yes; no Work history ever reporting being self-employed since 1992
1992 & later (measured in Wave 4)
Engagement in entrepreneurial employees; economically activities inactive; working Combination of main and secondary entrepreneur; nonworking economic activity entrepreneur
Wave 4
Work in private or state sector private sector; state sector State sector (state or municipal firm; organization financed by the state budget; kolkhoz); private sector (stock company; private firm; private person)
Wave 4
© Heili Pals September 2006
165
Appendix I: English Translation of the Questions APPENDIX 1. CONTINUED… Variables
Answer Choices
Occupational strata of current job
manager; professional; semi-professional; clerk, Occupational strata of first full-time job service, or skilled worker; before 1992 operative or unskilled worker; agricultural worker; no job
Timing Wave 4 Individual timing (measured in Wave 4)
Control Variables Gender
female; male
Wave 4
Region
Belarus; Estonia; Latvia; Kharkiv; Kurgan
Wave 1
Place of residence
city; town; village
Wave 4
Respondent’s years of education
years
Wave 4
Type of secondary education
vocational; specialized secondary; general secondary; academic
Wave 1
Parents’ years of education
years
Wave 4
Grade point average
range 3.5-5
Wave 1
Self-esteem Compared to your peers, evaluate your abilities in: 1) humanities, 2) sciences, 3) music, 4) arts, 5) organizing work, 6) intellectual matters, 7) motivation and willpower, 8) time management, 9) dealing with other people, 10) technical knowledge, and 11) manual skills.
sum of 4-point Likert scales rescaled to 0-1 (0 indicating low self-esteem, 1 indicating high selfesteem)
Wave 1
© Heili Pals September 2006
166
Appendix I: English Translation of the Questions APPENDIX 1. CONTINUED… Variables
Answer Choices
Timing
Locus of control You chose one of three types of secondary school: general, specialized, or vocational secondary school. Whose advice or what circumstances influenced your choice?: 1) teachers, 2) the school principal, 3) school counselor, 4) mother, 5) father, 6) classmates, 7) examples of acquaintances, 8) interest in a particular subject, and 9) interest in a hobby.
sum rescaled to 0-1 (0 indicating external locus of control, 1 indicating internal locus of control)
Wave 1
© Heili Pals September 2006
167
APPENDIX II: NOTE ABOUT MULTINOMIAL LOGISTIC REGRESSION Multinomial logistic regression is the most widely used method in the social sciences to analyze effects on a categorical polytomous variable. However, a multinomial logistic regression model assumes independence between the log odds of each category of the outcome. This assumption is often not met in empirical social science data. Hausman and McFadden (1984) were the first to propose a test of the assumption of independence of irrelevant alternatives (IIA). Hausman and McFadden’s test compares the parameters of a model estimated using a full set of outcome alternatives and a model with one alternative left out. If IIA holds, the parameters between the full model and the restricted model should be approximately the same. An alternative test developed by Small and Hsiao (1985) divides the sample into two sub-samples of approximately equal size. It then compares the log-likelihood for a full set of alternatives averaged across two sub-samples with the log-likelihood of the restricted model computed using the second sub-sample with one of the alternatives eliminated. I have used both Hausman-McFadden and Small-Hsiao tests of the IIA assumption in all cases when I analyze a polytomous dependent variable. Table A1 summarizes the results of these tests. Table A1 illustrates a frequent problem with the Hausman-McFadden and Small-Hsiao tests: they yield contradictory conclusions regarding the null hypothesis. For most of the polytomous outcomes the Hausman-McFadden tests hold and do not indicate a serious violation of IIA. However, the Small-Hsiao tests often suggest that
© Heili Pals September 2006
168
Appendix II: Note about Multinomial Logistic Regression there is a considerable violation of IIA for most polytomous outcomes analyzed here. Noting the frequent contradictory results from both tests and citing Monte Carlo testing (Cheng & Long 2005), Long and Freese (2006: 243) discourage the use of either test of IIA, doubting the results of both. They suggest the use of multinomial logistic regression models only in the theoretically obvious cases where IIA should hold. Table A1. Hausman-McFadden and Small-Hsiao Tests for the Independence of Irrelevant Alternatives (Source: Paths of a Generation, Waves 1 and 4). IIA Tests HausmanSmall-Hsiao McFadden
Dependent Variable Economic Activity in 1997-1998
Partly for H0
Against H0
Economic Sector in 1997-1998
Partly for H0
Against H0
For H0
Against H0
Partly for H0
Against H0
For H0
Against H0
Occupational Status of the First Job before 1992 Occupational Status in 1997-1998 Occupational Status by Economic Sector in 1997-1998 Note: H0: outcomes are independent of other alternatives.
The most frequently suggested and simplest solution is to combine some categories of the dependent variable. I collapsed all polytomous outcomes to three categories, keeping the reference category the same and collapsing the remaining categories besides the main category affected by noncompliance into one. For example, for economic sector, I created three categories: 1) employee in the private sector, 2) the rest of the categories collapsed (entrepreneur in the private sector, stock company, and not working), and 3) state sector as the reference category. Collapsing the categories of the labor market outcomes used here does not change the conclusions
© Heili Pals September 2006
169
Appendix II: Note about Multinomial Logistic Regression regarding the effects of adolescent noncompliance. Both IIA tests become more favorable for the null hypothesis; however, the tests still show some violation of the IIA assumption. An exception is economic sector, with both IIA tests showing no violation after collapsing the categories. As a test of robustness, I replicated the results of the multinomial logistic regression models presented here with separate binary logistic regression models20. For example, instead of one multinomial logistic regression model for the type of economic sector, I estimated five different binary logistic regression models where the dependent variable was a binary variable comparing each category against a combined group of all the remaining categories. The results were largely congruent with the results from the multinomial logistic regression models. However, this test does not directly solve the problem of IIA because the alternatives of the dependent variable are not considered simultaneously. At most, the largely congruent results from the binary logistic regression models increase confidence in the robustness of the results of the multinomial logistic regression models presented here. To fully overcome the problem of the lack of IIA, alternative methods of analysis should be used. Multinomial probit is one method that overcomes the problem of lack of IIA. This model has only recently been implemented in some statistical programs because estimating it is computationally very intensive. Consequently, it has been recommended that the outcome should have only few categories so that the multinomial probit model can converge. Indeed, my attempts to estimate the full models presented here using a multinomial probit model failed. 20
The tables are not presented here but are available from author upon request.
© Heili Pals September 2006
170
Appendix II: Note about Multinomial Logistic Regression However, I did estimate multinomial probit models using the collapsed versions of the main labor market outcomes using the MNP package in R. MNP fits the Bayesian multinomial probit model via Markov chain Monte Carlo (Imai & van Dyk 2005). In all multinomial probit models, I used three-category dependent variables: the reference category, the main category of interest, and the rest of the original categories collapsed to one. I estimated the following five models comparing: 1) Employees in the private sector with those in the state sector, 2) Entrepreneurs in the private sector with those in the state sector, 3) Working entrepreneurs with employees, 4) Managers with workers in 1997-1998, 5) Managers with workers at the first job before 1992. Two-sided tests at p<.05 reveal that adolescent noncompliance is significantly positive for employees and entrepreneurs in the private sector and for working entrepreneurs. The coefficient for managers, both in 1997-1998 and at the first job, fails to achieve the conservative two-sided test at p<.05. However, the coefficient for managers in 1997-1998 does approach significance. The coefficients for the control variables closely replicate those estimated by multinomial logistic regressions. Thus, multinomial probit analysis of the collapsed outcome variables does not contradict the results achieved here using multinomial logistic regression. A reanalysis using multinomial probit with detailed labor market outcomes as dependent variables should be conducted once software becomes available that is able to replicate the analysis with all the outcome categories.
© Heili Pals September 2006
171
REFERENCES Akers, Ronald L. 1991. “Self-Control as a General Theory of Crime.” Journal of Quantitative Criminology 7(2): 201-211. Alexander, Karl L., Doris R. Entwisle, and Carrie S. Horsey. 1997. “From First Grade Forward: Early Foundations of High School Dropout.” Sociology of Education 70: 87-107. Arum, Richard. 1997. “Trends in Male and Female Self-Employment: Growth in a New Middle Class or Increasing Modernization of the Labor Force?” Research in Stratification and Mobility 15: 209–78. Asch, Solomon E. 1956. “Studies of Independence and Conformity: A Minority of One against a Unanimous Majority.” Psychological Monographs 70 (9, Whole No. 416). ______. 1952. Social Psychology. Englewood Cliffs, NJ: Prentice-Hall. Aslund, Anders. 1994. “The Case for Radical Reform.” Journal of Democracy 5(4): 63-74. ______. 1993. “Comment on 'Gradual versus Rapid Liberalisation in Socialist Economies' by McKinnon.” Proceedings of the World Bank Annual Conference on Development Economies. Washington, DC: World Bank. ______. 1992. Post-Communist Economic Revolutions. How Big a Bang? Washington, DC: Center for Strategic and International Studies. Babakus, Emin, Carl E. Ferguson Jr., and Karl G. Jöreskog. 1987. “The Sensitivity of Confirmatory Maximum Likelihood Factor Analysis to Violations of Measurement Scale and Distributional Assumptions.” Journal of Marketing Research 24: 222-228. Bandura, Albert. 1995. “Exercise of Personal and Collective Efficacy in Changing Societies.” Pp. 1-45 in Self-Efficacy in Changing Societies, edited by A. Bandura. New York: Cambridge University Press. _______. 1992. “Exercise of Personal Agency through the Self-Efficacy Mechanism.” Pp. 3-38 in Self-Efficacy: Thought Control of Action, edited by R. Schwarzer. Washington, DC: Hemisphere. Becker, Howard S. 1963. Outsiders: Studies in the Sociology of Deviance. New York: The Free Press.
© Heili Pals September 2006
172
References Blau, Peter. 1963. Dynamics of Bureaucracy: A Study of Interpersonal Relations in Two Government Agencies. Chicago: University of Chicago Press. _______. 1960. “Patterns of Deviation in Work Groups.” Sociometry 23: 245-261. Bollen, Kenneth A. and Kenney H. Barb. 1981. “Pearson’s R and Coarsely Categorized Measures.” American Sociological Review 46: 232-239. Böröcz, József and Ákos Róna-Tas. 1995. “Small Leap Forward: Emergence of New Economic Elites.” Theory and Society 24: 751-781. Bowles, Samuel and Thomas E. Weisskopf. 1998. Economics and Social Justice: Essays on Power, Labor, and Institutional Change. Northampton, MA: Edward Elgar Publications. Bradley, Robert H. and Robert F. Corwyn. 2002. “Socioeconomic Status and Child Development.” Annual Review of Sociology 53: 371-399. Chen, Zeng-yin and Howard B. Kaplan. 2003. “School Failure in Early Adolescence and Status Attainment in Middle Adulthood: A Longitudinal Study.” Sociology of Education 76: 110-127. Cheng, S. and J. Scott Long. 2005. “Testing for IIA in the Multinomial Logit Model.” University of Connecticut, Working Paper. Chernysh, Mikhail F. 1995. “Changing Values: Russian Youth in Transition.” Pp. 161-176 in Values and Post-Soviet Youth: The Problem of Transition, edited by Luigi Tomasi. Milan, Italy: FrancoAngeli. Chubais, Anatoly B. and Maria Vishnevskaya. 1993. “Main Issues of Privatization in Russia.” Pg. 89-98 in Changing the Economic System in Russia, edited by Anders Aslund and Richard Layard. New York: St. Martin's Press. Clarke, Simon. 1999. The Formation of a Labour Market in Russia. Northampton, MA: Edward Elgar Publications. Clausen, John A. 1991a. “Adolescent Competence and the Life Course, or Why One Social Psychologist Needed a Concept of Personality.” Social Psychology Quarterly 54: 4-14. _______. 1991b. "Adolescent Competence and the Shaping of the Life Course." American Journal of Sociology 96: 805-842. Cleckley, Hervey Milton. 1976. The Mask of Sanity: An Attempt to Clarify Some Issues about the So-Called Psychopathic Personality (5th edition). St. Louis: Mosby.
© Heili Pals September 2006
173
References Clement, Cynthia and Peter Murrell. 2001. “Assessing the Value of Law in Transition Economies: An Introduction.” Pp. 1-20 in Assessing the Value of Law in Transition Economies, edited by Peter Murrell. Ann Arbor, MI: The University of Michigan Press. Cohen, Albert K. 1966. Deviance and Control. Englewood Cliffs, N.J.: Prentice-Hall. Crowne, D.P. and S. Liverant. 1963. “Conformity under Varying Conditions of Personal Commitment.” Journal of Abnormal Social Psychology 66: 547-555. Crutchfield, R. S. 1955. “Conformity and Character.” American Psychologist 10: 191198. DiStefano, Christine. 2002. “The Impact of Categorization With Confirmatory Factor Analysis.” Structural Equation Modeling: A Multidisciplinary Journal 9(3): 327-346. Dittes, James E. and Harold H. Kelley. 1956. “Effects of Different Conditions of Acceptance upon Conformity to Group Norms.” Journal of Abnormal and Social Psychology 53: 100-107. Dolan, Conor V. 1994. “Factor Analysis of Variables with 2, 3, 5 and 7 Response Categories: A Comparison of Categorical Variable Estimators Using Simulated Data.” The British Journal of Mathematical & Statistical Psychology 47: 309326. Du Toit, Mathilda (ed.). 2003. IRT from SSI: Bilog-MG, Multilog, Parscale, Testfact.. SSI Scientific Software International. Durkheim, Emile. 1915. The Elementary Forms of the Religious Life: A Study in Religious Sociology. Translated by Joseph Ward Swain. New York: Macmillan. Elder, Glen H., Jr. 1987. "War Mobilization and the Life Course: A Cohort of World War II Veterans." Sociological Forum 2(3): 449-472. ______. 1986. "Military Times and Turning Points in Men's Lives." Developmental Psychology 22(2): 233-245. ______. 1980. “Social Structure and Personality: A Life-Course Perspective.” Pp. 108-190 in Quantification and Psychology: Toward a New History, edited by Harvey J. Graff and Paul Monaco. Washington, D.C.: University Press of America. ______. 1974. Children of the Great Depression: Social Change in Life Experience. Chicago: University of Chicago Press.
© Heili Pals September 2006
174
References Elder, Glen H., Jr. and Robert Crosnoe. 2002. “The Influence of Early Behavior Patterns on Later Life.” Pp. 157-176 in Paths to Successful Development: Personality in the Life Course, edited by Lea Pulkkinen and Avshalom Caspi. New York: Cambridge University Press. Eyal, Gil, Iván Szelényi, and Eleanor Townsley. 1998. Making Capitalism Without Capitalists: The New Ruling Elites in Eastern Europe. New York: Verso. Featherman, David L. and Annemette Sorensen. 1983. “Societal Transformation in Norway and Change in the Life Course Transition into Adulthood.” Acta Sociologica 26(2): 105-126. Futo, Peter, Paul Hoggett, and Laszlo Kallay. 1997. “Small Firms and Economic Transformation in Hungary.” Innovation 10(2): 171-184. Gecas, Victor. 1982. “The Self-Concept.” Annual Review of Sociology 8: 1–33. Gerber, Theodore P. 2002. “Structural Change and Post-Socialist Stratification: Labor Market Transitions in Contemporary Russia.” American Sociological Review 67(5): 629-659. ______. 2001a. “The Selection Theory of Persisting Party Advantages in Russia: More Evidence and Implications.” Social Science Research 30(4): 653-671. ______. 2001b. “Paths to Success: Individual and Regional Determinants of SelfEmployment Entry in Post-Communist Russia.” International Journal of Sociology 31(2): 3-37. ______. 2000. “Membership Benefits or Selection Effects? Why Former Communist Party Members Do Better in Post-Soviet Russia.” Social Science Research 29(1): 25-50. Gerber, Theodore P. and Michael Hout. 1998. “More Shock than Therapy: Market Transition, Employment, and Income in Russia, 1991-1995.” The American Journal of Sociology 104(1): 1-50. Gimpleson, Vladimir and Douglas Lippoldt. 2001. The Russian Labour Market: Between Transition and Turmoil. New York, NY: Rowman & Littlefield Publishers, Inc. Gottfredson, Michael R. and Travis Hirschi. 1990. A General Theory of Crime. Stanford: Stanford University Press. Hambleton, Ronald K., Hariharan Swaminathan, and Jane H. Rogers. 1991. Fundamentals of Item Response Theory. Newbury Park, CA: Sage Publications.
© Heili Pals September 2006
175
References Hamil-Luker, Jenifer, Kenneth C. Land, and Judith Blau. 2004. “Diverse Trajectories of Cocaine Use through Early Adulthood among Rebellious and Socially Conforming Youth.” Social Science Research 33: 300-321. Hanley, Eric and Donald J. Treiman. 2005. “Recruitment into the Eastern European Communist Elite: Dual Career Paths.” Research in Social Stratification and Mobility 23: 35-66. Harvey O. J. and Conrad Consalvi. 1960. “Status and Conformity to Pressures in Informal Groups.” Journal of Abnormal and Social Psychology 60: 182-187. Hausman, J. and D. McFadden. 1984. “Specification Tests for the Multinomial Logit Model.” Econometrica 52: 1219-1240. Heckert, Alex and Heckert, Druann. 2002. “A New Typology of Deviance: Integrating Normative and Reactivist Definitions of Deviance.” Deviant Behavior 23: 449-479. Hemker, Bas T., Klaas Sijtsma, and Ivo M. Molenaar. 1995. “Selection of Unidimensional Scales from a Multidimensional Item Bank in the Polytomous Mokken IRT Model.” Applied Psychological Measurement 19(4): 337-352. Hollander, Edwin P. 1960. “Competence and Conformity in the Acceptance of Influence.” Journal of Abnormal and Social Psychology 61: 365-369. Homans, George C. 1961. Social Behavior: Its Elementary Forms. New York: Harcourt, Brace & World. Imai, Kosuke and David A. van Dyk. 2005. “MNP: R Package for Fitting the Multinomial Probit Model.” Journal of Statistical Software 14(3): 1-32. Inglehart, Ronald and Paul R. Abramson. 1994. “Economic Security and Value Change.” American Political Science Review 88(2): 336-353. Jessor, Richard and Shirley I. Jessor. 1975. “Adolescent Development and the Onset of Drinking.” Journal of Studies on Alcohol 36: 27-51. Jessor, Richard, John E. Donovan, and Frances Marie Costa. 1991. Beyond Adolescence: Problem Behavior and Young Adult Development. Cambridge: Cambridge University Press. Johnson, David Richard and James C. Creech. 1983. “Ordinal Measures in Multiple Indicator Models: A Simulation Study of Categorization Error.” American Sociological Review 48: 398-407. Kim, Jae-on and Charles W. Mueller. 1978. Factor Analysis: Statistical Methods and Practical Issues. Beverly Hills, CA: Sage Publications.
© Heili Pals September 2006
176
References Kolenikov, Stanislav and Gustavo Angeles. 2004. “Theory, Simulations, and Applications to Socioeconomic Indices.” Working Paper. Korzhov, Gennady. 1999. “Historical and Cultural Factors of Entrepreneurship ReEmergence in Post-Socialist Ukraine.” Polish Sociological Review 4(128): 503-532. Krohn, Marvin D., Ronald L. Akers, Marcia J. Radosevich, and Lonn Lanza-Kaduce. 1980. “Social Status and Deviance: Class Context of School, Social Status, and Delinquent Behavior.” Criminology 18(3): 303-318. Kryshtanovskaya, Olga and Stephen White. 1996. “From Soviet Nomenklatura to Russian Elite.” Europe-Asian Studies 48: 711-33. Kyvelidis, Ioannis. 2001. “Measuring Post-materialism in Post-Socialist Societies.” European Integration online Papers (EIoP), vol. 5. http://ssrn.com/abstract=302775. Laub, John H. and Robert J. Sampson. 1993. “Turning Points in the Life Course: Why Change Matters in the Study of Crime?” Criminology 31: 301-326. Laub, John H., Daniel S. Nagin, and Robert J. Sampson. 1998. “Trajectories of Change in Criminal Offending: Good Marriages and the Desistance Process.” American Sociological Review 63: 225-238. Lauristin, Marju and Peeter Vihalemm. 1994. “Social Change and Changing Values in Post-Communist Estonia.” International Sociological Association (ISA). Lauristin, Marju and Triin Vihalemm. 1997. “Changing Value Systems: Civilizational Shift and Local Differences.” Pp. 243-264 in Changing Value Systems: Civilizational Shift and Local Differences: Return to the Western World, edited by Marju Lauristin and Peeter Vihalemm, with Karl Erik Rosengren and Lennart Weibull. Tartu, Estonia: Tartu University Press. Lightfoot, Cynthia. 1997. The Culture of Adolescent Risk Taking. New York: The Guilford Press. Long, J. Scott and Jeremy Freese. 2006. Regression Models for Categorical Dependent Variables Using Stata (Second Edition). College Station, TX: Stata Press Publication. Looker, Dianne E. and Peter C. Pineo. 1983. “Social Psychological Variables and Their Relevance to the Status Attainment of Teenagers.” American Journal of Sociology 88(6): 1195-1219. Marshall, Victor W. 2000. “Agency, Structure, and the Life Course in the Era of Reflexive Modernization.” Symposium on “The Life Course in the 21st
© Heili Pals September 2006
177
References Century”, Annual Meetings of the American Sociological Association, Washington D.C., August 2000. http://www.aging.unc.edu/infocenter/resources/2000/marshallv.pdf. Retrieved 6/25/04. McFarland, Daniel. 2004. “Resistance as a Social Drama – A Study of ChangeOriented Encounters.” American Journal of Sociology 109(6): 1249–1318. ______. 2001. “Student Resistance: How the Formal and Informal Organization of Classrooms Facilitate Everyday Forms of Student Defiance.” American Journal of Sociology 107(3): 612-78. McFarland, Daniel A. and Reuben J. Thomas. 2006. “Bowling Young: On Youth Voluntary Associations.” American Sociological Review 71(3): 401-425. McLeod, Jane D. and Timothy J. Owens. 2004. “Psychological Well-Being in the Early Life Course: Variations by Socioeconomic Status, Gender, and Race/Ethnicity.” Social Psychology Quarterly 67(3): 257-278. Menzel, Herbert. 1960. “Innovation, Integration, and Marginality: A Survey of Physicians.” American Sociological Review 25: 704-713. Merton, Robert King. 1968. Social Theory and Social Structure. New York: The Free Press. Meulman, Jacqueline J. and Willem H. Heiser. 1999. SPSS Categories 10.0. Chicago: SPSS Inc. Milgram, Stanley. 1969. Obedience to Authority. New York: Harper & Row, Publishers. Murrell, Peter (ed.). 2001. Assessing the Value of Law in Transition Economies. Ann Arbor: The University of Michigan Press. ______. 1993. “What is Shock Therapy? What Did It Do in Poland and Russia?” Post-Soviet Affairs 9(2): 111-140. Olsson, Ulf. 1979. “The Robustness of Factor Analysis against Crude Classification of the Observations.” Multivariate Behavioral Research 14: 485-500. Ostini, Remo and Michael L. Nering. 2006. Polytomous Item Response Theory Models. Thousand Oaks, CA: Sage Publications. Otto, Luther B. and Archibald O. Haller. 1979. “Evidence for a Social Psychological View of the Status Attainment Process: Four Studies Compared.” Social Forces 57(3): 887-913.
© Heili Pals September 2006
178
References Pals, Heili and Nancy Brandon Tuma. 2004. "Entrepreneurial Activities in Post-Soviet Societies: Impacts of Social Psychological Characteristics." International Journal of Sociology 34(2): 11-38. Phillips, Damon J. and Ezra W. Zuckerman. 2001. “Middle-Status Conformity: Theoretical Restatement and Empirical Demonstration in Two Markets.” American Journal of Sociology 107(2): 379-429. Polishchuk, Leonid. 2001. “Legal Initiatives in Russian Regions: Determinants and Effects.” Pp. 330-369 in Assessing the Value of Law in Transition Economies, edited by Peter Murrell. Ann Arbor, MI: The University of Michigan Press. Rafky, David M. 1979. “School Rebellion: A Research Note.” Adolescence 14(55): 451-464. Rettig, Richard P. 1979. A Study in Student Development: Rebellion and Delinquency as Alternative Responses to Schooling. Palo Alto, CA: R & E Research Associates, Inc. Ridgeway, Cecilia L. 1981. “Nonconformity, Competence, and Influence in Groups: A Test of Two Theories.” American Sociological Review 46(3): 333-347. Róna-Tas, Ákos. 1994. “The First Shall Be Last? Entrepreneurship and Communist Cadres in the Transition from Socialism.” American Journal of Sociology 100: 4069. Rosenberg, Morris. 1965. Society and the Adolescent Self-Image. Princeton, NJ: Princeton University. Rostowski, Jacek. 1993. "Comment on 'Gradual versus Rapid Liberalisation in Socialist Economies' by McKinnon," Proceedings of the World Bank Annual Conference on Development Economies. Washington DC: World Bank. Rotter, Julian B. 1966. “Generalized Expectancies for Internal Versus External Control of Reinforcement.” Psychological Monographs: General and Applied 80: 1–28. Samejima, Fumiko. 1969. “Estimation of Latent Ability Using a Response Pattern of Graded Scores.” Psychometrika Monograph Supplement 17. Sampson, Robert J. and John H. Laub. 1990. “Crime and Deviance Over the Life Course: The Salience of Adult Social Bonds.” American Sociological Review 55: 609-627. Sampson, Robert J., Stephen Raudenbush, and Felton Earls. 1997. “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 277: 918-924.
© Heili Pals September 2006
179
References Schiff, Michael and Richard C. Lewontin. 1986. Education and Class: The Irrelevance of IQ Genetic Studies. Oxford, England: Clarendon Press. Scull, Andrew T. 1988. “Deviance and Social Control.” Pages 667-693 in Handbook of Sociology edited by Neil J. Smelser. Newbury Park, CA: Sage. Sharlet, Robert. 1993. “Russian Constitutional Crisis: Law and Politics Under Yel’tsin.” Post-Soviet Affairs 9(4): 314-336. Sherif, Muzifer. 1936. The Psychology of Social Norms. New York: Harper. Shkaratan, Ovsei. 1992. “The Old and the New Masters of Russia: From Power Relations to Proprietary Relations.” Sociological Research 31: 55-69. Simons, Ronald L., Christine Johnson, Rand D. Conger, and Glen H. Elder, Jr. 1998. “A Test of Latent Trait Versus Life-Course Perspectives on the Stability of Adolescent Antisocial Behavior.” Criminology 36: 217-242. Simons, Ronald L., Eric Stewart, Leslie C. Gordon, Rand D. Conge, and Glen H. Elder. 2002. “A Test of Life Course Explanations for Stability and Change in Antisocial Behavior from Adolescence to Young Adulthood.” Criminology 40: 901-934. Słomczyński, Kazimierz M. and Bogdan W. Mach. 1996. “The Impact of Psychological Resources on Status Attainment: Poland, 1978-1980 and 19921993.” Polish Sociological Review 4(116): 337-352. Slomczynski, Kazimierz M. and Irina Tomescu-Dubrow. 2005. “Friendship Patterns and Upward Mobility: A Test of Social Capital Hypothesis.” Polish Sociological Review 3(151): 221-235. Small, K. A. and C. Hsiao. 1985. “Multinomial Logit Specification Tests.” International Economic Review 26: 619-627. Solnick, Steven Lee. 1993. “Growing Pains: Youth Policies and Institutional Collapse in the Former Soviet Union.” Harvard University: Doctoral Dissertation. Spreitzer, Gretchen and Scott Sonenshein. 2004. “Toward A Construct Definition of Positive Deviance.” American Behavioral Scientist 47: 828-847. Stacey, Barrie and John Davies. 1972. “Drinking Behavior in Childhood and Adolescence: An Evaluative Review.” Journal of Alcohol and Drug Education 17(3): 1-11. Sterlin, Andrei R. 1993. “The Conditions for Entrepreneurship in the CIS (Former USSR).” Pp. 215-236 in Dynamic Entrepreneurship in Central and Eastern
© Heili Pals September 2006
180
References Europe, edited by Derek F. Abell and Thomas Kõllermeier. DELWEL Publishers.
The Hague:
Stinchcombe, Arthur L. 1964. Rebellion in a High School. Chicago: Quadrangle Books. Tanner, Julian, Scott Davies, and Bill O’Grady. 1999. “Whatever Happened to Yesterday’s Rebels? Longitudinal Effects of Youth Delinquency on Education and Employment.” Social Problems 46(2): 250-274. Tikhomirov, Yurii A. 1996. "Problems of Executive-Legislative Relations in the Russian Federation." Pages 272-275 in Democratization in Russia. The Developments of Legislative Institutions, edited by Jeffrey W. Hahn. M.E. Sharpe. Titma, Mikk and Lyudmila Koklyagina (eds.). 1989. Nachalo Puti: Pokolenie so Srednim Obrazovaniem (Beginning the Life Path: A Cohort with Secondary Education). Nauka, Moscow. Titma, Mikk and Rein Murakas. 2004. “Income Inequalities in the Baltic States.” International Journal of Sociology 34(2): 39-71. Titma, Mikk and Brian D. Silver. 1996. “Transitions from Totalitarian Society.” International Journal of Sociology 26(1): 20-43. Titma, Mikk, Brian D. Silver, and Barbara Anderson. 1994. Estonia on the Eve of Independence: Nationalities and Society in the Transition from State Socialism. Washington D.C.: The National Council for Soviet and East European Research. Titma, Mikk, Liina Mai Tooding, and Nancy Brandon Tuma. 2004. “Communist Party Members: Incentives and Gains.” International Journal of Sociology 34(2): 7299. Titma, Mikk and Denis Trapido. 2002. “Prediction of Success in Post-Communist Societies: Evidence from Latvia and Estonia.” Society and Economy 24(3): 297-331. Titma, Mikk and Nancy Brandon Tuma. 2005. "Human Agency in the Transition from Communism: Perspectives on the Life Course and Aging." Pp. 108-143 in Historical Influences on Lives and Aging, edited by K. Warner Schaie and Glenn Elder. New York: Springer. ______. 1995. “Paths of a Generation: A Comparative Longitudinal Study of Young Adults in the Former Soviet Union.” Technical Report, Department of Sociology, Stanford University.
© Heili Pals September 2006
181
References Titma, Mikk, Nancy Brandon Tuma, and Brian D. Silver. 1998. “Winners and Loosers in the Post-Communist Transition: New Evidence from Estonia.” Post-Soviet Affairs 1(2): 114-136. Titma, Mikk, Nancy Brandon Tuma, and Kadi Roosma. 2003. "Education as a Factor in Intergenerational Mobility in Soviet Society." European Sociological Review 19(3): 281-297. Tuma, Nancy Brandon and Mikk Titma. 2003. “The Educational Attainment Process in the Former Soviet Union: Factors Promoting Women's Success.” Paper presented at the Meeting of Research Committee 28 of the International Sociological Association on “Education and Social Inequality.” New York City, August 22-24, 2003. Tuma, Nancy Brandon, Mikk Titma, and Valery B. Yakubovich. 1995. “Sampling Design and Realization of Paths of a Generation.” Technical Report, Department of Sociology, Stanford University, Stanford, CA 94305. Uebersax, John. 2000. “The Tetrachoric and Polychoric Correlation Coefficients.” Retrieved Oct 10, 2005 from http://ourworld.compuserve.com/homepages/jsuebersax/tetra.htm#ex3, revised at July 27, 2000. Van Zon, Hans. 1998. “The Mismanaged Integration of Zaporizhzhya with the World Economy: Implications for Regional Development in Peripheral Regions.” Regional Studies 32(7): 607-618. Verhoeven, Willem-Jan, Wim Jansen, and Jos Dessens. 2005. “Income Attainment during Transformation Processes: A Meta-Analysis of the Market Transition Theory.” European Sociological Review 21(3): 201-226. Wang, Li-Ya, Edward Kick, James Fraser, and Thomas Jerome Burns. 1999. “Status Attainment in America: The Roles of Locus of Control and Self-Esteem in Educational and Occupational Outcomes.” Sociological Spectrum 19: 281298. West, Brad. 2004. "Synergies in Deviance: Revisiting the Positive Deviance Debate." Electronic Journal of Sociology 4(1). Willis, Paul E. 1977. Learning to Labor: How Working Class Kids Get Working Class Jobs. Westmead, UK: Saxon House. Wilson, James Q. and Richard J. Herrnstein. 1985. Crime and Human Nature. New York: Simon and Schuster.
© Heili Pals September 2006
182
References Wong, Raymond Sin-Kwok. 2004. “Egalitarianism versus Social Reproduction: Stratification in Eastern Europe.” Research in Sociology of Education 14: 139170. ______. 2002. “Occupational Attainment in Eastern Europe under Socialism.” Research in Social Stratification and Mobility 19: 191-230. Yakubovich, Valery. 2001. Between Exchange and Reciprocity: Matching Workers with Jobs in a Local Russian Labor Market. Unpublished Ph.D. dissertation, Stanford University.
© Heili Pals September 2006
183
