Teaching and Teacher Education 27 (2011) 831e840

Contents lists available at ScienceDirect

Teaching and Teacher Education journal homepage: www.elsevier.com/locate/tate

Grading styles and disciplinary expertise: The mediating role of the teacher’s perception of the subject matter Liat Biberman-Shalev a, *, Clara Sabbagh a, Nura Resh b, Bracha Kramarski c a

Faculty of Education, University of Haifa, Haifa, Israel School of Education, Hebrew University of Jerusalem, Jerusalem, Israel c School of Education, Bar-Ilan University, Ramat Gan, Israel b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 January 2010 Received in revised form 14 December 2010 Accepted 13 January 2011

Based on a sample of 312 high school teachers who participated in the Israeli PISA assessment of student academic achievement in 2002, the current study examines the mediatory role of their perception of the subject matter (as “open/flexible” or “closed/hierarchical”) in the relation between their disciplinary expertise (language, mathematics or science) and their grading style (performance-output or effortinput). The mediation hypothesis was completely supported for all disciplines in the case of perception of subject matter as open/flexible. With respect to the closed/hierarchical perception, it was supported only for the comparison of science vs. mathematics. Ó 2011 Elsevier Ltd. All rights reserved.

Keywords: Grades Subject matter Disciplinary expertise Mediation Teachers Distributive justice

1. Introduction In his seminal book, Spheres of Justice, Walzer (1983) suggests that education is a distinct sphere of justice in which different kinds of valued “goods” are distributed (e.g., the right to education, teaching methods, teacherestudent relationships and grades) (see Sabbagh, Resh, Mor, & Vanhuysse, 2006). Following that rationale, the current study focuses on a major educational distribution sphere e namely, the just distribution of grades, as perceived by Israeli teachers from different disciplinary backgrounds. The significance of grades as a valued good is manifold. They serve as a major selection mechanism in schools, which ultimately determines the student’s life chances and earnings (Resh, 1998; Kelly, 2008; Miller, 1998; Vanfossen, Jones, & Spade, 1987). Further, they shape the student’s self-image, motivation and popularity among peers, as well as parents’ expectations (Entwisle, Alexander, & Olson, 2007; Nisan, 1985; Roscigno & Ainsworth-Darnell,1999). Finally, they are used by teachers as control mechanisms to maintain authority and discipline (Deutsch, 1985; Nevo, Wolfe, & Goldblatt, 1988). Under the assumption that grade distribution has a wide range of effects on * Corresponding author. Department of Leadership & Policy in Education, University of Haifa, Mount Carmel, Haifa 31905, Israel. Tel.: þ972 4 8240867; fax: þ972 4 9240911. E-mail address: [email protected] (L. Biberman-Shalev). 0742-051X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tate.2011.01.007

student welfare, our study attempts to unveil what Israeli teachers’ distribution preferences are in this regard and to reveal a possible mediating mechanism underlying this preference. Various studies have examined students’ and teachers’ grade distribution preferences (Sabbagh, Faher-Aladeen, & Resh, 2004; Deutsch, 1985; Nisan, 1985; Tata, 1999; Thorkildsen, 1989). The literature on teaching practices suggests that the grading styles of high school teachers (preferences for just distribution rules when distributing grades) are mainly guided by a universal meritocratic principle that focuses on the performance-output aspect of learning e i.e., stresses academic performance or knowledge exhibition in tests over effort (Brookhart, 2004; Linn & Miller, 2005). Notwithstanding the strong consensus regarding this principle (Bell, 1977; Conley, 1996; Hurn, 1985), the rules for determining the meritocratic criteria in grading may vary. Moreover, the various rules are differentially weighted and considered simultaneously (Sabbagh, Cohen, & Levy, 2003; Leventhal, 1976). This preference for performance-output is challenged by a more radical educational critique of the grading system that calls for an alternative assessment practice focusing on effort-input aspects of learning, stressing the combination of rules of effort, class behavior and students’ need for encouragement over academic performance (Atkin, Black, & Coffey, 2001; Palm, 2008). Teachers’ grading styles are not necessarily uniform, but may be affected by personaleprofessional characteristics (e.g., gender,

832

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

seniority) (Sabbagh, Biberman-Shalev, & Resh, 2009; DeBoer et al., 2007) and other socio-cultural factors (Doran, Lawrenz & Hegelson, 1994). For example, several studies have indicated that academic performance and student ability receive more emphasis in China and Hong Kong than in the USA (Brown et al., 2009; Hutchison & Bailey, 2006; McCargar, 1993). These findings are attributed to the centralized structure of the educational system in Asian countries, promoting an exam-oriented culture where teachers are constantly held accountable for student achievements. Related studies have indicated that, when considering the cut-off points of grades for selection purposes, as well as the relative importance assigned to final exams (academic performance) versus continuous assessment during an academic year (effort), teachers in Nepal, Saudia Arabia and various African countries stress lower cut-off points and final exams more than their counterparts in the USA and New Zealand (Al-Sadan, 2000; Crooks, 2002; Rajbhandari & Wilmut, 2000). These differences are attributed to higher selection standards in the latter group of countries, which tend to be characterized by greater levels of economic and technological development. In sum, grading styles are a complex and multidimensional phenomenon whereby teachers simultaneously consider, and differently weigh, distribution rules in different situations (Sabbagh, et al., 2003; Carbonaro, 2005; Leventhal, 1976; Randall & Engelhard, 2010). While research has examined the effect of teachers’ personaleprofessional and cultural backgrounds on their grading distribution preferences, relatively little attention has been given to the effect of teachers’ disciplinary expertise on these preferences (Duncan & Noonan, 2007). With this in mind, and focusing on the context of Israeli society, our study attempts to further explore this association and to add to this equation the possible mediatory role of teachers’ perception of their subject matter. 2. Literature review 2.1. Grading style and disciplinary expertise Two relatively recent pioneering studies have examined the extent to which teachers’ grading styles are affected by their disciplinary expertise. McMillan (2001) investigated the grading practices and types of assessments used by 1483 high school and middle school teachers in the USA across four subject matter domains: science, social studies (history), mathematics and English. Specifically, teachers were asked to indicate the degree to which they apply different distribution rules (effort, academic performance) and types of assessment (tests, essays, portfolios) when distributing grades. Findings indicated that these two forms of grading practices are affected by the kind of subject matter teachers teach. When grading, social studies teachers tended to emphasize academic performance and tests less than mathematics and English teachers, focusing more on such academic enablers as effort, class participation, completion of homework, essay writing and preparing portfolios. Similarly, a study of grade distribution preferences among 372 Israeli high school teachers found that language and science teachers tended to emphasize effort more and attributed lower importance to academic performance than mathematics teachers (Resh, 2009). The study findings were attributed to the “inherent nature” of the subject matter: whereas mathematics can be characterized as hierarchically structured, language and science are perceived as having a more modular and flexible structure. However, neither of the above studies delved into the mechanisms that shape the “inherent nature” of the subject matter. Elaborating on McMillan (2001) and Resh (2009), we thus attempt to go a step further by bringing to light one of the mechanisms that may explain the association between disciplinary expertise and grading style. Specifically, we suggest that teachers’ perceptions of

the subject matter they teach e i.e., their perception of curricular aims and what they consider adequate pedagogical practices of imparting knowledge in their specific disciplines e may be the mediating factor in this association (Becher, 1989; Ylijoki, 2000). We base this assumption on the notion that teachers are not passive actors, who behave in accordance with some objective disciplinary structural constraints, but rather are active agents who dynamically interpret and implement these relatively “objective” disciplinary contents in their pedagogical practices, including grading. 2.2. Disciplinary expertise and perception of subject matter Several studies have examined how disciplinary expertise affects teachers’ perceptions of their subject matter (Grossman & Stodolsky, 1995; McLaughlin, Talbert, & Bascia, 1990; Shulman, 1986; Tyack & Tobin, 1994). In one such study, Stodolsky and Grossman (1995) examined the perceptions of American high school teachers in various disciplines in keeping with a dominant approach within the pedagogical tradition that focuses on the “internal structure” of knowledge. Their findings revealed that, in comparison to mathematics and foreign language teachers, English and science teachers perceived their subject matter as more flexible, modular and dynamic (hereafter, open/flexible structure). These teachers also reported high levels of control over curricular contents in the process of teaching (e.g., Cunningham, Zibulsky, & Callahan, 2009; Siskin, 1994). In contrast, mathematics and foreign language teachers perceived their discipline as static (fixed), more rigidly defined, standardized and sequential (hereafter, closed/hierarchical structure) (see also Loewenberg-Ball, Thames, & Phelps, 2008; Smith, 1996). In these disciplines, students are required to imitate teachers’ demonstrations and to memorize material through repeated drilling and practice. Some studies suggest that perception of the subject of matter when teaching language has ambivalent meaning. On the one hand, the subject matter is loaded with technical elements, such as spelling and grammar, eliciting a more “closed” perception of the subject matter structure that leads teachers to apply practices of dictation and error correction (Snow, Griffin, & Burns, 2005). On the other hand, it offers an opportunity for students to enrich their use of language and literacy, thus eliciting an “open” perception focusing on inquiry and discussion, which enable reflection on the social norms produced and reproduced in texts (Taylor & Otinsky, 2007). Similarly, studies among science teachers have unveiled an ambivalent perception of their subject matter. For instance, Van-Driel, Bulte and Verloop (2005) found that Dutch high school chemistry teachers not only perceived the teaching of science to emphasize fundamental theoretical concepts on the basis of facts, but also to be more learner oriented, based on activities in which knowledge is considered a resource for understanding societal issues and choosing meaningful social actions that further the welfare of human beings. These findings suggest that teachers in certain disciplines mix perception of an open/flexible structure with a closed/hierarchical one, whereas those in other disciplines tend to focus on the closed structure (Bernstein, 1971). In keeping with the sociology of knowledge tradition, which assumes that a discipline’s structure and content are socially constructed (i.e., are affected by the socio-cultural context) (Kuhn, 1970; Mannheim, 1991), some research suggests that the perception of subject matter may vary across countries. For example, several studies have indicated that teachers’ perception of subject matter in science, mathematics and language is more open and flexible in Australia and the USA than in Taiwan, China and Hong Kong (Aldridge, Fraser, & Huang, 1999; Lee, 2008). These differences are explained by the post-modern pedagogy that characterizes the Western world, whereby students are more involved in the learning process, teaching stresses exploration and inquiry more, and students are more autonomous and able to express their opinions

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

833

Direct effect:

Teachers’ grading style (Y): Disciplinary expertise (X): Language Mathematics Science

0. Performance-output 1. Effort-input

Path c

Mediation effects:

Path c'

Disciplinary expertise (X): Language Mathematics Science

Path a

Teachers’ perception of the subject matter as open/flexible (M1)

Teachers’ grading style (Y): Path b

0. Performance-output 1. Effort-input

Teachers’ perception of the subject matter as closed/hierarchical (M2)

Control variables: Gender Educational level Seniority Fig. 1. Study model: Perceptions of subject matter structure (open/flexible and closed/hierarchical) as mediating between disciplinary expertise and grading style.

with greater freedom (Trelstad, 2008). Given that Israel is a developed country which has been strongly affected by the Western world (Cohen, 1989), we assume that its disciplinary culture is closer to the one characterizing the USA and Australia.1 In such a learning context, the mix of open/flexible and closed/hierarchical structure in some disciplines may be more salient.

consider students’ efforts in the learning process and their welfare and needs when distributing grades; that is, they are more likely to prefer the effort-input grading style.

2.3. Perception of the subject matter and grading style

In examining the mediatory role of the teacher’s perception of the subject matter in the association between disciplinary expertise and grading style, we offer a mediation model (Fig. 1) that follows Baron and Kenny’s (1986) three-path scheme. The first path (path c) establishes the relationship between teachers’ disciplinary expertise and their grading styles. As mentioned above, this path has received some empirical support in previous studies (Resh, 2009; McMillan, 2001). In the current research, we both replicate and expand these findings by applying a holistic approach to the evaluation of grading styles (the dependent variable), which takes into account the simultaneous use of different distribution rules. The second and third paths of the model constitute our main contribution by establishing the association between disciplinary expertise and perception of the subject matter (the mediation variables; path a), as well as the effect of these perceptions on their grading style (the dependent variable; path b). As individual background variables have been found to affect distribution preferences (Hegtvedt, 1992; Robinson & Bell, 1978; Wagner, 1995), we include gender, educational level and seniority as control variables across the mediation steps. In line with our mediation model, we offer the following research hypotheses

To the best of our knowledge, the association between teachers’ perception of the subject matter structure and their grading styles has not been empirically tested. Following Deutsch (1985), we suggest that grading styles are context bound; that is, they are likely to reflect teachers’ perceptions of the subject matter and pedagogical practices. Thus, it is reasonable to expect teachers who perceive their subject matter as closed/hierarchical to be more taskoriented and promote pedagogical practices that further students’ linear progress in the material. This type of knowledge is organized in linear-developmental terms, and teachers have to follow certain sequential steps in order to cover the material, implying that students have to master certain tasks before they can continue to the next level. Moreover, teachers must constantly evaluate students’ proficiency and achievement at each step by means of tests. This task-oriented approach fits the equitarian (output-based) rule preference for grade distribution (Deutsch, 1985). Moreover, perception of a closed/hierarchical structure, especially in mathematics and foreign languages, is in keeping with the organization of learning in schools in Israel and other countries into ability groups and tracks, which also represents an output-oriented, meritocratic ideology (Alpert & Bechar, 2008; Grossman & Stodolsky, 1995). In contrast, teachers who perceive their subject matter as open/ flexible are more likely to promote pedagogical practices that focus on student welfare and creativity. The “nature” and organization of knowledge in these subjects allow greater discretion in the choice and amount of content to be covered, as well as the order of topics presented. The learning process is based more on experimental inquiry, discussion, negotiation of contrasting approaches and variegated teaching methods. Thus, teachers are more likely to

1 A systematic investigation of this issue is beyond the scope of the present study and should be examined in future research.

3. Research model and hypotheses

Hypothesis 1 (disciplinary expertise and grading style; see Fig. 1, path c): Mathematics teachers will prefer the performanceoutput grading style more than language and science teachers, ascribing greater weight to academic achievement than to effort, class behavior and need. Conversely, language and science teachers will favor the effort-input grading style more than their mathematics counterparts, placing more weight on effort, behavior and need than achievement. Hypothesis 2 (disciplinary expertise and perception of subject matter; see Fig. 1, path a): Language and science teachers will perceive their subject matter as more open/flexible (dynamic, modular) than mathematics teachers, thus enabling greater autonomy in curriculum choices and teaching practices. Conversely,

834

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

mathematics teachers will perceive their subject matter as more closed/hierarchical (linear, sequential and task-oriented in its learning practices) than language and science teachers. Hypothesis 3 (perception of subject matter and grading style; see Fig. 1, path b): When distributing grades, teachers who hold an open/flexible perception of their subject matter will prefer the effort-input grading style over the performance-output grading style. Conversely, teachers who hold a closed/hierarchical perception of their subject matter will prefer the performanceoutput grading style. These hypotheses can be integrated into the following mediation hypothesis: Mediation Hypothesis (Fig. 1, path c’): Language and science teachers will perceive their subject matter as open/flexible, and so will prefer the effort-input grading style, more than mathematics teachers. Conversely, mathematics teachers will perceive their subject matter as closed/hierarchical, and so will prefer the performance-output grading style, more than language and science teachers.

4. Method 4.1. Sample The study was carried out in a national sample of 165 high schools, proportionally representing the four sectors of the Israeli public educational system: Jewish secular (99 schools: 60%), Jewish religious (23 schools: 14%), Jewish ultra-orthodox (16 schools: 10%) and Arab (27 schools: 16%). This was the same national school sample drawn for the Israeli PISA assessment of student academic achievement. In each school, three teachers were randomly drawn from the list of tenth-grade teachers in each of the three tested disciplines: language (Hebrew or Arabic), mathematics and science (total of 3  165 ¼ 495 teachers). The response rate was 77%, yielding a total of 380 respondents. We explain this moderate, partial response rate by the fact that recruitment was on a voluntary basis. Moreover, Israeli teachers often carry a heavy load, teach large classes and receive low salaries, all of which may negatively impact their willingness to participate in the study. In addition, as a result of missing data on disciplinary expertise and grading style, our final sample included 312 respondents: 107 language teachers, 98 mathematics teachers and 107 science teachers. Despite these missing cases, the distribution of teachers across sectors of the educational system in this final sample is similar to the one obtained for the original sample of 495 teachers. Of the final sample, 71.47% were females and 28.53% males; 62.18% held a bachelor’s degree and 37.82% held a master’s degree or higher. The average seniority of teaching was M ¼ 18.31 (SD ¼ 8.96) years. Comparison of our sample to the population of high school teachers in the country (Israeli Central Bureau of Statistics, 2002) revealed that it represents this population to some extent in terms of gender, level of education and seniority. Out of 62,838 teachers who worked in 1494 schools, 65.6% were female and 34.4% male; 75.0% hold a bachelor’s degree or less and 25.0% hold a master’s degree or more. The average seniority of teaching was M ¼ 17.8 years. Moreover, out of the total population of teachers, about 22% (N ¼ 13,903) taught (Hebrew/Arabic) language; 14% (N ¼ 8803) taught mathematics and about 20% (N ¼ 12,876) taught various types of science-related subject matter (biology, chemistry, physics, technology and computers). The rest (44%) taught other subjects, such as history, art, foreign language and Bible, which are not examined in the current study.

4.2. Instrument Teachers were asked to fill out an anonymous questionnaire that included items about different aspects of their teaching role. In our study, we included only items pertaining to teachers’ personaleprofessional background, perceptions of their subject matter and rules of grade distribution. 4.3. Measures 4.3.1. Dependent variable: grading style (teachers’ rule preferences) Teachers were asked to assign a relative weight (in percentages) to each of five rules when distributing grades: (1) ability, (2) tests, (3) effort invested in the learning process, (4) behavior in class and (5) student’s need for encouragement. These five variables are interdependent, and the distribution of percentages had to add up to 100%, representing the final grade (cases that did not reach this total were excluded from the sample). These preferences yielded two main grading styles that were coded as a binary categorical variable: 0 ¼ performance-output style and 1 ¼ effort-input style. 4.3.2. Independent variable: disciplinary expertise This categorical variable was coded into two dummy variables (Hardy, 1993): (a) language, (b) science. Mathematics was used as the comparison point. 4.3.3. Mediation variable: perception of the subject matter as open and/or closed Exploratory factor analysis (PCA) with varimax (orthogonal) rotation2 was applied to 15 items pertaining to teachers’ perceptions of their subject matter (see Appendix A). Possible responses ranged from 1 ¼ not at all to 4 ¼ very much. The analysis yielded two orthogonal factors (r ¼ .048, p > .05) with Eigenvalues > 1 that explained 37% of the variance (Brown, 2001). Thus, these factors are not two opposite points of one continuum, but rather represent two distinct perceptions of the subject matter (see Kerlinger, 1958). Accordingly, we constructed two measures, defined as mean scores across items, that corresponded to teachers’ perceptions of subject matter as open/flexible and closed/hierarchical (see Appendix A). The open/flexible perception included 8 items referring to the modular, flexible and autonomous use of subject matter e.g., “The subject matter enables critical evaluation of the material” and “Curricular planning for the subject matter is based on the teaching material’s relevance to students” (Cronbach’s a ¼ .72; M ¼ 3.06, SD ¼ .46). The closed/hierarchical perception included 7 items referring to the static, linear, hierarchical nature of the subject matter, requiring drilling e.g., “The subject matter is based on systematic ordinal thinking” and “Use of a teaching strategy of memorizing and instilling information” (Cronbach’s a ¼ .73; M ¼ 3.25, SD ¼ .47). 4.3.4. Control variables: teachers’ personaleprofessional characteristics As personaleprofessional characteristics were found to affect distribution preferences in a wide variety of studies (Hegtvedt, 1992; Robinson & Bell,1978; Wagner,1995), they are held as control variables. (a) Gender: 0 ¼ male; 1 ¼ female. (b) Educational level: 0 ¼ bachelor’s degree; 1 ¼ master’s degree or higher.

2 This statistical technique transforms a number of possibly correlated items (i.e., items about teachers’ perceptions of the subject matter) into a smaller number of uncorrelated factors (open/flexible or closed/hierarchical perception) (Costello & Osborne, 2005; Thompson, 2004).

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

835

Table 1 Descriptive values of grading rules and rule weight distribution frequencies (%).

Grading rules Mean % (SD) Min Max Rule weights (%): 0e10 11e20 21e30 31e40 41e50 51e60 61e70 71e80 81e90 91e100 Total N

Academic performance

Effort-related rules

Ability

Effort

5.47 (9.56) 0 74 88 7 3 1 0 0 0 1 0 0 100% 312

Tests 61.96 (21.43) 0 100 4 2 4 7 9 21 18 22 11 2 100% 312

19.58 (14.93) 0 95 36 37 17 5 1 0 0 3 0 1 100% 312

(c) Seniority: number of years working as a teacher. 5. Results 5.1. Grading styles In presenting results with regard to grading styles, we begin with a descriptive account of distribution preferences, which gives an overall picture of grading practices. This is followed by a presentation of cluster analysis results, which provide a more parsimonious view of grading styles. 5.1.1. Descriptive analysis of distribution preferences Table 1 displays the average weight that teachers assigned to the five grading rules, as well as the frequencies (in percentages) of these weights broken down into deciles. The table reveals that tests were the teachers’ most preferred rule for distributing grades (M ¼ 61.96, SD ¼ 21.43), followed by preference for the rule of effort (M ¼ 19.58, SD ¼ 14.93). Lower weights were assigned to students’ need for encouragement (M ¼ 6.86, SD ¼ 6.91), class behavior (M ¼ 6.13, SD ¼ 6.43) and, finally, ability (M ¼ 5.47, SD ¼ 9.56). Notwithstanding the strong weight (80%) given to tests and effort, there is considerable variance among teachers that remains to be explained. Looking at the breakdown into deciles, almost all teachers (99%) assigned low weight (0e40) to ability, behavior and need, and a very large majority (95%) ascribed low weight (0e40) to effort. In contrast, most teachers (85%) gave moderate to strong weight (41e100) to tests, with the bulk falling in the 51e80 range. Finally, Table 1 presents an overall measure for describing the relation between academic performance and effort. Based on a previous study (Sabbagh, Cohen, & Levy, 2003), this measure was defined as the difference between two groups of rules: academic performance (tests and ability) and effort-related rules (effort, behavior and need).3 Findings indicate an average score of 34.86

3 The rationale for this measure, which was supported by a factor analysis applied to a sample of 9000 students, is that distribution rules can be distinguished by the degree of differentiation (inequality) they promote. As opposed to academic performance and ability, rules such as effort, class behavior and need are likely to promote greater equality among students because they are easily applicable to all students and depend on their will and motivation. In contrast, academic performance and ability are more differentiating, because they require innate talents that are unevenly distributed and much less dependent on students’ active investment (Sabbagh, Cohen, & Levy, 2003).

Behaviour 6.13 (6.43) 0 30 88 9 3 0 0 0 0 0 0 0 100% 312

Need 6.86 (6.91) 0 50

Total

Academic performance e Effort

100.00

34.86 (38.22) 100 100

88 9 2 0 1 0 0 0 0 0 100% 312

(SD ¼ 38.22), which points to a clear preference for academic performance over effort-related rules. 5.1.2. Cluster analysis and the definition of grading styles While the descriptive analysis of average weights that teachers assign to rules when distributing grades portrays a detailed account of their considerations, it has the disadvantage of focusing on overall tendencies, thereby ignoring the possibility that teachers weigh several rules simultaneously (rather than switching between one single rule and another) and combine them together in different ways to arrive at decisions with respect to grades. Accordingly, we applied k-means clustering (Everitt, Landau, & Lesse, 2001) to capture teachers’ simultaneous rule preferences for distributing grades. This technique reduces the multidimensionality of the teachers’ combinations of rule preferences into a minimal number of distinct k clusters or profiles of grading styles. It does so via an algorithm that partitions observations into clusters that maximize between-cluster variation and minimize within-cluster variation while preserving the proportion of weight assigned to each of the five rules. This type of analysis thus enables the comparison of teachers’ grading styles in a parsimonious form: rather than considering the five-rule preferences separately, it captures them in a minimum number of clusters or grading styles. Application of cluster analysis to our data revealed that, when specifying k > 2 clusters, the obtained solutions did not distinguish well between grading styles. Thus, in keeping with our theoretical assumptions, we specified k ¼ 2 clusters. To determine the extent to which these two clusters (grading styles) significantly differ from each other, we conducted a MANOVA, with the five distribution rules as dependent variables and the two clusters as independent variables. Findings indicated significant differences between the two clusters as predictors of the five dependent variables (distribution rules) (Hotelling’s Trace ¼ 131.664, F(4,312) ¼ 10,105.218, p < .001, h2 ¼ .992). Table 2 presents the two grading styles, together with mean scores of each distribution rule. As the table shows, the most dominant grading style was performance-output, favored by 243 out of 312 teachers (80%). In this style of grading, teachers assigned the strongest weight to tests (mean ¼ 71.11), much less weight to effort (mean ¼ 14.91) and minor weight to the remaining rules (means: ability ¼ 3.31; behavior ¼ 4.73, need ¼ 5.94). In contrast, the effort-input style was favored by only 69 out of 312 teachers (20%). In this style, student effort was the most preferred rule when considering grade distributions (mean ¼ 36.01), higher than

836

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

Table 2 Distribution of grading style clusters and mean scores (standard deviation) per distribution rule. Grading style

Effort-input Performance-output t

Code

1 0

N

69 243

Proportion

.22 .78

Mean scores (SD) of the five rules Ability

Tests

Effort

Behavior

Need

Total

13.09 (15.96) 3.31 (4.96) .83***

29.72 (16.34) 71.11 (11.62) 23.69***

36.01 (20.89) 14.91 (8.05) 12.79***

11.07 (7.82) 4.73 (5.19) 7.92***

10.10 (9.90) 5.94 (5.46) 4.55***

100% 100%

Variance between clusters ¼ .17; Variation ratio ¼ .22; Index of diversity ¼ .35; Index of qualitative variance ¼ .69. ***p < .001.

preference for tests (mean ¼ 29.72). The remaining rules received similar weights (means: ability ¼ 13.09; behavior ¼ 11.07; need ¼ 10.10). Thus, the effort-input style reflects a more balanced profile of preferences. As Table 2 shows, t-tests examining differences between clusters (grading styles) with respect to each of the five rules yielded significant results (p < .001). Variance between the two clusters, tested by the Index of Diversity and the Index of Qualitative Variance, was considerable (D ¼ .35, IQV ¼ .69, respectively) (Weisberg, 1992).4 5.2. Mediation model The methodology developed by Baron and Kenny (1986) has proven useful for testing mediation models; however, it only applies to linear regressions. Owing to the dichotomous nature of the dependent variable, a binary logistic regression was more suitable for use with grading style (for details, see MacKinnon & Dwyer, 1993). In our mediation model, we therefore applied a binary logistic regression to the first and third paths (c and c’ in Fig. 1), where the dependent variable is the dichotomous grading style, and used an ordinary linear regression with respect to subject matter perceptions, a continuous variable (path a). 5.2.1. Path c: the direct effect of disciplinary expertise on grading style We first carried out a hierarchical binary logistic regression, with disciplinary expertise as the independent variable and gender, education and seniority as the personaleprofessional variables (controls). The control variables were entered in the first step and disciplinary expertise in the second step. Findings in Table 3 (Model 1) indicate that the personaleprofessional (control) variables had no significant effect on grading style (c2 ¼ 4.53, p > .05). However, when disciplinary expertise was added to the equation (Model 2), it significantly affected respondents’ preferences for a grading style (c2 ¼ 6.73, p < .05). Specifically, the result for language teachers yielded an odds ratio 5of 2.39 (B ¼ .87, p < .05, 95% CI [1.14, 4.99]); for science teachers the odds ratio was 2.23 (B ¼ .80, p < .05, 95% CI [1.07, 4.66]). This suggests that language teachers are 2.39 times more likely than mathematics teachers to adopt the effort-input grading style, while science teachers are 2.23 times more likely than mathematics teachers to use that grading style, corroborating Hypothesis 1. 5.2.2. Path a: the effect of disciplinary expertise on subject matter perception We first carried out a hierarchical linear logistic regression, with disciplinary expertise as independent variable and gender, education

4 The authors are grateful to an anonymous reviewer for his/her comments on the variance of binary variables. 5 The odds ratio reflects effect size in a binary logistic regression. It represents the probability that teachers would prefer the effort-input grading style divided by the probability that they would prefer the performance-output grading style (see Haddock, Rindskopf, & Shadish, 1998).

and seniority as the controls. The control variables were entered in the first step and disciplinary expertise in the second step. 5.2.2.1. The open/flexible perception. Supporting Hypothesis 2, findings in Table 4 (Model 1 for open/flexible perception) indicate that the control variables had no significant effect, but the two dummy disciplinary expertise variables significantly affected perception of the subject matter as open/flexible (Model 2), accounting for 6% of the variance (F(5,306) ¼ 3.91, p < .001). Specifically, language and science teachers perceived their subject matter as more open/flexible than mathematics teachers (b ¼ .18, p < .01, 95% CI [.05, .30]; b ¼ .25, p < .001, 95% CI [.12, .37], respectively). 5.2.2.2. The closed/hierarchical perception. With respect to the second half of Hypothesis 2, findings in Table 4 (Model 1 for closed/ hierarchical perception) indicate that the control variables accounted for only 2% of the variance of teachers’ closed/hierarchical perception (F(3,308) ¼ 3.41, p < .05). The only significant control variable in this model was gender (b ¼ .16, p < .01, 95% CI [.04, .26]), suggesting, somewhat unexpectedly, that female teachers have a greater tendency to perceive their subject matter as closed/hierarchical. The addition of disciplinary expertise to the equation significantly predicted teachers’ closed/hierarchical perception (Model 2), accounting for 13% of the variance (F(5,306) ¼ 10.64, p < .001). As expected, science teachers had significantly less closed/hierarchical perceptions than mathematics teachers (b ¼ .35, p < .001, 95% CI [.47, .22]). Language was also negatively associated with closed/ hierarchical perception, though not significantly (b ¼ .01, p > .05, 95% CI [.13, .12]). That is, language teachers did not significantly differ from their mathematics colleagues in the extent to which they perceived their subject matter as closed/hierarchical. Thus, the second part of Hypothesis 2 was corroborated only for science teachers. 5.2.3. Path c’: the mediating role of subject matter perception Finally, we examined the extent to which the effect of disciplinary expertise on grading style (path c’) is mediated by the perception of subject matter as open/flexible and closed/hierarchical. Accordingly, we applied a binary logistic regression, entering the control variables in the first step, the mediators in the second step and disciplinary expertise (as two dummy variables) in the last step. Table 3 Hierarchical binary logistic regression of disciplinary expertise (controlling for personaleprofessional variables) on grading style. Model 1 B

SE

Step 1: Background variables: Gender .24 .37 Educational level .55 .28 Seniority .01 .02 Step 2: Disciplinary expertise: Language Science Nagelkerke R2 *p < .05; **p < .01; ***p < .001.

Model 2 Odds ratio

B

SE

Odds ratio

1.27 1.72 1.00

.20 .54 .01

.32 .29 .02

1.22 1.71 1.01

.87* .80*

.38 .38

2.39* 2.23* .05

.02

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

837

Table 4 Hierarchical regressions of disciplinary expertise (controlling for personaleprofessional variables) on open/flexible and closed/hierarchical perceptions. Open/flexible perception Model 1

b Step 1: Background variables: Gender .01 Educational level .10 Seniority .04

df)

Model 1

Model 2

SE

b

SE

b

SE

b

SE

.06 .05 .01

.01 .09 .06

.06 .05 .01

.16** -.09 .02

.06 .05 .01

.15** .04 .01

.06 .05 .01

.18** .25***

.06 .06

.01 .35***

.06 .06

Step 2: Disciplinary expertise: Language Science R2 F (df,

Closed/hierarchical perception Model 2

.01 F(3,308) ¼ 1.32

.06 F(5,306) ¼ 3.91***

.02 F(3,308) ¼ 3.41*

.13 F(5,306) ¼ 10.64***

*p < .05; **p < .01; ***p < .001.

Table 5 (Model 1) indicates that the control variables had no significant effect on grading style (c2 ¼ 4.53, p > .05). Confirming Hypothesis 3, the two mediators (Model 3) did have a significant effect on grading style (c2 ¼ 20.93, p < .001). Specifically, the open/ flexible perception had an odds ratio of 3.57 (p < .001, 95% CI [1.73, 7.14]) and the closed/hierarchical perception had an odds ratio of .43 (p < .01, 95% CI [.24, .78]). This means that teachers who perceived their subject matter as open/flexible were 3.57 times more likely to adopt the effort-input grading style than the performance-output style (B ¼ 1.27, p < .001), while those who perceived their subject matter as closed/hierarchical were .43 times (less) likely to adopt the effort-input grading style (B ¼ .85, p < .01). Finally, examination of the mediation hypothesis (Model 4) reveals that, when perception of the subject matter is introduced into the model of effects on grading style, the estimated effect of both science and language is no longer significant. In the case of science, the odds ratio drops from a significant value of 2.23 to 1.25 (p > .05, 95% CI [.56, 2.79]), and for language the odds ratio drops from a significant value of 2.39 to 1.90 (p > .05, 95% CI [.88, 4.09]). While the significant effect of disciplinary expertise on grading style was neutralized, the mediation effects remained significant: The open/flexible perception had significant odds ratios of 3.57 and 3.31 (p < .001, 95% CI [1.62, 6.76]) and the closed/hierarchical perception had significant odds ratios of .43 and .42 (p < .01, 95% CI [.22, .79]). However, because language does not predict a perception of subject matter as closed/hierarchical (see Table 4, Model 2), it cannot be claimed that it plays a mediating role in our model. Therefore, in the case of closed/hierarchical perception, the mediation hypothesis was supported only for the comparison of science vs. mathematics. In contrast, for the perception of subject matter as open/flexible, findings completely supported the mediation hypothesis for all disciplines.

6. Discussion In line with Grossman and Stodolsky’s (1995) argument that disciplinary expertise structures teachers’ pedagogical beliefs and practices, as well as Walzer’s (1983) conceptualization of education as a “sphere of justice,” we investigated teachers as justice agents who are guided by a variety of rules when distributing grades (Sabbagh et al., 2006). Specifically, the study examined the extent to which teachers’ grading styles are associated with their disciplinary expertise, as well as the degree to which this relation is mediated by their disciplinary culture e their perception of the subject matter as open/flexible or closed/hierarchical. Based on the assumption that, when grading, teachers simultaneously apply a number of distribution rules in some combined multivalent weighted fashion (Sabbagh, Cohen, & Levy, 2003; Leventhal, 1976), we defined a priori a two-dimensional grading style which appeared to fit the empirical structure obtained in a cluster analysis (for details, see Sabbagh, Biberman-Shalev, & Resh, 2009) the performance-output style (predominantly favored), which stresses actual performance over effort, class behavior and need, and the effort-input style, which favors effort and behavior over performance. This twodimensional definition is an important contribution to current research, as it can facilitate a systematic cross-cultural examination of the multidimensionality of teachers’ grading styles. Further investigation could determine whether this two-dimensional cluster structure can be applied universally or is culture-specific. Moreover, if such a parsimonious structure proves to be universal, future research could be devoted to a detailed description of the complexity of these grading styles, examining the extent to which weights assigned to the different rules vary within and between grading styles (clusters) across different socio-cultural personal or collective social categories

Table 5 Hierarchical binary logistic regression for mediation variables predicting grading style. Model 1

Model 2

Model 3

B

SE

Odds ratio

B

SE

Odds ratio

Step 1: Background variables: Gender Educational level Seniority

.24 .55 .01

.32 .28 .02

1.27 1.72 1.00

.20 .54 .01

.32 .29 .02

1.22 1.71 1.01

Step 2: Mediation variables: Open/flexible perception Closed/hierarchical perception

e e

e e

e e

e e

e e

e e

.87* .80*

.38 .38

2.39* 2.23* .05

Step 3: Disciplinary expertise: Language Science Nagelkerke R2 *p < .05; **p < .01; ***p < .001.

.02

B

Model 4 SE

Odds ratio

.33 .30 .02

1.38 1.64 1.00

1.27*** .85**

.35 .31

3.57*** .43**

e e

e e

e e .12

.32 .40 .00

B .27 .44 .01 1.20*** .87** .64 .22

SE

Odds ratio

.33 .30 .02

1.30 1.56 1.01

.37 .33

3.31*** .42**

.39 .41

1.90 1.25 .13

838

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

(e.g., gender, countries). It is worth noting that cluster analysis, as used in this study, will enable the systematic comparison of these averaged combinations of distribution rules between sub-groups. Such an approach to teachers’ grading styles goes beyond existent analyses, which tend to focus on global averages obtained from separate examination of the various distribution rules. In our study, teachers’ preference for either grading style was found to be related to their disciplinary expertise: Mathematics teachers favored the performance-output grading style more, while science and language teachers were more likely to prefer the effort-input grading style. More significant is the finding that perception of subject matter mediated these associations. In regard to the open/flexible perception, complete mediation was found: both science and language teachers perceived their subject matter as more open/flexible than mathematics teachers, which affected their stronger tendency for the effort-input grading style. In the case of closed/hierarchical perceptions, only partial mediation was observed: mathematics teachers perceived their subject matter as more closed/hierarchical and favored the performance-output style more strongly than science teachers, but not language teachers. The lack of difference found between language and mathematics teachers in their perception of the subject matter as closed/ hierarchical may stem from the composition of the language teachers’ sample, which included both literature and grammar teachers. Teaching grammatical principles may seem similar to the teaching of mathematical procedures, thus enhancing the tendency to hold a closed/hierarchical perception of the subject matter. This should be further investigated in larger samples of teachers that include a wider variety of major academic disciplines and sub-disciplines. The above findings also challenge the conventional dichotomy between the “scientific-hard” disciplines, such as mathematics/science and the more “humanistic-soft” disciplines, such as language and art (Becher, 1989; Pollio, 1996). Science teachers seem to be more similar to their language-teaching colleagues and different from mathematic teachers in terms of both grading style and perception of the discipline. A plausible interpretation for this finding lies in the radical changes that have occurred in past decades in the science discipline in general and in the Israeli context in particular (Stofflett & Stoddart, 1993) e from a “hard-pure” discipline (Becher, 1989; Biglan, 1973) that focuses on fact verification and absolute truth to an applied discipline that is being studied and taught as an interdisciplinary domain of knowledge, open to the introduction of “relevant” and “practical” topics into the curriculum. Specifically, the Israeli science curriculum currently stresses active inquiry, teamwork and out-of-class learning (LevyNahum et al., 2010; Sadeh & Zion, 2009). Moreover, in some of the science disciplines (e.g., biology, bio-technology), evaluation procedures and methods (e.g., lab reports, projects and portfolios) reflect students’ efforts, and final exams in high school are written by the school teachers themselves rather than an external authority. Science has thus become a dynamic discipline which is constantly changing and being supported by new technologies (e.g., computers, the Internet). Accordingly, young students rapidly become proficient in these technologies, often even better than their teachers, so that the discipline requires flexible pedagogical practices. In contrast, the structure of the mathematics discipline continues to be considered as more hierarchical and sequential, static and stable over time. As a result, mathematics teachers preserve more conservative practices that emphasize drills and linear progress in learning as a condition for success. In such a hierarchical discipline, tracking and ability grouping are usually favored by teachers (Dar, 1985; Siskin, 1994). It is worth noting that gender had a significant effect on grading style. The finding that women perceived their subject matter as more closed than men does not fit the common view that women attribute higher significance to democratic values in the classroom

and are more supportive of student-oriented practices (Kesici, 2008). Future research should delve into this issue. In sum, our study highlights the importance of disciplinary culture in structuring teachers’ roles and pedagogical practices in general, and their grading practices in particular. Rather than defining a discipline as a “uniform,” “objective” body of knowledge, we suggest that these bodies of knowledge are also actively and subjectively constructed within and across cultures (see, e.g., Kuhn, 1970; Mannheim, 1991), either through socialization and training of teachers in their specific discipline, or in their actual practices in school and interactions with departmental colleagues. Thus, by answering a few questions about the factors affecting grade distribution, new ones arise that should be addressed in future research. Especially important is the need to widen the scope of disciplinary culture to include additional dimensions in its definition and measurement, investigating the mediating effect of this construct in the process of grading. Moreover, as cultural background has been found to affect grading styles and perceptions of the subject matter, future studies examining the role of disciplines in shaping grading practices should focus on the degree to which the mediatory role found in this study is universal.

Acknowledgement We thank the anonymous reviewers for their insightful comments and Helene Hogri for her invaluable editorial help.

Appendix A. Final factor loadings of the principal component analysis following varimax rotation on the measure of subject matter perception. Items

Subject matter perception Open/flexible (Factor 1)

The subject matter is based on systematic ordinal thinking. The subject matter deals with daily situations. Success in the subject matter requires remembering concepts and processes. Success in the subject matter requires drilling the material. The subject matter promotes creative thinking. The subject matter helps find solutions to daily problems. The subject matter furthers logical thinking. The subject matter enables critical evaluation of the material. The subject matter encourages verbal and written communication. Curricular planning for the subject matter is based on an external examination system. Curricular planning for the subject matter is based on covering all the material. Curricular planning for the subject matter is based on the teaching material's relevance to students. Teaching plan of the subject matter is based on state curriculum. The subject matter's teaching strategies involve memorizing and instilling information. The subject matter's teaching strategies enable the creation of connections to other types of knowledge. Eigenvalue (total variance explained by each factor) Mean (SD) Variance (%) explained after varimax rotation Total variance (%) explained

Closed/ hierarchical (Factor 2) .51

.64 .56 .66 .58 .69 .44 .67 .62 .65 .71 .48

.58 .60 .62

2.61

2.98

3.06 (.46) 3.25 (.47) 18.14% 19.11% 37.25%

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

References Aldridge, J. M., Fraser, B. J., & Huang, T. I. (1999). Investigating classroom environments in Taiwan and Australia with multiple research methods. Journal of Educational Research, 93, 48e61. Alpert, B., & Bechar, S. (2008). School organizational efforts in search for alternatives to ability grouping. Teacher and Teaching Education, 24, 1599e1612. Al-Sadan, I. A. (2000). Educational assessment in Saudi Arabian schools. Assessment in Education: Principles, Policy and Practice, 7(1), 143e155. Atkin, J. M., Black, P., & Coffey, J. (2001). Classroom assessment and the national science education standards. Washington, DC: National Academy Press. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173e1182. Becher, T. (1989). Academic tribes and territories: Intellectual enquiry and the cultures of disciplines. Buckingham: Open University Press. Bell, D. (1977). On meritocracy and equality. In J. Karabel, & A. H. Halsey (Eds.), Power and ideology in education (pp. 607e635). New York: Oxford University Press. Bernstein, B. (1971). On the classification of educational knowledge. In M. F. Young (Ed.), Knowledge and control: New directions for the sociology of education (pp. 47e69). New York: Collier Macmillan. Biglan, A. (1973). The characteristics of subject matters in different academic areas. Journal of Applied Psychology, 57, 195e203. Brookhart, S. (2004). Grading. Upper Saddle River. NJ: Pearson Education. Brown, J. D. (2001). Using surveys in language programs. Cambridge: Cambridge University Press. Brown, G. T. L., Kennedy, K. J., Fok, P. K., Kin Sang, C. J., & Ming, Y. W. (2009). Assessment for student improvement: understanding Hong Kong teachers’ conceptions and practices of assessment. Assessment in Education: Principles, Policy & Practice, 16(3), 347e363. Carbonaro, W. (2005). Tracking, students’ effort and academic achievement. Sociology of Education, 78, 27e49. Cohen, E. (1989). The changing legitimations of the state of Israel. Studies in Contemporary Jewry, 5, 148e165. Conley, P. (1996). Local justice in the distribution of college admissions: a statistical study of beliefs versus practice. Social Justice Research, 9(3), 239e258. Costello, A. B., & Osborne, J. W. (2005). Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis. Practical Assessment, Research & Evaluation, 10(7). Available online http://pareonline.net/getvn.asp? v¼10&n¼7. Crooks, T. J. (2002). Educational assessment in New Zealand schools. Assessment in Education, 9(2), 237e253. Cunningham, A. E., Zibulsky, J., & Callahan, M. D. (2009). Starting small: building preschool teacher knowledge that supports early literacy development. Reading and Writing: An Interdisciplinary Journal, 22, 487e510. Dar, Y. (1985). Teachers’ attitudes toward ability grouping: educational considerations and social and organizational influences. Interchange, 16, 17e38. DeBoer, B. V., Anderson, D. M., & Elfessi, A. M. (2007). Grading styles and instructor attitudes. College Teaching, 55(2), 57e65. Deutsch, M. (1985). Distributive justice. New Haven, CT: Yale University Press. Doran, R. L., Lawrenz, F., & Hegelson, S. (1994). Research on assessment in science. In D. Gabel (Ed.), Handbook of research of science teaching and learning (pp. 389e442). New York: Macmillan. Duncan, C. R., & Noonan, B. (2007). Factors affecting teachers’ grading and assessment practices. Alberta Journal of Educational Research, 53(1), 1e21. Entwisle, D. R., Alexander, K. L., & Olson, L. S. (2007). Early schooling: the handicap of being poor and male. Sociology of Education, 80, 114e138. Everitt, B. S., Landau, S., & Lesse, M. (2001). Cluster analysis. London: Arnold. Grossman, P. L., & Stodolsky, S. S. (1995). Content as context: the role of school subjects in secondary school teaching. Educational Researcher, 24(8), 5e11. Haddock, C. K., Rindskopf, D., & Shadish, W. R. (1998). Using odds ratio as effect sizes for meta-analysis of dichotomous data: a primer on methods and issues. Psychological Methods, 3(3), 339e353. Hardy, M. A. (1993). Regression with dummy variables. London: Sage Publishers. Hegtvedt, K. A. (1992). Bargaining for justice: the resolution of competing justice claims. Social Justice Research, 5, 155e172. Hurn, C. (1985). The Limits and possibilities of schooling: An introduction to the sociology of education. Boston: Allyn and Bacon. Hutchison, C. B., & Bailey, L. M. (2006). Cross-cultural perception of assessment of selected international science teachers in American high schools. Cultural Studies of Science Education, 1, 657e680. Israeli Central Bureau of Statistics. (2002). Teaching staff survey. Jerusalem: Central Bureau of Statistics. (Hebrew). Kelly, S. (2008). What types of students’ effort are rewarded with high marks? Sociology of Education, 81, 32e52. Kerlinger, F. N. (1958). Progressivism and traditionalism: basic factors of educational attitudes. Journal of Psychology, 48, 111e135. Kesici, S. (2008). Democratic teacher beliefs according to the teacher’s gender and locus of control. Journal of Instructional Psychology, 35(1), 62e92. Kuhn, T. (1970). The structure of scientific revolutions. Chicago: University of Chicago Press. Lee, I. (2008). Understanding teachers’ written feedback practices in Hong Kong secondary classrooms. Journal of Second Language Writing, 17, 69e85.

839

Leventhal, G. S. (1976). Fairness in social relations. In J. Thibaut, T. Spence, & R. C. Carson (Eds.), Contemporary topics in social psychology (pp. 211e239). Morristown, NJ: General Learning Press. Levy-Nahum, T., Ben-Chaim, D., Azaiza, I., Herskovitz, O., & Zoller, U. (2010). Does STES-oriented science education promote 10th-grade students’ decisionmaking capability? International Journal of Science Education, 32(10), 1315e1336. Linn, R., & Miller, M. (2005). Measurement and assessment in teaching. Upper Saddle River, NJ: Pearson Prentice Hall. Loewenberg-Ball, D., Thames, M. H., & Phelps, G. (2008). Content knowledge for teaching: what makes it special? Journal of Teacher Education, 59(5), 389e407. MacKinnon, D. P., & Dwyer, J. H. (1993). Estimating mediated effects in prevention studies. Evaluation Review, 17(2), 144e158. Mannheim, K. (1991). Ideology and utopia. London: Routledge. McCargar, D. F. (1993). Teacher and student role expectations: cross-cultural differences and implications. The Modern Language Journal, 77(2), 192e207. McLaughlin, M. W., Talbert, J. E., & Bascia, N. (Eds.). (1990). The contexts of teaching in secondary schools: teachers’ realities. New York and London: Teacher College, Columbia University. McMillan, H. J. (2001). Secondary teachers’ classroom assessment and grading practices. Educational Measurement: Issues and Practice, 20(1), 20e32. Miller, S. (1998). Shortcut: high school grades as a signal of human capital. Educational Evaluation and Policy Analysis, 20, 299e311. Nevo, D., Wolfe, Y., & Goldblatt, T. A. (1988). Concepts of evaluation and school basedassessment in active schools. Research report. Ministry of Education and Tel Aviv University, Department of Education. (Hebrew). Nisan, M. (1985). Justice in the classroom: teachers’ and students’ perceptions of principles that guide grading distribution. In Z. Lam (Ed.), School and education (pp. 141e155). Jerusalem: Magnes, (Hebrew). Palm, T. (2008). Performance assessment and authentic assessment: a conceptual analysis of the literature. Practical Assessment Research & Evaluation, 13(4), 1e11. Pollio, H. R. (1996). The two cultures of pedagogy: teaching and learning in the natural sciences and the humanities. Teaching and Learning Issues, 75, 3e33. Rajbhandari, P., & Wilmut, J. (2000). Assessment in Nepal. Assessment in Education, 7(2), 255e269. Randall, J., & Engelhard, G. (2010). Examining the grading practices of teachers. Teaching and Teacher Education, 30, 1e9. Resh, N. (1998). Track placement: How the “sorting machine” works in Israel. American Journal of Education, 106, 416e438. Resh, N. (2009). Justice in grades allocation: Teachers’ perspective. Social Psychology of Education: An International Journal, 12(3), 315e325. Robinson, R. V., & Bell, W. (1978). Equality, success and social justice in England and the United States. American Sociological Review, 43, 125e144. Roscigno, V. J., & Ainsworth-Darnell, J. W. (1999). Race, cultural capital, and educational resources: persistent inequalities and achievement returns. Sociology of Education, 72(3), 158e178. Sabbagh, C., Cohen, E. H., & Levy, S. (2003). Styles of social justice judgments as portrayed by partial-order scalogram analysis. Acta Sociologica, 46(4), 323e338. Sabbagh, C., Faher-Aladeen, R., & Resh, N. (2004). Evaluation of grade distributions: A comparison of Druze and Jewish pupils in Israel. Social Psychology of Education, 7(3), 313e337. Sabbagh, C., Resh, N., Mor, M., & Vanhuysse, P. (2006). Spheres of justice within schools: Reflections and evidence on the distribution of educational goods. Social Psychology of Education, 9, 97e118. Sabbagh, C., Biberman-Shalev, L., & Resh, N. (2009). Teachers’ evaluation styles when distributing grades: Do individual-status variables matter? In M. Ortiz, & C. Rubio (Eds.), Educational Evaluation: 21st Century Issues and Challenges (pp. 333e348) New York: Nova Science Publishers. Sadeh, I., & Zion, M. (2009). The development of dynamic inquiry performances within an open inquiry setting: a comparison to guided inquiry setting. Journal of Research in Science Education, 46(10), 1137e1160. Shulman, L. (1986). Those who understand: knowledge growth in teaching. Educational Researcher, 15(2), 4e14. Siskin, L. S. (1994). Realms of knowledge: Academic departments in secondary school. Washington, DC: Falmer Press. Smith, J. P. (1996). Efficacy and teaching mathematics by telling: a challenge for reform. Journal for Research in Mathematics Education, 27(4), 387e402. Snow, C. E., Griffin, P., & Burns, M. S. (2005). Knowledge to support the teaching of reading: A model of professional growth in reading education. San Francisco: Jossey-Bass. Stodolsky, S. S., & Grossman, P. A. (1995). The impact of subject matter on curricular activity: an analysis of five academic subjects. American Educational Research Journal, 32, 227e249. Stofflett, R. T., & Stoddart, T. (1993). The ability to understand and use conceptual change pedagogy as a function of prior content learning experience. Journal of Research in Science Teaching, 31(1), 31e51. Tata, J. (1999). Grade distributions, grading procedures and students’ evaluations of instructors: a justice perspective. Journal of Psychology, 133(3), 263e271. Taylor, M., & Otinsky, G. (2007). Becoming whole language teachers and social justice agents: pre-service teachers inquire with sixth graders. International Journal of Progressive Education, 3(2), 1e16. Thompson, B. (2004). Exploratory and confirmatory factor analysis. Washington: American Psychological Association. Thorkildsen, T. A. (1989). Justice in the classroom: the students’ view. Child Development, 60, 323e334. Trelstad, M. (2008). The ethics of effective teaching: challenges from the religious right and critical pedagogy. Teaching, Theology and Religion, 11(4), 191e202.

840

L. Biberman-Shalev et al. / Teaching and Teacher Education 27 (2011) 831e840

Tyack, D., & Tobin, W. (1994). The “grammar” of schooling: why has it been so hard to change? American Educational Research Journal, 31(3e4), 453e479. Van-Driel, J. H., Bulte, A. M. W., & Verloop, N. (2005). The conceptions of chemistry teachers about teaching and learning in the context of a curriculum innovation. International Journal of Science Education, 27(3), 303e322. Vanfossen, B., Jones, J., & Spade, J. (1987). Curriculum tracking and status maintenance. Sociology of Education, 43, 355e376.

Wagner, D. G. (1995). Gender differences in reward preference: a status-based account. Small Group Research, 26(3), 353e371. Walzer, M. (1983). Spheres of justice. New York: Basic Books. Weisberg, H. F. (1992). Central tendency and variability. Newbury Park, CA: Sage. Ylijoki, O. H. (2000). Disciplinary cultures and the moral order of studying: a case study of four Finnish university departments. Higher Education, 39, 339e362.