Crossroads of Science and Religion. Acceptance of Evolution in Ten Post-Communist European Societies
Cosima Rughiniş University of Bucharest
Please do not cite without permission! Submitted for publication on September 9, 2009
Abstract: The article analyzes public acceptance of evolution in ten post-communist European countries. A structural equation model is used to estimate influences of scientific literacy, religiosity, education, age, and gender, using data from the 2002 Eurobarometer 55.2 “New Europeans, Science and Technology”. Results indicate that scientific literacy is the most powerful predictor for acceptance of evolution, significant in seven out of ten countries, but its influence is weaker for religious people. Church attendance is only significant in four countries. Education has a mediated effect on acceptance of evolution, via scientific literacy, but it does not have a direct effect.
Key words: scientific literacy, evolution, creationism, Eurobarometer
1
Public acceptance of evolution, divine creation and creationism has received relatively little attention in European survey research. The topic comes at the intersection of religion studies and science studies, in particular the Public Understanding of Science paradigm, and it occupies a marginal position in both. It has attracted considerable more attention in the United States of America, especially in education studies – probably as a result of intense public debates on teaching evolution and creationism in US schools. In this paper I use the 2002 Eurobarometer 55.2 data in order to explore the factors that influence public acceptance of evolution. Starting from a review of previous research, I propose and test a structural equation model that investigates the influence of scientific literacy, education, religious denomination and church attendance on acceptance of evolution, in ten post-communist European countries: Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia and Slovenia.
Previous findings There is considerable literature on acceptance of evolution and creationism. I follow Fulljames et. al (1991, 173) in distinguishing between acceptance of creationism, understood as a belief in the literal truth of religious genesis texts and rejection of evolutionary theories, and acceptance of the divine creation of the world, which may be compatible with evolutionary theories. A proof of this compatibility is offered by the acceptance of the Roman Catholic Church of the scientific legitimacy of evolution, declared by Pope John Paul II “more than just a hypothesis” and compatible with Christian faith (Mazur 2004, 57). I will also follow Deniz et al. (2008, 422) in using “acceptance of evolution” rather than “belief in evolution”, since “belief” covers common-sense, scientific and religious knowledge as well, blurring important differences 2
between them. A person’s “acceptance of evolution” thus refers to his or her agreement with the main tenets of evolutionary theories, justified by current scientific support in favor of evolution, and by his / her acknowledgement of scientific performance in describing and explaining the world. Previous research has focused on three main topics in the study of evolution and creationism: building measurement instruments, identifying worldview variables conceptually and empirically linked to this acceptance, and the research of broader sociodemographic correlates such as education, religiosity and age.
Measurement for acceptance of evolution and creationism Eve and Dunn (1990) analyze acceptance of creationism and pseudo-scientific beliefs among high school science teachers in the US. They conduct a factor analysis of a series of items and they identify a common factor of “Biblical literalism”, including items concerning creationism and religious doctrine; other items, concerning various pseudoscientific beliefs, failed to aggregate in a factor structure. The “Biblical literalism” factor includes the following items (idem, 16): 1. Satan is a real personality working in the world today. 2. Regardless of the fact that the earth may be extremely old, the fossil record does not provide persuasive evidence that there has been any significant evolutionary change through time. 3. Heaven really exists. 4. Black magic exists and has real power. 5. Each person has a soul that survives after his or her physical death. 6. One can believe in the Bible and creation, or in atheistic evolution; there is 3
really no middle ground. 7. Dinosaurs and humans lived at the same time. 8. Adam and Eve were the first human beings and were created by God. 9. The Bible's account of creation should be taught in public schools as an explanation of origins. 10. The Bible is an authoritative and reliable source of information with respect to such scientific issues as the age of the earth and the origin of life. 11. There is a good deal of scientific evidence against evolution and in favor of the Bible's account of creation. The authors could not identify strong demographic correlates of acceptance of Biblical literalism within the population of science teachers. Significant correlations were found with the type of college or university where teachers have received their final degree, since graduates of private religious institutions were more likely to endorse Biblical literalism than graduates of other private or public institutions (idem, 17). Also, respondents who have attended science courses during college were less likely to support this views, while respondents who have attended Bible study courses were more likely to support it (idem, 17). Francis and Greer (1999) review previous research on creationism, focusing on measurement instruments. They conclude that most research uses single-item measurement of creationism, and they propose and test on a sample of pupils a scale that measures acceptance of evolution. The scale (idem, 96) includes the following items: 1. I believe in the scientific view of the origins of the world. 2. God created the world as described in the Bible.
4
3. Everything in the world was made by natural forces. 4. The world was made by God in six days each of 24 hours. 5. God created the universe, including living creatures, out of nothing. 6. I accept the idea of evolution creating everything over millions of years. 7. God formed man out of the dust of the earth. 8. God made woman out of man’s rib. 9. God rested on the seventh day after he had finished his work of creation. 10. Science disproves the biblical account of creation. 11. Scientists have discovered how the world was made. Another dedicated scale is the Measure of Acceptance of the Theory of Evolution – MATE scale, developed by Rutledge and Warden (1999, apud Deniz et. al. 2008). The instrument includes items such as: ‘‘Evolution is a scientifically valid theory’’, ‘‘Organisms existing today are the result of evolutionary processes that have occurred over millions of years’’, and ‘‘The theory of evolution brings meaning to the diverse characteristics and behaviors observed in living things.’’ Deniz, Donnelly and Yilmaz (2008) use this instrument in order to identify the conceptual ecology of acceptance of evolution among college biology students. They explore the relationships between acceptance of evolution and broader epistemological beliefs, understanding of evolution and thinking dispositions – in particular open-mindedness, controlling for parents’ level of education and students’ year in the educational program. Still, the authors do not investigate the influence of students’ religiosity. Fulljames, Gibson and Francis (1991) study adolescent pupils’ attitude towards Christianity and its relationships with two variables: “scientism”, defined as “the view
5
that scientific methods and scientific theories can attain to absolute truth” (p. 173), and the perception of Christianity as necessarily involving creationism. While the authors do not investigate acceptance of evolution or creationism, these variables could also be useful in this respect.
Influence of scientific literacy, education and religiosity Mazur (2004) focuses on belief in evolution in the US, using data from the General Social Surveys of 1993, 1994 and 2000. Mazur investigates the relationship between acceptance of evolution and scientific literacy, education and religiosity, using Pearson correlations and multiple regression models. In order to measure belief in evolution, the author uses the GSS item that requires respondents to answer how true is the statement: “Human beings evolved from earlier species of animals” (“definitely true”, “probably true”, “probably not true”, and “definitely not true”). The 1993 and 1994 GSS also includes four other items of scientific literacy: whether it was true or not that “all radioactivity was made by humans” (not true); that “antibiotics killed bacteria, not viruses” (true); that “astrology had some scientific truth” (not true); and that “all manmade chemicals can cause cancer if eaten in sufficient quantity” (not true) (Mazur 2004, 56). The author uses three religiosity indicators: perception of the Bible as: fable, inspired, or God’s word; the degree of religious fundamentalism of respondent’s denomination, and church attendance. Standardized coefficients of each religiosity indicator were stronger than the influence of education, leading the author to conclude that “Religious belief is the primary correlate - and may be the chief determinant - of Americans' attitudes toward the life-sciences claim that humans evolved from earlier species” (idem, 60). Given the low Pearson correlations (R < 0.15) between the evolution 6
item and the other scientific literacy items, Mazur concludes that the influence of education on acceptance of evolution “cannot confidently be attributed to improved scientific literacy alone” (idem, 60) and it probably reflects other mediating factor. Mazur highlights the finding that political views are significantly associated with acceptance of evolution, even after controlling for education, religiosity and other sociodemographic variables. We may see in this association an indicator of the intense public and political debate around creationism and evolution in the United States. Political variables are also used by Woodrum and Hoban (1992), who analyze support for teaching creationism in schools, in the US. The main predictors are belief in biblical literalism, religiosity, political activism, education and a conservative orientation; biological knowledge does not have a direct effect on support for creationism teaching, after controlling for the other variables. Pardo and Calvo (2008, 14) have investigated a similar topic, insofar it is shaped by both science and religion: the moral status of the embryo, studied as predictor for attitudes towards embryo research. Findings from a survey of 9 West- and CentralEuropean countries indicate that biological literacy is a significant predictor of the moral status of the embryo. Also, Italy and Poland displayed a particular profile, with a majority of respondents believing in the human status of the embryo, thus indicating the influence of the Catholic Church (idem, 32-33). Brossard et.al. (2008) identifies a moderating role of religiosity in the influence of scientific knowledge on support for nanotechnology. Specifically, the influence of knowledge is significantly stronger for less religious respondents. The authors conclude that religiosity acts as an “interpretive tool” (idem, 10) for audiences trying to make sense
7
of nanotechnology. We may summarize previous research by highlighting the factors that influence individual acceptance of evolution, creation and / or creationism. The main forces that compete in shaping individual views on the creation of Earth, life and humans are science and religion. Acceptance of evolution is shaped by a person’s knowledge and understanding of science, since evolution is a scientific theory. At the same time, acceptance of creation and / or creationism is shaped by individual religious beliefs and relationships with religious communities. Religiosity may also act as a moderator for the influence of scientific literacy. A person’s acceptance of evolution is also shaped by his or her understanding of the relationship between evolution and creation/ism, and more generally the relationship between science and religion. The relationship between the two may be more or less salient for a given person, depending on local conditions and on individual background and interests: some people may not be aware of conflicts between science and religion, including the one between evolution and creationism, while others may take active part in public debates. Also, the nature of this relationship may be differently perceived: for example, some people may consider creationism to be endorsed by the scientific community, while others may consider it un-scientific. Biblical literalism also seems to be a relevant variable for understanding relationships between religion and science. Acceptance of evolution and creation/ism also depends on factors at community and social level – such as the information from school curricula, public circulation of scientific information, discourses and debates on evolution and creationism in the public sphere and in local religious organizations and communities. For example, Park (2001)
8
argues that in the United States, starting with the 60’s, public debates and especially campus debates have had a significant contribution into legitimizing creationism assertions at the expense of evolution. Mazen (2004) indicates that political affiliation and the type of religious community significantly shape individual belief in evolution. Education is expected to be associated with acceptance of evolution, since school is the first significant contact of young people with science and scientific findings. Given the broad social processes of differentiation between generations and gender categories, we may hypothesize that age and gender are associated with scientific literacy, religiosity, and indirectly with acceptance of evolution. For example, Mazur reports a significant positive effect of education and a significant negative effect of age on beliefs in evolution, after controlling for three indicators of religiosity, for political orientation and for rural/urban environment – but not for gender or scientific literacy (2004, 59). Also, Hayes (2001, 664) reports a significant influence of education, gender and age on scientific literacy after controlling for manual occupation, marital status, religious denomination and belief, and political ideology, in United States, Japan and four European countries.
The factual knowledge of science scale The debate between evolution and creationism has received less public coverage and political weight in Europe, including the post-communist countries, than in the United States. Therefore, I aim to analyze the relationship between acceptance of evolution and scientific literacy and education, on one hand, and religiosity, on the other hand, using available data from the 2002 Eurobarometer 55.2 “New Europeans, Science and
9
Technology”1. The Eurobarometer includes two indicators of acceptance of evolution as part of the Factual Knowledge of Science Scale, also used by the National Science Foundation, consisting of several items that measure public knowledge of general scientific facts. In the 2002 Eurobarometer the following items are included (of which items 9-7 and 9-11 are potentially useful measurements of acceptance of evolution):
Here is a little quiz. For each of the following statements, please tell me if you think it is true or false. If you don’t know, say so, and we will go on to the next one (SHOW CARD) (1. True, 2. False, 8. Don’t know, 9. Refusal) Q9-1. The centre of the Earth is very hot Q9-2. The oxygen we breathe comes from plants Q9-3. Radioactive milk can be made safe by boiling it Q9-4. Electrons are smaller than atoms Q9-5. The continents on which we live have been moving for millions of years and will continue to move in the future Q9-6. It is the father’s genes that decide whether the baby is a boy or a girl Q9-7. The earliest humans lived at the same time as the dinosaurs Q9-8. Antibiotics kill viruses as well as bacteria Q9-9. Lasers work by focusing sound waves Q9-10. All radioactivity is man-made Q9-11. Human beings, as we know them today, developed from earlier species of
1
Data is available online in the GESIS – ZACAT library (ZA4325 dataset): http://zacat.gesis.org/webview/index.jsp
10
animals Q9-12. The Sun goes around the Earth Q9-13. It takes 1 month for the Earth to go around the Sun
The same items (with the exception of item Q9-6, which was reformulated to refer mother’s genes) were also included in the 2005 Eurobarometer 63.1. I have chosen to work with the 2002 Eurobarometer because it includes information on respondents’ educational achievement, while the 2005 Eurobarometer only includes information on the age of the respondent when finishing full-time education. The Factual Knowledge of Science scale has been used by the National Science Foundation in the US since 1979, in a longer form, including 16 closed-answer questions (including all Eurobarometer items, except for 9-7) and two open-ended questions (Bann and Schwerin 2004). An equivalent shorter scale, tested by Bann and Schwerin, includes 11 items distributed across the same domains: life science / biology (items 9-6 and 9-8), evolution / history of Earth (items 9-11 and an additional item: “Universe began with explosion”, which is not included in the Eurobarometer); geology / geography (items 9-1 and 9-5); astronomy (items 9-12 and 9-13), radioactivity (item 9-10), physics (item 9-9), and chemistry (item 9-4). In comparison with the short scale, the Eurobarometer scale includes an additional item for biology (item 9-2), an additional item for radioactivity (item 9-3) and an additional item for evolution / history of Earth (item 9-7, which was not part of the NSF items). Bann and Schwerin have concluded that items 9-2, 9-3 and 9-6 have low factor loadings and low discrimination power, but decided to maintain item 9-6 in the short scale because it is the only one on which women generally score higher than
11
men, and removing it might affect equivalence with the long form (idem, 15). The factual knowledge of science scale has received detailed analysis, especially in relation to attitudes towards science (see for example Allum at al. 2008, Bauer et. al. 2007, Miller 2004, Pardo and Calvo 2004). Nevertheless, the topic of creationism has not been studied it its own, other than to pinpoint the deviation of item Q9-11 from the general scale due to the influence of religiosity (Pardo and Calvo 2004, 209). My analysis follows to a certain extent Mazur (2004), sharing a similar dependent variable (Q9-11) and several predictors. There are some differences in item formulation, and in methods. While the item on evolution was formulated on a four-level scale in the GSS, it is formulated as True/False/Don’t know in the Eurobarometer. Mazur does not report treatment for the “Don’t know” (DK) answers. Because of high rates of DK response (for example, for item Q9-11 they reach 14-29% in the 2002 Eurobarometer 55.2), I have chosen to recode them into the same category as false response, instead of considering them missing values. Two of the scientific literacy items used by Mazur are similar with items from the Eurobarometer, with slight differences in formulation and scale: the item on radioactivity and the one on antibiotics. Mazur also uses respondent education and church attendance as predictors, controlling for age and urban / rural environment, among others. The main difference refers to data analysis methods. While Mazur uses multiple regression models, I have chosen to use a structural equation model, in order to test model fitness for the scientific literacy construct and to estimate the direct effects of scientific literacy, education, religiosity, gender and age, and also the indirect effect of education. The structural model (Figure 5) includes relationships between acceptance of
12
evolution and the following independent variables: scientific literacy, church attendance, religious denomination, education, gender, age, and locality type. Scientific literacy and church attendance are also modeled as dependent on education, religious denomination, gender, age, and locality type. Therefore, my analysis attempts to address the following research questions: 1. How is acceptance of evolution related to scientific literacy, church attendance, religious denomination, education, gender and age, controlling for locality type? 2. How does the direct effect of church attendance compare with the direct effect of scientific literacy, the indirect effect of education via scientific literacy, and the direct effect of education? 3. Is there a moderating effect of religiosity on the influence of scientific literacy on acceptance of evolution?
Data and methods In order to use the 2002 Eurobarometer 55.2 factual knowledge indicators, I have recoded them into dichotomous variables so that the scientifically correct answer has the value “1”, the scientifically incorrect answer has the value “0”, and “Don’t know” has value “0”. Therefore, the variables in the model merge the incorrect answers and the “Don’t know” answers into a single category. Besides the 13 items of the scientific knowledge scale, I have also used the other two indicators of scientific literacy available in the dataset:
Q10. Try to imagine that a scientist wants to know if a certain drug is effective against a disease for which there is no means of prevention or treatment. Which do you 13
think is the scientifically correct approach for testing the drug’s effectiveness? 1. Administer the drug to 1000 people suffering from the disease and see how many show signs of recovery 2. Administer the drug to 1000 people suffering from the disease, tell another 500 to continue without drug treatment to see in which of the two groups there are more people showing signs of recovery 3. Administer the drug to 1000 people suffering from the disease and treat another 5000 with a totally harmless, identical-looking substance to see in which of the two groups there are more people showing signs of recovery. (correct answer) 9. Don’t know Q11. Suppose doctors tell a couple that their genetic make-up means that they’ve got a one in four chance of having a child with an inherited illness. Does this mean that…? 1. If they have only three children, none of them will have the illness 2. If their first child has the illness, the next three children will not 3. Each of their children has the same risk of having the illness (correct answer) 4. If their first three children do not have the illness, the fourth will 9. Don’t know For items Q10 and Q11 I have recoded the correct answer into 1 and all other answers, including “Don’t know”, into 0. I have used item Q9-11, recoded into a dichotomous variable, as indicator of acceptance of evolution. Although item Q9-7 is also compatible with a creationist viewpoint and incompatible with an evolutionary theory, the two items are weakly
14
correlated or not at all, in 2002 as well as 2005 (Table 1). Therefore, I have only used the more explicit question Q9-11 to measure acceptance of evolution, and I have included Q9-7 in the scientific literacy scale, to replace item Q9-11 for the “Evolution / History of Earth” domain. (Table 1 around here) Acceptance of evolution is used as a dependent variable in a structural equation model in which it is regressed on scientific literacy, religious denomination, church attendance, education, age, sex, and type of locality. Independent observed variables are presented in Table 2, and the unobserved variable of scientific literacy is discussed in the measurement model section.
(Table 2 around here)
(Table 3 around here) The measurement and structural models have been estimated using Mplus v.5. Since the dependent variable Q9-11 and all indicators for scientific literacy are dichotomous variables, model estimates are based on the WLSMV estimator (Muthén and Muthén 2007, 249)2.
Findings In Figure 1, Figure 2 and Figure 3 we can see the cross-country distributions for three
2
Complete Mplus output results for measurement and structural model estimates are available upon request.
15
indicators of factual knowledge of science, including acceptance of evolution. We can see that there is a considerable proportion of “Don’t know” answers for all items – including the apparently obvious “The Sun goes around the Earth”. For example, in Romania 14% of respondents answered that they do not know, and an additional 29% answered that the assertion is true. (Figure 1 around here) Proportions of “Don’t know” answers are higher for item Q9-7, ranging from a minimum of 19% in Estonia to 42% in Romania. These high frequencies indicate that “Don’t know” is a socially significant answer which cannot be treated as a missing value, for substantial reasons and for methodological reasons as well. Listwise deletion of all respondents with “don’t know” or system missing answers for items in the scientific literacy scale leads to the exclusion of 73% of the total number of 10, 247 cases from the ten countries.
(Figure 2 around here) As regards item Q9-11, the proportion of respondents who accept evolution ranges from 40% in Lithuania to 67% in Slovenia, while the proportion of respondents who reject evolution, thus implicitly adhering to a creationist view, ranges from 19% in Slovenia to 39% in Latvia. Still, for all positively formulated items such as Q9-11, we should take into account that a certain proportion of affirmative, thus correct, answers may be due to acquiescence bias.
(Figure 3 around here) 16
Measurement model The fit indicators for the final measurement model are presented in Table 4. With the exception of Estonia and Lithuania, for all models the CFI and TLI values are equal or higher than 0.95, and for all models the RMSEA is lower than 0.05, indicating an overall good model fit (Hu and Bentler, 1999). In all countries, all or almost all indicators for scientific literacy have standardized loadings larger than 0.400. In order to improve model fit, I have removed items Q9-2, Q9-6, and Q10, which had low loadings on the literacy factor, and I have added correlations between the error terms of Q9-4 and Q9-5, Q9-12 and Q9-13, and Q9-7 and Q-13. The first two had modification indices higher than 10 for several countries, while the third considerably improved model fit for Lithuania. The correlation of Q9-4 and Q9-5 probably appears because both are positively formulated items, while all others, except Q9-1, are negatively formulated items. The correlation between Q9-12 and Q9-13 may be due to the fact that both items refer to the Sun and the Earth, thus tapping into the same knowledge reservoir. Since items Q9-7 and Q9-13 correlate strongly only in Lithuania, the explanation is probably context-specific.
(Figure 4 around here)
Overall, the scientific literacy scale that I will use is similar with the shorter form of the Factual Knowledge of Science scale proposed by Bann and Schwering (2004, 17), with the following changes:
17
1. Item 9-11 is replaced with item 9-7, because item 9-11 is used as a dependent variable in the structural model; 2. Item 9-6 has been eliminated to improve model fit, since equivalence with the longer NSF scale was not relevant for the study; 3. There are two items for radioactivity: 9-3 and 9-10, since both of them have displayed high factor loadings; 4. The scale also includes item Q11, measuring respondents’ understanding of the concept of probability (Bann and Schwerin 2004, 8).
Structural model Indicators of structural model fit for all countries are presented in Table 4. With the exception of Lithuania, Slovakia and Romania, for all countries the values of CFI and TLI are equal or higher than 0.95, and for all models the RMSEA is lower than 0.05, indicating an overall good model fit (Hu and Bentler 1999). Standardized results of structural models are presented in Table 5. Scientific literacy is the strongest predictor for acceptance of evolution, and it is positively and significantly associated with the dependent variable for 7 out of 10 countries. In all countries, scientific literacy is positively related to education and negatively to age, and men display higher scores than women. There seems to be no systematic direct effect of education on acceptance of evolution. With the exception of Slovenia, which has an unexpected negative direct influence of education, in all other countries the influence of education on the dependent variable is totally mediated by scientific literacy.
18
(Figure 5 around here)
Church attendance has a significant negative influence in four countries, while men and younger people have higher acceptance of evolution in three out of the ten countries. It is interesting that three out of the seven countries which have a significant Romano-Catholic population display a significant effect of church attendance on acceptance of evolution (Hungary, Poland and Slovakia); while only one (Estonia) of the four countries with a sizeable Orthodox population have this influence (Lithuania is counted for both categories). This may indicate a stronger relevance of the creationism issue for the Romano-Catholic church and societies with a Catholic influence. The direct effect of church attendance is stronger than the effect of education mediated by scientific literacy, in the three countries (Hungary, Poland, Slovakia) where both effects are present. Still, in these cases the direct effect of scientific literacy is stronger than the direct effect of church attendance.
(Table 4 around here)
(Table 5 around here)
In relation with the study of Mazur (2004), a common finding concerns the stronger influence of church attendance in relation with education. The most important divergence concerns the influence of scientific literacy. Although bivariate correlations between Q9-11 and the items on radioactivity and antibiotics are indeed absent or low in
19
the 2002 Eurobarometer as well (Table 6), the structural models indicates that in general the most important predictor of acceptance of evolution is scientific literacy, which mediates completely the influence of education. Gender and age have direct effects on acceptance of evolution only in 3 out of 10 countries.
(Table 6 around here)
In Table 7 we can see the results of the structural model without the two denomination variables (“Majority denomination” and “No denomination”), estimated for the entire sample, for the group of respondents who declare “no denomination”, and for the group of respondents who declare that they belong to a religion. Because in Romania and Poland there are too few cases of respondents who do not belong to a denomination, the models have been estimated only for the other eight countries.
(Table 7 around here)
In Czech Republic, Hungary and Slovakia we can see that the standardized coefficients for the influence of scientific literacy are higher for the non-religious people than for members of religious denominations. We can see that in these countries religiosity has a moderating relationship on the influence of scientific literacy, insofar as this influence is weaker for the religious people. Lithuania is an exception that is difficult to explain, because scientific literacy is only significant for the religious respondents, and there is a negative influence of church attendance, significant for p=0.05, for the group of non-religious respondents. In Bulgaria there seems to be no difference between the two 20
categories; it is probable that the lack of statistical significance for the group of nonreligious respondents is due to the low number of cases (117 respondents). Also, in Estonia, Hungary, Latvia and Slovakia we can see that, as expected, church influence only has an effect for the group of respondents that are members of a denomination.
Discussion The structural equation model based on the 2002 Eurobarometer 55.2 data indicates that scientific literacy is the most powerful predictor for acceptance of evolution, followed by church attendance, gender and age. There seems to be no additional direct effect of education, after controlling for scientific literacy. These conclusions are nevertheless conditioned by several limitations of the present analysis. The clearest limitation concerns the summary information on religiosity, which is only included as denomination and church attendance; the 2002 Eurobarometer does not include measures of religious beliefs, biblical literalism, or private forms of religious practice. Acceptance of evolution has also been measured with a single item, originally included in the scientific literacy scale. The second item from this domain available in the dataset cannot be used to measure acceptance of evolution, because it is not strongly correlated with the explicit item. A more complex measurement may lead to different results. Given the prominent influence of religiosity on acceptance of evolution in the United States (Mazur 2004), it is interesting that in six out of ten countries church attendance does not have a significant influence. Off all four dominantly Orthodox countries, only in Estonia is church attendance significant, with a probability of error p=0.05. This may reflect the absence of a public debate on the creationism / evolution issue, or at least the absence of an active involvement of churches, especially the 21
Orthodox churches, in such debates. For example, the issue of creationism in public education has had a relatively low prominence in Romania until nowadays; controversies have arisen around the inclusion of religion classes in schools (Stan and Turcescu 2005), and the presence of religious icons in schools. In 2007 the decision of the Ministry of Education to take out of the curriculum lessons especially dedicated to the theory of evolution have stirred some public debate (Cernea 2009a), but the issue has gradually fallen into oblivion. Recently, in March 2009, a controversy has arisen around two books used in public schools that endorsed creationist viewpoints (Cernea 2009b) – but there have been no public interventions in favor of strengthening or defending the presence of creationism in public education. At a more general level, it would be interesting to see whether there is any association between religious confession, the intensity of public debates on creationism, and the influence of church attendance on acceptance of evolution, across different European countries. Religiosity seems to have a moderating effect on the influence of scientific literacy: the association is stronger for the non-religious people. It is possible that a key variable in understanding acceptance of evolution is people’s understanding of the relationship between religion and science, including evolutionary theories, or the related variable of biblical literalism. Further research should investigate the hypothesis that scientific literacy is associated with acceptance of evolution for non-literalist religious respondents.
22
References 1. Allum, Nick, Patrick Sturgis, Dimitra Tabourazi and Ian Brunton-Smith. 2008. Science knowledge and attitudes across cultures: a meta-analysis. Public Understanding of Science 17: 35-54. 2. Bann, Carla M. and Michael J. Schwerin. 2004. Public knowledge and attitudes scale construction: development of short forms. RTI International. 3. Bauer, Martin W., Nick Allum and Steve Miller. 2007. What can we learn from 25 years of PUS survey research? Liberating and expanding the agenda. Public Understanding of Science 16:79-85 4. Brossard, Dominique, Dietram A. Scheufele, Eunkyung Kim and Bruce V. Lewenstein. 2008. Religiosity as a perceptual filter: examining processes of opinion formation about nanotechnology. Public Understanding of Science OnlineFirst, published on October 1, 2008 as doi:10.1177/0963662507087304 5. Cernea, Remus. 2009a. Învăţămîntul românesc şi evoluţia – interviu cu Remus Cernea, preşedintele Asociaţiei Umaniste Române. Dilema Veche, no. 285. 6. Cernea, Remus. 2009b. Noul Ev Mediu în învăţământul românesc: Stiinţa este sclava religiei în programe, manuale şi accesorii curriculare. Available online on August 14, 2009 on http://www.secularhumanism.ro/index.php/langro/comunicate-de-presa/77-stiinta-este-sclava-teologiei 7. Rutledge, Michael L., Melissa A. Warden. 1999. Development and validation of the measure of acceptance of the theory of evolution instrument. School Science and Mathematics 99: 13–18. 8. Deniz, Hasan, Lisa A. Donnelly, Irfan Yilmaz. 2008. Exploring the Factors 23
Related to Acceptance of Evolutionary Theory Among Turkish Preservice Biology Teachers: Toward a More Informative Conceptual Ecology for Biological Evolution. Journal of Research in Science Teaching 45, 4: 420-443 9. Eve, Raymond E. and Dana Dunn. 1990. Psychic Powers, Astrology & Creationism in the Classroom? Evidence of Pseudoscientific Beliefs among High School Biology & Life Science Teachers. The American Biology Teacher 52, 1: 10-21 10. Hayes, Bernadette. 2001. Gender, Scientific Knowledge, and Attitudes toward the Environment: A Cross-National Analysis. Political Research Quarterly 54: 657671 11. Hu, Li-tze and Peter M. Bentler. 1999. Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives. Structural Equation Modeling 6: 1-55. 12. Leslie J. Francis and John E. Greer. 1999. Attitudes towards creationism and evolutionary theory: the debate among secondary pupils attending Catholic and Protestant schools in Northern Ireland. Public Understanding of Science 8: 93-103 13. Mazur, Allan. 2004. Believers and Disbelievers in Evolution. Politics and the Life Sciences 23: 55-61 14. Miller, Jon D. 2004. Public understanding of, and attitudes toward, scientific research: what we know and what we need to know. Public Understanding of Science 13: 273-294. 15. Muthén, Linda K. and Muthén, Bengt O. 1998-2007. Mplus User’s Guide. Fifth Edition. Los Angeles, CA: Muthén & Muthén
24
16. Pardo Rafael, Félix Calvo. 2004 The cognitive dimension of public perceptions of science: methodological issues. Public Understanding of Science 13: 203–227 17. Pardo, Rafael, and Félix Calvo. 2008. Attitudes Toward Embryo Research, Worldviews, and the Moral Status of the Embryo Frame. Science Communication 30: 8-47. 18. Stan Lavinia, and Lucian Turcescu. Religious education in Romania. 2005. Communist and Post-Communist Studies 38: 381-401 19. Woodrum, Eric and Thomas Hoban. 1992. Support for Prayer in School and Creationism. Sociological Analysis, 53: 309-321
25
Tables
Table 1. Pearson correlations of Q9-7 and Q9-11 in the 2002 and 2005 Eurobarometers
Czech Bulgaria Rep. Estonia Hungary Latvia Lith. EB 2002 EB 0.18 2005 (0.01)
-
-
-
-
-
-
0.11 (0.01)
-
Poland 0.09 (0.05)
Romania 0.22 (0.01)
Slovakia 0.09 (0.05)
Slovenia 0.11 (0.01)
-
-
-
-
-
The absence of statistically significant influences is indicated by “-“. Statistical significance is specified under each coefficient.
Table 2. Observed independent variables in the structural model for acceptance of evolution
Observed
Abbreviation Item
independent variables Church
Church
attendance
“Do you attend religious services other than weddings or funerals… several times a week, once a week, a few times a year, once a year or less, or never?”
Respondent Education
Educ
“What is your level of education?”(country-specific response codes) Education variables from each country have been recoded and merged into a single variable with three values: 1. Primary education or less
26
2. Secondary education 3. Tertiary education Gender
Sex
“Masculine” is coded as 1 and “feminine” is coded as 0.
Age
Age100
Respondent age was divided by 100 in order to avoid model convergence problems (Muthén and Muthén 2007, 382).
Locality size
Urban
“Would you say you live in a… 1 – rural area or village 2 – small or middle sized town 3 – large town”
Denomination Catholic
“Do you consider yourself as belonging to a particular
Orthodox
religion? (If YES) Which one? (Show card – one answer
None
only) 1. Roman Catholic 2. Greek Catholic 3. Protestant 4. Orthodox 5. Jewish 6. Muslim 7. Buddhist 8. Hindu 9. Other 10. None, I do not consider myself belonging to a
27
particular religion” The variable was recoded in three dichotomous variables: - “Roman Catholic” (Yes/No), - “Orthodox” (Yes/No), - “No denomination” (Yes/No). Each model includes the “No denomination” variable and one of the other two, according to the majority denomination in each country: Bulgaria, Estonia, Latvia and Romania include the “Orthodox” variable, and the others include instead the “Roman Catholic” variable (see Error! Reference source not found.). The reference category is composed of all minority denominations, and it is country specific.
Table 3. Affiliation with religious denominations across ten countries Czech Bulgaria Latvia Lithuania Rep. Estonia Hungary Roman 1.2 34.8 0.4 60.4 21.4 80.2 Catholic Orthodox** 77.3 25.5 0.3 26.9 2.2 Greek 0.3 0.2 2.6 0.2 0.5 Catholic Protestant* 0.7 1.6 15.8 19.0 20.2 1.6 Jewish 0.5 0.6 0.1 0.1 Muslim 8.7 0.2 Buddhist 0.1 0.2 0.3 0.1 0.1 Other 1.5 4.0 1.1 2.0 2.0
Poland
Romania
Slovakia
Slovenia
94.7
6.1
68.0
59.2
0.1
86.2
0.3
2.8
0.2
1.4
5.5
-
0.2 0.8
3.1 0.1 0.1 2.1
6.5 0.1 4.8
0.4 1.4 0.6
None
11.9
61.6
53.1
15.9
29.1
13.3
4.0
1.0
14.8
35.6
Total
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
28
Table 4. Indicators of model fit for measurement models and for structural models Czech Bulgaria Rep. Estonia Hungary Latvia 79.50 58.85 100.16 62.24 97.86 Measurement Chi Square model Degrees of 31 34 36 32 37 freedom 0.973 0.974 0.938 0.979 0.951 CFI 0.977 0.977 0.941 0.982 0.954 TLI 0.040 0.026 0.042 0.031 0.040 RMSEA 1.057 0.879 1.134 0.901 1.104 WRMR 155.47 150.71 192.44 117.95 222.23 Structural Chi Square model Degrees of 66 65 78 73 88 freedom 0.947 0.947 0.945 0.974 0.983 CFI 0.959 0.950 0.950 0.979 0.983 TLI 0.037 0.035 0.038 0.025 0.039 RMSEA 0.993 1.030 1.083 0.833 1.109 WRMR
Lithuania 121.29
Poland 44.17
Romania 73.61
Slovakia 94.47
Slovenia 61.22
34 0.929 0.931 0.050 1.272 196.40
34 0.991 0.992 0.017 0.753 92.29
32 0.980 0.985 0.035 0.960 175.80
33 0.948 0.956 0.041 1.114 163.87
34 0.979 0.982 0.028 0.893 147.46
77 0.920 0.929 0.039 1.084
67 0.971 0.976 0.022 0.809
65 0.944 0.955 0.041 1.050
70 0.938 0.949 0.035 1.022
71 0.964 0.969 0.033 0.955
Table 5. Standardized (STDYX) estimates of direct and indirect effects from the structural model Czech Bulgaria Rep. Estonia Hungary Latvia Lithuania 0.239 0.216 0.297 0.266 Q9-11 Literacy on: -0.129 Church (0.05) -0.205 Orthodox NA NA NA NA NA NA Catholic 0.324 No denomination Education 0.086 0.099 0.108 Sex (0.05) (0.05) (0.05) -0.119 Age100 (0.05) -0.141 0.108 -0.172 Urban
29
Poland 0.302
Romania 0.357
Slovakia 0.314
Slovenia
-0.158
-0.267
-
NA -
0.131 (0.05) NA -
NA -
NA -0.144
-
-
-
-
-0.160 -
-
-
-
-
Bulgaria Literacy on
Czech Rep. (0.05)
Hungary
0.266
Education Sex Age100 Urban Orthodox Catholic No denomination Education
Poland
Romania
Slovakia
Slovenia
NA
0.093 (0.05)
NA
NA
-
NA
NA
-
NA
-
-
NA
-
-
0.364
0.314
0.384
0.332
0.368
0.230 -0.194 0.088 (0.05) -0.108 NA -0.426
0.182 -0.283
0.189 -0.248
0.149 -0.151
0.362 0.091 (0.05) -0.197
0.114 NA 0.148 -0.304 -0.066 (0.05)
0.128 NA -0.407 -0.077 (0.05)
0.227 -0.270 NA -0.247
NA -0.518
-
-
0.244 0.420 0.089 (0.05) -0.318 0.077 (0.05) NA -0.608 -0.062 (0.05)
-0.281 0.070 (0.05)
-0.225
-0.147
-0.220
-0.143
-
0.233
0.129
0.113
-
-0.121 0.162 0.390 0.166 -0.017 (0.05)
0.049 0.207 0.295
0.136 0.329 0.360
-0.162 0.169 0.319 0.295
0.135 -0.080 (0.05) 0.167 0.344 0.192
-0.098 0.238 0.260 0.462
-0.072 0.034 0.392 0.468
-
-
-
-
-
-
0.108
-
0.102
0.100
0.131
0.114
-
NA 0.494
0.419
NA -0.114 (0.05) 0.376
0.131 -0.177
0.129 -0.137
0.163 -0.289
0.220 -0.226
NA -0.271
0.113 NA -0.764 -
-0.090 NA -0.624 0.072 (0.05)
-0.186
-
-0.115
-
-0.058 (0.05) 0.189 0.257 0.398
-
0.135 NA -0.293 0.082 (0.05) -0.085 (0.05) 0.077 (0.05)
Sex Age100 Urban
Lithuania
-0.148
NA -0.299 (0.05)
No denomination
0.129 0.080 0.048 Q9-11 0.462 0.278 Literacy 0.203 0.377 Church Indirect Education via effects Church Education via 0.118 0.091 Literacy The absence of statistically significant influences is indicated by “-”. R2
Latvia
Orthodox Catholic
Church on
Estonia
-
All coefficients are significant for p=0.010, unless otherwise specified in parentheses.
30
Table 6. Significant Pearson correlations of item Q9-11 with items Q9-3 and Q9-8 q9-3 Knowledge: q9-8 Knowledge: Radioactive milk safe Antibiotics kill viruses as after boiling well as bacteria q9-11: Knowledge: Human beings developed from animals .077* Bulgaria -.064* Czech Rep. Estonia .066* 0.058 Hungary Latvia -.064* Lithuania .106** .094** Poland .123** .069* Romania .140** .074* Slovakia Slovenia
Flagged coefficients are significant for p=0.05 (*) and p=0.01 (**)
Table 7. Interaction of denomination with religiosity and church attendance in the structural model Model Q9-11 Czech on… Bulgaria Rep. Estonia Hungary Latvia Lithuania Slovakia 0.253 0.210 0.292 0.269 0.333 Total Literacy sample -0.190 -0.134 -0.174 -0.281 Church 0.220 No Literacy (0.200) 0.292 0.658 0.621 denomin ation -0.259 Church (0.05) 0.224 0.233 0.308 0.240 All Literacy denomin -0.136 Church ations -0.174 -0.195 (0.05) -0.251 The absence of statistically significant influences is indicated by “-“. All coefficients are significant for p=0.010, unless otherwise specified in parentheses.
Slovenia -
