- Email: [email protected]
The Flynn effect in the Czech Republic
INTELL-01173; No of Pages 4 Intelligence xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Intelligence
The Flynn effect in the Czech Republic Jiří Laciga, Hynek Cígler ⁎ Department of Psychology, Faculty of Social Sciences, Masaryk University, Brno, Czech Republic
a r t i c l e
i n f o
Article history: Received 25 July 2016 Received in revised form 18 November 2016 Accepted 22 November 2016 Available online xxxx Keywords: Flynn effect Rule-dependence model
a b s t r a c t The Flynn effect in the Czech population presented here was estimated based on the re-standardization of two short intelligence tests on a sample of 133 eighth-grade students. For over 44 years, the average performance in these tests had been improving by 0.23 and 0.42 IQ points per year respectively. There had been a significant difference between the tests at p b 0.05, although both tests had been designed to measure fluid intelligence. We propose that the gains in the test scores were generally brought about by changes in the Czech education system, while Armstrong and Woodley's (2014) rule-dependence model provides a reasonably accurate explanation for the difference between the tests. © 2016 Elsevier Inc. All rights reserved.
1. Introduction Since the early 1980s, J. R. Flynn, a political scientist, had been first to systematically investigate IQ test score changes in time. The increasing level of performance in intelligence tests was later coined as the Flynn effect (Herrnstein & Murray, 1996); some prior studies had shown similar results (e.g. Schaie & Strother, 1968). Flynn (1984) documented the occurrence of this phenomenon in different revisions of Stanford-Binet (SB) and Wechsler intelligence tests (WAIS, WISC). Flynn's study (1984) revealed a 13.8-point increase in IQ scores between 1932 and 1978, amounting to an approximately 0.3-point increase annually. Flynn (2009a) calculated IQ score gains between 1972 and 2006 based on normative samples of SB and Wechsler tests for adults and children. The average annual increase in IQ scores was 0.31, which proved to be consistent with Flynn's original findings. A meta-analysis was conducted by Trahan, Stuebing, Fletcher, and Hiscock (2014) using a wide range of 27 individually administered multifactorial intelligence tests (e.g. SB, WAIS, DAS (Differential Ability Scales)), Kaufman tests, WJ (Woodcock Johnson Test of Cognitive Ability). The individual comparative groups were tested with time gaps of at least five years and the meta-analysis included 378 comparisons between the older and the newer versions of the tests. The main finding was the meta-analytic estimation of the Flynn effect at 2.31 IQ points per ten years, with a 95% CI [2.0, 2.6] (Trahan et al., 2014). This number is slightly less than the 3.11 which Flynn had originally presented (Flynn, 1984). Nonetheless, considering that the ⁎ Corresponding author at: Department of Psychology, Faculty of Social Sciences, Masaryk University, Joštova 10, 602 00 Brno, Czech Republic. E-mail address: [email protected] (H. Cígler).
standard deviation of effects across 285 studies in the meta-analysis equaled to 2.5, it does not differ unexpectedly from Flynn's estimate. The lowest performance group showed IQ gains higher than the average; however, the pattern of the effect in the lowest performance group was unexpectedly bimodal. In addition, different intelligence tests produced different values of the Flynn effect: the highest gains were found in Wechsler tests and SB as well as in modern multifactorial intelligence tests, with the exception of Kaufman test (KABC). (Trahan et al., 2014) In the most recent meta-analysis, Pietschnig and Voracek (2015) estimated a mean effect of about 2.8 IQ points per decade with the highest gain in fluid (4.1 points per decade) and the lowest in crystalized IQ (2.1 points per decade). However, the effect was non-linear and it has been slowing down over the last decades; this challenges the aforementioned linearity assumption in Trahan et al.'s (2014) results. Furthermore, a decrease in IQ scores is present in some countries. In France, WAIS III (1999) and WAIS IV (2008–9) showed a decline of 3.8 IQ points on the total scale (Dutton & Lynn, 2015), a decrease was also observed in Germany and Austria (Pietschnig & Gittler, 2015), and Denmark (Teasdale & Owen, 2005). The Flynn effect differed across countries and was more prominent in adults than in children. Pietschnig and Voracek (2015) discuss proposed theories and factors presumably accounting for the effect. Their results provide support for the theory of life history speed (Woodley, 2012) as the main cause of the global trajectory of the Flynn effect, while the social multipliers theory (Dickens & Flynn, 2001) and economic prosperity (Rindermann, 2008) can in part explain the differences between countries. Our study is the first attempt to map the Flynn effect in the Czech population. The data were obtained during the re-standardization of two short intelligence tests. While other intelligence tests (e.g.
http://dx.doi.org/10.1016/j.intell.2016.11.005 0160-2896/© 2016 Elsevier Inc. All rights reserved.
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
2
J. Laciga, H. Cígler / Intelligence xxx (2016) xxx–xxx
Fig. 1. The 1st and the 10th item (of the 30 in total) in the IPT test.
Wechsler's) are widely used in the Czech Republic, they do not provide sufficient relevant data to determine the magnitude of the Flynn effect. 2. Methods 2.1. Materials The two original multiple-choice screening intelligence tests were designed by Pavel Říčan. Both are group-administered, paper-and-pencil methods; the respondent is to choose a single correct answer out of 5 options. The intellectual potential test (IPT; Říčan, 1971) aims to maximize rule finding and to minimize the visual factor. It involves geometric patterns, which combine principles of relation and sequence. The correlation of the IPT test with Raven's matrices was r = 0.64 (n = 52) with a total score of Amthauer's IST1 r = 0.55 (n = 57). Fig. 1 offers two exemplary test items. The test of number sequences (NS; Říčan, 1973) requires the participant to complete arithmetic and geometric, simple and combined number sequences. Fig. 2 shows an example of two test items. In the original study, the test correlated adequately with the IPT test (r = 0.63; the sample size was not mentioned). As pointed out by the author, the test measures a fluid intelligence component rather than the crystallized component. Both tests were conducted with students aged 12 and older. The reliability of the original test version, which was estimated using the splithalf method with the Spearman-Brown correction, was rxx′ = 0.919 for seventh-grade pupils (n = 51) and rxx′ = 0.885 for children aged 13;0– 14;0 (n = 45) (Říčan, 1971); other data are not available. The reliability of the IPT test in the current sample was estimated for five separate age categories used in the analysis (a timespan of roughly seven months, see below) using Cronbach's alpha, which ranged between 0.81 and 0.88 with the median of Md = 0.86. The internal consistency within the whole sample was α = 0.86. The internal consistency of the NS test in the original study on the comparative sample of eighth-grade pupils was α = 0.88 (n = 292); in our sample, Cronbach's alpha was α = 0.91. 2.2. Sample and procedure According to the author of the former standardization studies (Říčan, 1971, 1973), the original samples were representative of the Czech population (both norms were conducted for the Czech Republic only, although at the time it was a part of former Czechoslovakia). However, the method of sampling and data gathering was not described and both tests were standardized separately. The new sample was gathered with the purpose of updating norms. The distributor of both tests approached school psychologists and educational and psychological counselling centers' psychologists who had been involved in the distributor's earlier standardizations with the offer to become involved in the new study. The new norms have not been published to date. 1 The Intelligence Structure Test (IST) is a pencil-and-paper intelligence battery common in some European countries and it involves nine subtests with three indexes: verbal, numerical, and figural intelligence.
The Czech Republic being geographically homogenous, respondents were stratified merely by the size of municipality (5 categories). Both tests were group-administered and conducted in a bundle with two other short paper-and-pencil tests. They were taken by the whole class of students during standard teaching times, as in the original standardization. Additionally, the psychologists responsible for the testing were required to gather the same number of classes for each grade tested. The sampling procedure was not completely randomized, however, the correspondence of the new sample to the national census (age, sex, size of residence, and type of school) was reviewed and no significant difference was observed. As for the IPT test, the original sample contained 1025 children without sex distinction, aged 12;7–15;5. In addition, the research sample was split into five age subcategories of unspecified size. For the purposes of the following analyses we therefore assume that all of the groups were of the same size. In the new standardization study of the IPT test, the sample consisted of 359 children aged 12;7–15;5. This corresponds to the original sample of 187 boys and 168 girls (4 respondents did not specify the sex). The mean of the boys' overall scores in the IPT test, M = 19.48 (SD = 5.67), was lower than the girls' mean score, M = 20.61 (SD = 4.81), t(352) = − 2.04, and p = 0.042. However, the effect size was low, Cohen's d = 0.2. Having conducted the analysis for all grades individually, we observed a significant difference merely for age categories 13;2–13;8, t(62) = − 2.26 (boys achieved lower scores here) and 14;11–15;5, t(54) = 2.03, p = 0.048 (here, on the contrary, boys scored higher). Having no sex-differentiated score results for the original study, the differences observed being rather small, and bearing in mind that the Flynn effect should apply equally to men and women (Flynn, 1998; Pietschnig, Voracek, & Formann, 2011) we made the decision to conduct all analyses together for both sexes. As for the NS test, the original standardization study involved 292 eighth grade pupils (aged approximately 13–14). The norms for the other grades were not based on a representative sample and therefore could not have been included in the analysis. In the original study, the age of pupils was not specified, in contrast to the new study which, on the other hand, lacked information concerning the grade; that is why we decided to compare the children who would have been in the eighth grade had they started their obligatory school attendance at the standard age of 6 or 7, i.e. children aged 12–12;11.2 The number of such children was 133, out of which there were 71 boys and 62 girls. The average scores for both boys and girls were identical, t(129) = − 1.36, p = 0.176. For this reason and because the original sample was sex non-specific, we carried out the following analysis irrespective of sex. 2.3. Statistical analysis In order to express the changes in test performance on the IQ scale, we used the pooled standard deviation based on both samples and Welch's t-test for unequal variances in all cases. The information on distribution of both tests in the original study is drawn entirely from the 2 Nevertheless, the results presented in this study are significant even when comparing younger children aged 11–11;11 (N = 143, M = 21.03, SD = 7.21) with the original study sample, t(256) = 11.02, p b 0.001; the difference in this case was 17.2 IQ points.
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
J. Laciga, H. Cígler / Intelligence xxx (2016) xxx–xxx
3
Fig. 2. The 1st and the 10th item (of the 36 in total) in the NS test.
test manuals (Říčan, 1971; Říčan, 1973); the raw data were not available to date. 3. Results Table 1 gives descriptive statistics of the intellectual potential test (IPT) in both the original and the current study as well as the t-test of the differences between the two studies. The increase in test performance was significant in all age groups at significance level p b 0.001 (with the exception of 13;2–13; 8 age group where p b 0.01). The estimated increase ranged from 7.1 to 12.7 IQ points in the 44 years with Md = 10.4. Assuming the same size of the effect between 12 and 15 years of age, the average effect weighted according to the number of respondents in the given age group was M = 10.14 IQ with a 95% CI [8.08, 12.21]. This 95% confidence interval included all point estimates of the effect in all age groups. The differences of the effect values across all age groups corresponded to random differences, χ2 = 1.90, df = 4, p = 0.75. Furthermore, in the test of number sequences (NS) a significantly higher score was achieved by the children in the current sample (M = 21.50, SD = 7.59) than by the children in the original study (M = 13.2, SD = 6.45), t(222) = 10.93, p b 0.001. The difference was 17.66 IQ points with a 95% CI [14.50, 20.85]. Transformed to the annual increase, the Flynn effect size was IQ = 0.23 with a 95% CI [0.18, 0.28] for the IPT test scores (across all age categories) and IQ = 0.42 [0.34, 0.50] for the NS test. Based on the comparison of the confidence intervals it is evident that the effect size for number sequences is higher (p b 0.05). 4. Discussion The results presented here confirm the presence of the Flynn effect in the Czech population between 1971 and 2015; the observed increase was estimated at 0.23 and 0.42 IQ points annually. At the same time, the effect appears to be higher for the number sequencing test (NS) than for the intellectual potential test (IPT). The effect in the IPT test was consistent with the average effect in Trahan et al. (2014) and Pietschnig and Voracek's (2015) meta-analyses, while the effect for the number sequencing test was slightly higher than what their studies had estimated; this is discussed in more detail below. The Czech population was very stable during the reference period. Before 1990, there was minimum immigration due to the communist regime; after 1989, the number of people with permanent residence in the country increased by 222,652 and after 2003, Czech citizenship was acquired by 42,685 persons of whom 31% were culturally and
linguistically related Slovaks. At the end of 2015, the total population of the Czech Republic was 10,553,843 (Czech Statistical Office, 2016a). Mingroni (2007) assumes that the most likely cause of the rising IQ is heterosis, i.e. the process of mating between genetically distinct individuals which creates new favorable traits in their offspring. However, his theory has been given serious consideration (e.g. Woodley, 2011) and it seems highly improbable to provide a viable explanation for the Flynn effect described here, as the effect values are much higher than what could possibly be explained by heterosis. Equally improbable is the increase of the g-factor as the rise in the frame of the Flynn effect either does not correlate with the saturation of the g-factor or does so only slightly negatively (Jensen, 1998; Woodley & Meisenberg, 2013; te Nijenhuis and van der Flier (2013). The following factors are likely to have contributed to the rising performance in the IQ tests: an improvement in the quality of nutrition (Colom, Lluis-Font, & Andrés-Pueyo, 2005), a lower occurrence of harmful parasites in the environment (Eppig, Fincher, & Thornhill, 2010), an economic growth (Rindermann, 2008), which has been remarkable since the fall of the communist regime in 1989, and the rising standards of healthcare, in particular at the early stage of child development – the infant mortality rate before the end of the first year of life has continuously decreased in the Czech Republic and from 1971 to 2015 dropped from 2.0% to 0.2% (Czech Statistical Office, 2016b). However, Rutter (2000, p. 223) argues that this should not imply a positive effect on the mean IQ: the lowering early-age mortality rate contributes to a higher chance of survival for children born with disabilities which, without the use of modern technologies, would otherwise impose low chances of survival. The stability of the social and economic environment has led to higher life expectancy (Figueredo, de Baca, & Woodley, 2013). Furthermore, the number of children per family has decreased and thus parents have been able to devote more time to their children's upbringing. The general level of education increased gradually in the observed period. The share of university educated parents (considering the typical age of a parent to be 20–30 years; Czech Statistical Office, 2016c) of the children in the samples in the older studies was estimated to be 7%; the share of university educated parents in the present studies (typical age of parents here being 25–35 years) amounts to about 20% (Czech Statistical Office, 2016c). Moreover, the Czech education system has shifted its focus from factual knowledge to “ways of thinking” (Flynn, 2009b). For these reasons, the significant difference between the testing in two distinct time periods might find general explanation in the rising level of education in the population. As for the differences between the two tests used, Armstrong and Woodley (2014) propose the rule-dependence model: the IQ score
Table 1 The IPT test descriptives and the estimate of the size of the Flynn effect. 1971 study a
Age
N
12;7–13;1 13;2–13;8 13;9–14;3 14;4–14;10 14;11–15;5
205 205 205 205 205
a b
t-Testb
2015 study
Differences in IQ
M
SD
SE
N
M
SD
SE
t
df
p
Effect
95% CI
15.38 15.84 16.23 17.29 17.53
5.91 5.13 5.04 5.35 5.33
0.41 0.36 0.35 0.37 0.37
90 70 72 71 56
19.86 18.44 20.35 20.38 21.13
5.46 5.92 4.70 5.28 5.08
0.58 0.71 0.55 0.63 0.68
−6.32 −3.28 −6.27 −4.23 −4.64
183 107 132 123 91
0.000 0.001 0.000 0.000 0.000
11.8 7.1 12.7 8.7 10.4
[8.1, 15.5] [2.8, 11.3] [8.7, 16.7] [4.6, 12.8] [5.9, 14.8]
Only an estimate; the sample sizes in the original study were collective for all age categories. Welch's test (t-test for unequal variances).
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
4
J. Laciga, H. Cígler / Intelligence xxx (2016) xxx–xxx
gains are analogous to the individual rise in performance via retesting. As well as the training effect, the Flynn effect is linked with the ability of rule induction. The speed at which people learn implicitly the mental operations necessary to solve IQ tests is influenced by the presence and, more importantly, the number of rules utilized in a test. Tests containing a small number of rules, e.g. Raven's Progressive Matrices, are more sensitive to rule induction and therefore show higher IQ gains than tests containing no rules or using fewer predictable rules (e.g. Cattell's Culture Fair Test). In tests with a small number of rules, working memory and implicit learning become relatively less dependent on the g-factor. (Armstrong & Woodley, 2014) This appears consistent with the higher Flynn effect visible in the results of the number sequences test as opposed to the test of intellectual potential, which contains a higher number of rather ambiguous, less clear rules. In addition, tasks similar to numerical sequences are sometimes used in mathematics courses at Czech basic schools. We presume that the increasing IQ could be attributed to the external factors described above. The question that remains is whether these favorable conditions will persist and if or when they might cease to exist or even precipitate a reverse effect: a decline in IQ, which has been the case in some developed countries. Nevertheless, to some extent our study reaches its limits. The main problem is posed by the data gathering for two distinct time periods only. Thus, it can hardly be concluded that the increase in performance in the tests was linear between 1971 and 2015. Another aspect to take into consideration favors steep changes, as opposed to the linear growth. That is, inter alia, the end of the communist dictatorship in 1989 and the subsequent liberalization of the society as a whole, connected with the transformation of both the public sphere and education. Therefore, the estimate of the average annual increase in IQ must be treated with caution, even though it appears that the Flynn effect did exist in the observed period in the Czech Republic. Another limit is imposed by insufficient information on the original standardization studies (Říčan, 1971, 1973). The guideline manuals contain no information on the sampling methods nor do they provide exact demographic parameters (e.g. the sex of the tested persons), and overall descriptives are presented only. Additionally, no other sources of information on these prior standardization studies are available (P. Říčan, personal communication, June 17, 2016). In spite of the limits discussed, it does appear that for over 44 years, from 1971 to 2015, there was a noticeable increase in Czech children's intelligence test performance, namely in the range of 0.23–0.42 IQ points annually. At the same time, our results support the rule dependence model (Armstrong & Woodley, 2014), which argues that the Flynn effect is stronger for tests containing a smaller number of rules. References Armstrong, E. L., & Woodley, M. A. (2014). The rule-dependence model explains the commonalities between the Flynn effect and IQ gains via retesting. Learning and Individual Differences, 29, 41–49. http://dx.doi.org/10.1016/j.lindif.2013.10.009. Colom, R., Lluis-Font, J., & Andrés-Pueyo, A. (2005). The generational intelligence gains are caused by decreasing variance in the lower half of the distribution: Supporting evidence for the nutrition hypothesis. Intelligence, 33(1), 83–91. http://dx.doi.org/10. 1016/j.intell.2004.07.010. Czech Statistical Office (2016a). Foreigners by type of residence; 1985 - 2015 (31. 12.). Retrieved from https://www.czso.cz/documents/11292/41862280/c01R03_2015. pdf/0d942f31-a85c-4bf99cf7-d96f857d0b9c?version=1.0
Czech Statistical Office (2016b). Infant deaths and infant mortality rates: 1950–2015. Retrieved from https://www.czso.cz/documents/10180/32846217/130055160809.pdf/ 3d51b571-1734-4076-9112-e0735713df97?version=1.0 Czech Statistical Office (2016c). Highest educational attainment of population aged 15+ by 1950–2011 censuses. Retrieved from https://www.czso.cz/documents/10180/ 32846217/130055160115.pdf/9b9d5434-0b1f-432f-9992-2e5d5418353f?version= 1.0 Dickens, W. T., & Flynn, J. R. (2001). Heritability estimates versus large environmental effects: The IQ paradox resolved. Psychological Review, 108(2), 346–369. http://dx.doi. org/10.1037//0033-295x.108.2.346. Dutton, E. L., & Lynn, R. (2015). A negative Flynn effect in France, 1999 to 2008–9. Intelligence, 51, 67–70. http://dx.doi.org/10.1016/j.intell.2015.05.005. Eppig, C., Fincher, C. L., & Thornhill, R. (2010). Parasite prevalence and the worldwide distribution of cognitive ability. Proceedings of the Royal Society B: Biological Sciences, 277(1701), 3801–3808. http://dx.doi.org/10.1098/rspb.2010.0973. Figueredo, A. J., de Baca, T. C., & Woodley, M. A. (2013). The measurement of human life history strategy. Personality and Individual Differences, 55, 251–255. http://dx.doi.org/ 10.1016/j.paid.2012.04.033. Flynn, J. R. (1984). The mean IQ of Americans: Massive gains 1932 to 1978. Psychological Bulletin, 95(1), 29–51. http://dx.doi.org/10.1037/0033-2909.95.1.29. Flynn, J. R. (1998). Israeli military IQ tests: Gender differences small; IQ gains large. Journal of Biosocial Science, 30(4), 541–553. http://dx.doi.org/10.1017/ S0021932098005410. Flynn, J. R. (2009a). The WAIS-III and WAIS-IV: Daubert motions favor the certainly false over the approximately true. Applied Neuropsychology, 16(2), 98–104. http://dx.doi. org/10.1080/09084280902864360. Flynn, J. R. (2009b). What is intelligence? (2nd ed.). New York: Cambridge University Press. Herrnstein, R. J., & Murray, C. (1996). The bell curve: Intelligence and class structure in American society. New York, NY: Simon & Schuster. Jensen, A. R. (1998). The g factor: The science of mental ability. Westport: Praeger. Mingroni, M. A. (2007). Resolving the IQ paradox: Heterosis as a cause of the Flynn effect and other trends. Psychological Review, 114(3), 806–829. http://dx.doi.org/10.1037/ 0033-295X.114.3.806. Pietschnig, J., & Gittler, G. (2015). A reversal of the Flynn effect for spatial perception in German-speaking countries: Evidence from a cross-temporal IRT-based meta-analysis (1977–2014). Intelligence, 53, 145–153. http://dx.doi.org/10.1016/j.intell.2015.10. 004. Pietschnig, J., & Voracek, M. (2015). One century of global IQ gains: A formal meta-analysis of the Flynn effect (1909–2013). Perspectives on Psychological Science, 10(3), 282–306. http://dx.doi.org/10.1177/1745691615577701. Pietschnig, J., Voracek, M., & Formann, A. K. (2011). Female Flynn effects: No sex differences in generational IQ gains. Personality and Individual Differences, 50(5), 759–762. http://dx.doi.org/10.1016/j.paid.2010.12.019. Říčan, P. (1971). Test intelektového potenciálu. Bratislava: Psychodiagnostické a didaktické testy. Říčan, P. (1973). Číselné řady. Bratislava: Psychodiagnostické a didaktické testy. Rindermann, H. (2008). Relevance of education and intelligence at the national level for the economic welfare of people. Intelligence, 36(2), 127–142. http://dx.doi.org/10. 1016/j.intell.2007.02.002. Rutter, J. M. (2000). Comments in discussion on James R. Flynn. In G. R. Bock, J. A. Goode, & K. Webb (Eds.), The nature of intelligence (pp. 222–223). Chichester: JohnWiley & Sons. Schaie, K. W., & Strother, C. R. (1968). A cross-sequential study of age changes in cognitive behavior. Psychological Bulletin, 70(6), 671–680. http://dx.doi.org/10.1037/h0026811. te Nijenhuis, J., & van der Flier, H. (2013). Is the Flynn effect on g?: A meta-analysis. Intelligence, 41(6), 802–807. http://dx.doi.org/10.1016/j.intell.2013.03.001. Teasdale, T. W., & Owen, D. R. (2005). A long-term rise and recent decline in intelligence test performance: The Flynn effect in reverse. Personality and Individual Differences, 39(4), 837–843. http://dx.doi.org/10.1016/j.paid.2005.01.029. Trahan, L. H., Stuebing, K. K., Fletcher, J. M., & Hiscock, M. (2014). The Flynn effect: A meta-analysis. Psychological Bulletin, 140(5), 1332–1360. http://dx.doi.org/10.1037/ a0037173. Woodley, M. A. (2011). Heterosis doesn't cause the Flynn effect: A critical examination of Mingroni (2007). Psychological Review, 118(4), 689–693. http://dx.doi.org/10.1037/ a0024759. Woodley, M. A. (2012). A life history model of the Lynn–Flynn effect. Personality and Individual Differences, 53(2), 152–156. http://dx.doi.org/10.1016/j.paid.2011.03.028. Woodley, M. A., & Meisenberg, G. (2013). In the Netherlands the anti-Flynn effect is a Jensen effect. Personality and Individual Differences, 54(8), 871–876. http://dx.doi. org/10.1016/j.paid.2012.12.022.
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
Contents lists available at ScienceDirect
Intelligence
The Flynn effect in the Czech Republic Jiří Laciga, Hynek Cígler ⁎ Department of Psychology, Faculty of Social Sciences, Masaryk University, Brno, Czech Republic
a r t i c l e
i n f o
Article history: Received 25 July 2016 Received in revised form 18 November 2016 Accepted 22 November 2016 Available online xxxx Keywords: Flynn effect Rule-dependence model
a b s t r a c t The Flynn effect in the Czech population presented here was estimated based on the re-standardization of two short intelligence tests on a sample of 133 eighth-grade students. For over 44 years, the average performance in these tests had been improving by 0.23 and 0.42 IQ points per year respectively. There had been a significant difference between the tests at p b 0.05, although both tests had been designed to measure fluid intelligence. We propose that the gains in the test scores were generally brought about by changes in the Czech education system, while Armstrong and Woodley's (2014) rule-dependence model provides a reasonably accurate explanation for the difference between the tests. © 2016 Elsevier Inc. All rights reserved.
1. Introduction Since the early 1980s, J. R. Flynn, a political scientist, had been first to systematically investigate IQ test score changes in time. The increasing level of performance in intelligence tests was later coined as the Flynn effect (Herrnstein & Murray, 1996); some prior studies had shown similar results (e.g. Schaie & Strother, 1968). Flynn (1984) documented the occurrence of this phenomenon in different revisions of Stanford-Binet (SB) and Wechsler intelligence tests (WAIS, WISC). Flynn's study (1984) revealed a 13.8-point increase in IQ scores between 1932 and 1978, amounting to an approximately 0.3-point increase annually. Flynn (2009a) calculated IQ score gains between 1972 and 2006 based on normative samples of SB and Wechsler tests for adults and children. The average annual increase in IQ scores was 0.31, which proved to be consistent with Flynn's original findings. A meta-analysis was conducted by Trahan, Stuebing, Fletcher, and Hiscock (2014) using a wide range of 27 individually administered multifactorial intelligence tests (e.g. SB, WAIS, DAS (Differential Ability Scales)), Kaufman tests, WJ (Woodcock Johnson Test of Cognitive Ability). The individual comparative groups were tested with time gaps of at least five years and the meta-analysis included 378 comparisons between the older and the newer versions of the tests. The main finding was the meta-analytic estimation of the Flynn effect at 2.31 IQ points per ten years, with a 95% CI [2.0, 2.6] (Trahan et al., 2014). This number is slightly less than the 3.11 which Flynn had originally presented (Flynn, 1984). Nonetheless, considering that the ⁎ Corresponding author at: Department of Psychology, Faculty of Social Sciences, Masaryk University, Joštova 10, 602 00 Brno, Czech Republic. E-mail address: [email protected] (H. Cígler).
standard deviation of effects across 285 studies in the meta-analysis equaled to 2.5, it does not differ unexpectedly from Flynn's estimate. The lowest performance group showed IQ gains higher than the average; however, the pattern of the effect in the lowest performance group was unexpectedly bimodal. In addition, different intelligence tests produced different values of the Flynn effect: the highest gains were found in Wechsler tests and SB as well as in modern multifactorial intelligence tests, with the exception of Kaufman test (KABC). (Trahan et al., 2014) In the most recent meta-analysis, Pietschnig and Voracek (2015) estimated a mean effect of about 2.8 IQ points per decade with the highest gain in fluid (4.1 points per decade) and the lowest in crystalized IQ (2.1 points per decade). However, the effect was non-linear and it has been slowing down over the last decades; this challenges the aforementioned linearity assumption in Trahan et al.'s (2014) results. Furthermore, a decrease in IQ scores is present in some countries. In France, WAIS III (1999) and WAIS IV (2008–9) showed a decline of 3.8 IQ points on the total scale (Dutton & Lynn, 2015), a decrease was also observed in Germany and Austria (Pietschnig & Gittler, 2015), and Denmark (Teasdale & Owen, 2005). The Flynn effect differed across countries and was more prominent in adults than in children. Pietschnig and Voracek (2015) discuss proposed theories and factors presumably accounting for the effect. Their results provide support for the theory of life history speed (Woodley, 2012) as the main cause of the global trajectory of the Flynn effect, while the social multipliers theory (Dickens & Flynn, 2001) and economic prosperity (Rindermann, 2008) can in part explain the differences between countries. Our study is the first attempt to map the Flynn effect in the Czech population. The data were obtained during the re-standardization of two short intelligence tests. While other intelligence tests (e.g.
http://dx.doi.org/10.1016/j.intell.2016.11.005 0160-2896/© 2016 Elsevier Inc. All rights reserved.
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
2
J. Laciga, H. Cígler / Intelligence xxx (2016) xxx–xxx
Fig. 1. The 1st and the 10th item (of the 30 in total) in the IPT test.
Wechsler's) are widely used in the Czech Republic, they do not provide sufficient relevant data to determine the magnitude of the Flynn effect. 2. Methods 2.1. Materials The two original multiple-choice screening intelligence tests were designed by Pavel Říčan. Both are group-administered, paper-and-pencil methods; the respondent is to choose a single correct answer out of 5 options. The intellectual potential test (IPT; Říčan, 1971) aims to maximize rule finding and to minimize the visual factor. It involves geometric patterns, which combine principles of relation and sequence. The correlation of the IPT test with Raven's matrices was r = 0.64 (n = 52) with a total score of Amthauer's IST1 r = 0.55 (n = 57). Fig. 1 offers two exemplary test items. The test of number sequences (NS; Říčan, 1973) requires the participant to complete arithmetic and geometric, simple and combined number sequences. Fig. 2 shows an example of two test items. In the original study, the test correlated adequately with the IPT test (r = 0.63; the sample size was not mentioned). As pointed out by the author, the test measures a fluid intelligence component rather than the crystallized component. Both tests were conducted with students aged 12 and older. The reliability of the original test version, which was estimated using the splithalf method with the Spearman-Brown correction, was rxx′ = 0.919 for seventh-grade pupils (n = 51) and rxx′ = 0.885 for children aged 13;0– 14;0 (n = 45) (Říčan, 1971); other data are not available. The reliability of the IPT test in the current sample was estimated for five separate age categories used in the analysis (a timespan of roughly seven months, see below) using Cronbach's alpha, which ranged between 0.81 and 0.88 with the median of Md = 0.86. The internal consistency within the whole sample was α = 0.86. The internal consistency of the NS test in the original study on the comparative sample of eighth-grade pupils was α = 0.88 (n = 292); in our sample, Cronbach's alpha was α = 0.91. 2.2. Sample and procedure According to the author of the former standardization studies (Říčan, 1971, 1973), the original samples were representative of the Czech population (both norms were conducted for the Czech Republic only, although at the time it was a part of former Czechoslovakia). However, the method of sampling and data gathering was not described and both tests were standardized separately. The new sample was gathered with the purpose of updating norms. The distributor of both tests approached school psychologists and educational and psychological counselling centers' psychologists who had been involved in the distributor's earlier standardizations with the offer to become involved in the new study. The new norms have not been published to date. 1 The Intelligence Structure Test (IST) is a pencil-and-paper intelligence battery common in some European countries and it involves nine subtests with three indexes: verbal, numerical, and figural intelligence.
The Czech Republic being geographically homogenous, respondents were stratified merely by the size of municipality (5 categories). Both tests were group-administered and conducted in a bundle with two other short paper-and-pencil tests. They were taken by the whole class of students during standard teaching times, as in the original standardization. Additionally, the psychologists responsible for the testing were required to gather the same number of classes for each grade tested. The sampling procedure was not completely randomized, however, the correspondence of the new sample to the national census (age, sex, size of residence, and type of school) was reviewed and no significant difference was observed. As for the IPT test, the original sample contained 1025 children without sex distinction, aged 12;7–15;5. In addition, the research sample was split into five age subcategories of unspecified size. For the purposes of the following analyses we therefore assume that all of the groups were of the same size. In the new standardization study of the IPT test, the sample consisted of 359 children aged 12;7–15;5. This corresponds to the original sample of 187 boys and 168 girls (4 respondents did not specify the sex). The mean of the boys' overall scores in the IPT test, M = 19.48 (SD = 5.67), was lower than the girls' mean score, M = 20.61 (SD = 4.81), t(352) = − 2.04, and p = 0.042. However, the effect size was low, Cohen's d = 0.2. Having conducted the analysis for all grades individually, we observed a significant difference merely for age categories 13;2–13;8, t(62) = − 2.26 (boys achieved lower scores here) and 14;11–15;5, t(54) = 2.03, p = 0.048 (here, on the contrary, boys scored higher). Having no sex-differentiated score results for the original study, the differences observed being rather small, and bearing in mind that the Flynn effect should apply equally to men and women (Flynn, 1998; Pietschnig, Voracek, & Formann, 2011) we made the decision to conduct all analyses together for both sexes. As for the NS test, the original standardization study involved 292 eighth grade pupils (aged approximately 13–14). The norms for the other grades were not based on a representative sample and therefore could not have been included in the analysis. In the original study, the age of pupils was not specified, in contrast to the new study which, on the other hand, lacked information concerning the grade; that is why we decided to compare the children who would have been in the eighth grade had they started their obligatory school attendance at the standard age of 6 or 7, i.e. children aged 12–12;11.2 The number of such children was 133, out of which there were 71 boys and 62 girls. The average scores for both boys and girls were identical, t(129) = − 1.36, p = 0.176. For this reason and because the original sample was sex non-specific, we carried out the following analysis irrespective of sex. 2.3. Statistical analysis In order to express the changes in test performance on the IQ scale, we used the pooled standard deviation based on both samples and Welch's t-test for unequal variances in all cases. The information on distribution of both tests in the original study is drawn entirely from the 2 Nevertheless, the results presented in this study are significant even when comparing younger children aged 11–11;11 (N = 143, M = 21.03, SD = 7.21) with the original study sample, t(256) = 11.02, p b 0.001; the difference in this case was 17.2 IQ points.
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
J. Laciga, H. Cígler / Intelligence xxx (2016) xxx–xxx
3
Fig. 2. The 1st and the 10th item (of the 36 in total) in the NS test.
test manuals (Říčan, 1971; Říčan, 1973); the raw data were not available to date. 3. Results Table 1 gives descriptive statistics of the intellectual potential test (IPT) in both the original and the current study as well as the t-test of the differences between the two studies. The increase in test performance was significant in all age groups at significance level p b 0.001 (with the exception of 13;2–13; 8 age group where p b 0.01). The estimated increase ranged from 7.1 to 12.7 IQ points in the 44 years with Md = 10.4. Assuming the same size of the effect between 12 and 15 years of age, the average effect weighted according to the number of respondents in the given age group was M = 10.14 IQ with a 95% CI [8.08, 12.21]. This 95% confidence interval included all point estimates of the effect in all age groups. The differences of the effect values across all age groups corresponded to random differences, χ2 = 1.90, df = 4, p = 0.75. Furthermore, in the test of number sequences (NS) a significantly higher score was achieved by the children in the current sample (M = 21.50, SD = 7.59) than by the children in the original study (M = 13.2, SD = 6.45), t(222) = 10.93, p b 0.001. The difference was 17.66 IQ points with a 95% CI [14.50, 20.85]. Transformed to the annual increase, the Flynn effect size was IQ = 0.23 with a 95% CI [0.18, 0.28] for the IPT test scores (across all age categories) and IQ = 0.42 [0.34, 0.50] for the NS test. Based on the comparison of the confidence intervals it is evident that the effect size for number sequences is higher (p b 0.05). 4. Discussion The results presented here confirm the presence of the Flynn effect in the Czech population between 1971 and 2015; the observed increase was estimated at 0.23 and 0.42 IQ points annually. At the same time, the effect appears to be higher for the number sequencing test (NS) than for the intellectual potential test (IPT). The effect in the IPT test was consistent with the average effect in Trahan et al. (2014) and Pietschnig and Voracek's (2015) meta-analyses, while the effect for the number sequencing test was slightly higher than what their studies had estimated; this is discussed in more detail below. The Czech population was very stable during the reference period. Before 1990, there was minimum immigration due to the communist regime; after 1989, the number of people with permanent residence in the country increased by 222,652 and after 2003, Czech citizenship was acquired by 42,685 persons of whom 31% were culturally and
linguistically related Slovaks. At the end of 2015, the total population of the Czech Republic was 10,553,843 (Czech Statistical Office, 2016a). Mingroni (2007) assumes that the most likely cause of the rising IQ is heterosis, i.e. the process of mating between genetically distinct individuals which creates new favorable traits in their offspring. However, his theory has been given serious consideration (e.g. Woodley, 2011) and it seems highly improbable to provide a viable explanation for the Flynn effect described here, as the effect values are much higher than what could possibly be explained by heterosis. Equally improbable is the increase of the g-factor as the rise in the frame of the Flynn effect either does not correlate with the saturation of the g-factor or does so only slightly negatively (Jensen, 1998; Woodley & Meisenberg, 2013; te Nijenhuis and van der Flier (2013). The following factors are likely to have contributed to the rising performance in the IQ tests: an improvement in the quality of nutrition (Colom, Lluis-Font, & Andrés-Pueyo, 2005), a lower occurrence of harmful parasites in the environment (Eppig, Fincher, & Thornhill, 2010), an economic growth (Rindermann, 2008), which has been remarkable since the fall of the communist regime in 1989, and the rising standards of healthcare, in particular at the early stage of child development – the infant mortality rate before the end of the first year of life has continuously decreased in the Czech Republic and from 1971 to 2015 dropped from 2.0% to 0.2% (Czech Statistical Office, 2016b). However, Rutter (2000, p. 223) argues that this should not imply a positive effect on the mean IQ: the lowering early-age mortality rate contributes to a higher chance of survival for children born with disabilities which, without the use of modern technologies, would otherwise impose low chances of survival. The stability of the social and economic environment has led to higher life expectancy (Figueredo, de Baca, & Woodley, 2013). Furthermore, the number of children per family has decreased and thus parents have been able to devote more time to their children's upbringing. The general level of education increased gradually in the observed period. The share of university educated parents (considering the typical age of a parent to be 20–30 years; Czech Statistical Office, 2016c) of the children in the samples in the older studies was estimated to be 7%; the share of university educated parents in the present studies (typical age of parents here being 25–35 years) amounts to about 20% (Czech Statistical Office, 2016c). Moreover, the Czech education system has shifted its focus from factual knowledge to “ways of thinking” (Flynn, 2009b). For these reasons, the significant difference between the testing in two distinct time periods might find general explanation in the rising level of education in the population. As for the differences between the two tests used, Armstrong and Woodley (2014) propose the rule-dependence model: the IQ score
Table 1 The IPT test descriptives and the estimate of the size of the Flynn effect. 1971 study a
Age
N
12;7–13;1 13;2–13;8 13;9–14;3 14;4–14;10 14;11–15;5
205 205 205 205 205
a b
t-Testb
2015 study
Differences in IQ
M
SD
SE
N
M
SD
SE
t
df
p
Effect
95% CI
15.38 15.84 16.23 17.29 17.53
5.91 5.13 5.04 5.35 5.33
0.41 0.36 0.35 0.37 0.37
90 70 72 71 56
19.86 18.44 20.35 20.38 21.13
5.46 5.92 4.70 5.28 5.08
0.58 0.71 0.55 0.63 0.68
−6.32 −3.28 −6.27 −4.23 −4.64
183 107 132 123 91
0.000 0.001 0.000 0.000 0.000
11.8 7.1 12.7 8.7 10.4
[8.1, 15.5] [2.8, 11.3] [8.7, 16.7] [4.6, 12.8] [5.9, 14.8]
Only an estimate; the sample sizes in the original study were collective for all age categories. Welch's test (t-test for unequal variances).
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005
4
J. Laciga, H. Cígler / Intelligence xxx (2016) xxx–xxx
gains are analogous to the individual rise in performance via retesting. As well as the training effect, the Flynn effect is linked with the ability of rule induction. The speed at which people learn implicitly the mental operations necessary to solve IQ tests is influenced by the presence and, more importantly, the number of rules utilized in a test. Tests containing a small number of rules, e.g. Raven's Progressive Matrices, are more sensitive to rule induction and therefore show higher IQ gains than tests containing no rules or using fewer predictable rules (e.g. Cattell's Culture Fair Test). In tests with a small number of rules, working memory and implicit learning become relatively less dependent on the g-factor. (Armstrong & Woodley, 2014) This appears consistent with the higher Flynn effect visible in the results of the number sequences test as opposed to the test of intellectual potential, which contains a higher number of rather ambiguous, less clear rules. In addition, tasks similar to numerical sequences are sometimes used in mathematics courses at Czech basic schools. We presume that the increasing IQ could be attributed to the external factors described above. The question that remains is whether these favorable conditions will persist and if or when they might cease to exist or even precipitate a reverse effect: a decline in IQ, which has been the case in some developed countries. Nevertheless, to some extent our study reaches its limits. The main problem is posed by the data gathering for two distinct time periods only. Thus, it can hardly be concluded that the increase in performance in the tests was linear between 1971 and 2015. Another aspect to take into consideration favors steep changes, as opposed to the linear growth. That is, inter alia, the end of the communist dictatorship in 1989 and the subsequent liberalization of the society as a whole, connected with the transformation of both the public sphere and education. Therefore, the estimate of the average annual increase in IQ must be treated with caution, even though it appears that the Flynn effect did exist in the observed period in the Czech Republic. Another limit is imposed by insufficient information on the original standardization studies (Říčan, 1971, 1973). The guideline manuals contain no information on the sampling methods nor do they provide exact demographic parameters (e.g. the sex of the tested persons), and overall descriptives are presented only. Additionally, no other sources of information on these prior standardization studies are available (P. Říčan, personal communication, June 17, 2016). In spite of the limits discussed, it does appear that for over 44 years, from 1971 to 2015, there was a noticeable increase in Czech children's intelligence test performance, namely in the range of 0.23–0.42 IQ points annually. At the same time, our results support the rule dependence model (Armstrong & Woodley, 2014), which argues that the Flynn effect is stronger for tests containing a smaller number of rules. References Armstrong, E. L., & Woodley, M. A. (2014). The rule-dependence model explains the commonalities between the Flynn effect and IQ gains via retesting. Learning and Individual Differences, 29, 41–49. http://dx.doi.org/10.1016/j.lindif.2013.10.009. Colom, R., Lluis-Font, J., & Andrés-Pueyo, A. (2005). The generational intelligence gains are caused by decreasing variance in the lower half of the distribution: Supporting evidence for the nutrition hypothesis. Intelligence, 33(1), 83–91. http://dx.doi.org/10. 1016/j.intell.2004.07.010. Czech Statistical Office (2016a). Foreigners by type of residence; 1985 - 2015 (31. 12.). Retrieved from https://www.czso.cz/documents/11292/41862280/c01R03_2015. pdf/0d942f31-a85c-4bf99cf7-d96f857d0b9c?version=1.0
Czech Statistical Office (2016b). Infant deaths and infant mortality rates: 1950–2015. Retrieved from https://www.czso.cz/documents/10180/32846217/130055160809.pdf/ 3d51b571-1734-4076-9112-e0735713df97?version=1.0 Czech Statistical Office (2016c). Highest educational attainment of population aged 15+ by 1950–2011 censuses. Retrieved from https://www.czso.cz/documents/10180/ 32846217/130055160115.pdf/9b9d5434-0b1f-432f-9992-2e5d5418353f?version= 1.0 Dickens, W. T., & Flynn, J. R. (2001). Heritability estimates versus large environmental effects: The IQ paradox resolved. Psychological Review, 108(2), 346–369. http://dx.doi. org/10.1037//0033-295x.108.2.346. Dutton, E. L., & Lynn, R. (2015). A negative Flynn effect in France, 1999 to 2008–9. Intelligence, 51, 67–70. http://dx.doi.org/10.1016/j.intell.2015.05.005. Eppig, C., Fincher, C. L., & Thornhill, R. (2010). Parasite prevalence and the worldwide distribution of cognitive ability. Proceedings of the Royal Society B: Biological Sciences, 277(1701), 3801–3808. http://dx.doi.org/10.1098/rspb.2010.0973. Figueredo, A. J., de Baca, T. C., & Woodley, M. A. (2013). The measurement of human life history strategy. Personality and Individual Differences, 55, 251–255. http://dx.doi.org/ 10.1016/j.paid.2012.04.033. Flynn, J. R. (1984). The mean IQ of Americans: Massive gains 1932 to 1978. Psychological Bulletin, 95(1), 29–51. http://dx.doi.org/10.1037/0033-2909.95.1.29. Flynn, J. R. (1998). Israeli military IQ tests: Gender differences small; IQ gains large. Journal of Biosocial Science, 30(4), 541–553. http://dx.doi.org/10.1017/ S0021932098005410. Flynn, J. R. (2009a). The WAIS-III and WAIS-IV: Daubert motions favor the certainly false over the approximately true. Applied Neuropsychology, 16(2), 98–104. http://dx.doi. org/10.1080/09084280902864360. Flynn, J. R. (2009b). What is intelligence? (2nd ed.). New York: Cambridge University Press. Herrnstein, R. J., & Murray, C. (1996). The bell curve: Intelligence and class structure in American society. New York, NY: Simon & Schuster. Jensen, A. R. (1998). The g factor: The science of mental ability. Westport: Praeger. Mingroni, M. A. (2007). Resolving the IQ paradox: Heterosis as a cause of the Flynn effect and other trends. Psychological Review, 114(3), 806–829. http://dx.doi.org/10.1037/ 0033-295X.114.3.806. Pietschnig, J., & Gittler, G. (2015). A reversal of the Flynn effect for spatial perception in German-speaking countries: Evidence from a cross-temporal IRT-based meta-analysis (1977–2014). Intelligence, 53, 145–153. http://dx.doi.org/10.1016/j.intell.2015.10. 004. Pietschnig, J., & Voracek, M. (2015). One century of global IQ gains: A formal meta-analysis of the Flynn effect (1909–2013). Perspectives on Psychological Science, 10(3), 282–306. http://dx.doi.org/10.1177/1745691615577701. Pietschnig, J., Voracek, M., & Formann, A. K. (2011). Female Flynn effects: No sex differences in generational IQ gains. Personality and Individual Differences, 50(5), 759–762. http://dx.doi.org/10.1016/j.paid.2010.12.019. Říčan, P. (1971). Test intelektového potenciálu. Bratislava: Psychodiagnostické a didaktické testy. Říčan, P. (1973). Číselné řady. Bratislava: Psychodiagnostické a didaktické testy. Rindermann, H. (2008). Relevance of education and intelligence at the national level for the economic welfare of people. Intelligence, 36(2), 127–142. http://dx.doi.org/10. 1016/j.intell.2007.02.002. Rutter, J. M. (2000). Comments in discussion on James R. Flynn. In G. R. Bock, J. A. Goode, & K. Webb (Eds.), The nature of intelligence (pp. 222–223). Chichester: JohnWiley & Sons. Schaie, K. W., & Strother, C. R. (1968). A cross-sequential study of age changes in cognitive behavior. Psychological Bulletin, 70(6), 671–680. http://dx.doi.org/10.1037/h0026811. te Nijenhuis, J., & van der Flier, H. (2013). Is the Flynn effect on g?: A meta-analysis. Intelligence, 41(6), 802–807. http://dx.doi.org/10.1016/j.intell.2013.03.001. Teasdale, T. W., & Owen, D. R. (2005). A long-term rise and recent decline in intelligence test performance: The Flynn effect in reverse. Personality and Individual Differences, 39(4), 837–843. http://dx.doi.org/10.1016/j.paid.2005.01.029. Trahan, L. H., Stuebing, K. K., Fletcher, J. M., & Hiscock, M. (2014). The Flynn effect: A meta-analysis. Psychological Bulletin, 140(5), 1332–1360. http://dx.doi.org/10.1037/ a0037173. Woodley, M. A. (2011). Heterosis doesn't cause the Flynn effect: A critical examination of Mingroni (2007). Psychological Review, 118(4), 689–693. http://dx.doi.org/10.1037/ a0024759. Woodley, M. A. (2012). A life history model of the Lynn–Flynn effect. Personality and Individual Differences, 53(2), 152–156. http://dx.doi.org/10.1016/j.paid.2011.03.028. Woodley, M. A., & Meisenberg, G. (2013). In the Netherlands the anti-Flynn effect is a Jensen effect. Personality and Individual Differences, 54(8), 871–876. http://dx.doi. org/10.1016/j.paid.2012.12.022.
Please cite this article as: Laciga, J., & Cígler, H., The Flynn effect in the Czech Republic, Intelligence (2016), http://dx.doi.org/10.1016/ j.intell.2016.11.005