The National Intelligence Tests (NIT) were originally developed from the Army Alfa and Beta tests, and modified for educational use. This test is based on a sound theory, has a strict testing structure, and most importantly -... [ view full abstract ]
The National Intelligence Tests (NIT) were originally developed from the Army Alfa and Beta tests, and modified for educational use. This test is based on a sound theory, has a strict testing structure, and most importantly - is appropriate for use decades later. The Estonian version of the NIT was developed in the 1930’s by Juhan Tork. This version has been the foundation of several Flynn Effect (FE) studies. The comparison of two cohorts of 13-14 years old students from 1933/1936 to 2006 was done.
The most important published findings are as follows:
The secular gains were most pronounced in the low g-loaded subtests.
The IQ scores of the different cohorts are not directly comparable: the duration of time between the measurements, and the societal/educational changes have occurred between the testing of the two cohorts. The meaning of subtests and the meaning of the g score have changed over time.
Rapid guessing in test-taking has increased over time and has had a pronounced influence on test scores. The increased guessing enhances the independence of the FE from g. The FE can partly be explained by the changes in test-taking behavior.
The FE is evident in all of the subtests, except one (computing). There is however, some variability in the magnitude of the FE as it can range from an effect size of 0.24 to 1.05.
If the confounds are controlled, then the abstractness of the test items can positively predict the FE magnitudes.
Word abstractness was found to be a strong predictor of all the testing results in both cohorts. The FE is primarily an indicator of the rise of abstract thinking ability.
The subtest versions consisting of only speeded items shows a large positive FE, with cohort differences up to 50%. The rise in test-taking speed between cohorts can be attributed to an increase in automated responses. It is apparent that the younger cohort is faster than the older one, but the differences in level cannot be determined with certainty.
Conclusion: the rise in IQ scores over time cannot simply be attributed to a difference in group means, and there are several additional factors that are shaping this phenomenon.
