Article summary with Massive IQ gains in 14 nations: What IQ tests really measure by Flynn - 1987

Introduction

This article presents the results of IQ trends that have been measured in various countries. The data was analyzed in order to acquire the benefits that comparative data can often give. A closer look is taken into which factors have the most powerful effect on IQ, what an IQ test actually measures, and finally, how these tests can be used when comparing between-groups.

Data Collection

Data was collected using letters, personal appeals, and/ or questionnaires to all researchers known to be interested in this field (165 researchers from 35 countries worldwide).

Data Presentation

Data was evaluated according to four general criteria:

1. That possible sample bias was eliminated by means of comprehensive samples (f.e. mass testing of registrants)

2. That the test stayed the same from one generation to the next, whereby estimates of trends are based on the raw score differences

3. That the tests that are culturally reduced (for example, tests such as the Ravens Progressive Matrices Test, should receive particular emphasis in the study)

4. That the subjects that have acquired the peak of their raw score performance should receive particular emphasis (this criterion was included as it could remove the possibility that the present generation reaches peak performance at a younger age that the generation before them so that, at full maturity, the scores of both the generations are approximately similar).

Tests used to test IQ are in the studies were: Ravens Progressive Matrices Test (Ravens), Shapes, Matrices, Horn-Ravens, Jenkins, Wechsler Intellignece Scale for r Children (Wechsler), California Test of Mental Maturity (CTMM), Wechsler-Binet, Otis Quick Scoring Beta, Form A and Otis-Lennon, Form J (Otis) and Verbal-Math

Test continuity, sample quality, and estimates of IQ gains were used to assess every dataset and give each dataset a status:

1. Status 1 was given to datasets that verify IQ gains well (thus beyond a reasonable doubt)

2. Status 2 was given to datasets that offer probable evidence of IQ gains

3. Status 3 was given to datasets that offer tentative evidence of IQ gains

4. Status 4 was given to datasets that offer only speculative evidence of IQ gains

By giving each dataset a status, it was possible to make a natural division, namely strong and weak datasets.

Strong data was derived from the Netherlands using Ravens (status 1); Belgium using Ravens (status 1), using Shapes (status 1), and using Verbal-Math (status 1); France using Ravens (status 3), and using Verbal-Math (status 3); Norway using Matrices (status 1) and using Verbal-Math (status 1); New Zealand using Otis (status 1), Canada – Edmonton using CTMM (status 1), Saskatchewan using Otis on 10-year-old children (status 2/3) and Saskatchewan using Otis on 13-year olds (status 2/3); and the United States using the Wechsler-Binet (status 2), and the Wechsler (status 3).

Weak data was derived from East Germany using Ravens (status 3/2); Great Britain using Ravens on children age 8-14 (status 3), and on ages 20-30 (status 3); Australia using Jenkins (status 3), and using Ravens (status 4); Japan using Wechsler (status 3/4); West Germany using Horn-Ravens (status 4), and Wechsler (status 3/4); Switzerland using Wechsler (status 4); Austria using Wechsler (status 4); and France using Ravens (status 3), and Verbal-Math (status 3).

For exact results including all rates of gain per country and ages tested, see Table 15 in the article.

Overall, the gains for most countries (discounting the most extreme cases and France a little bit), ranged from 5 to 25 points with a median of 15 points (1 standard deviation), which indicates quite massive gains (even in the countries that only have weak datasets these gains are present, their weak datasets do not indicate that they lack gains, only that there is no reliable estimate of the magnitude of these gains). However, even though the present generation shows these massive IQ gains, this cannot be generalized to the next generation as it is not certain that they will show massive gains as well.

Conclusions / Implications

Learned content and IQ gains

The conclusion regarding learned content and IQ gains is that learned content has not promoted, but inhibited, IQ gains. This applies only to the present generation and there are exceptions. This conclusion was supported by strong data.

Early maturation and IQ gains

The conclusion regarding early maturation and IQ gains is that in the Netherlands, Norway, and Belgium, IQ gains persist to full maturity. It seems that they may persist in Great Britain, the United States, and France. This conclusion is supported by strong data.

Aging and IQ gains

The conclusion regarding aging on IQ gains as a measure of effect, is that cross-sectional data is questionable. This conclusion is a logical consequence of the previous two conclusions, but is also strong in itself.

What does an IQ test actually measure

The Ravens test doesn’t measure intelligence: it seems to measure a correlate with a weak causal link to intelligence. This may be the case with all IQ tests. Thus, the conclusion is that the Ravens test doesn’t measure IQ and that this is perhaps also the case for all of the other test testing IQ.

What are the causes of large IQ gains

The majority of the IQ gains within the Dutch are caused by unidentified environmental variables. It seems that these unidentified variables could account for up until 15 points in the present generation. The conclusion is that familiar environmental variables are likely to play a significant role in IQ gains.

Between-groups differences and gains on IQ tests

Between-groups differences in scores on the tests (thus between the older and the present generation), may not be equal to differences in intelligence. The tentative conclusion made is that IQ differences can’t be used to show that there are no between-groups those differences.

The decline of Scholastic Aptitude Test (SAT) score and the gain in IQ

Diminishing SAT scores over time are compatible with the Wechsler-Benet IQ gain over the same period. The tentative conclusion is that Wechsler-Binet IQ gains over time can’t be related to gains in intelligence.

Summary of the abovementioned implications

The gains that the present generation has made on a variety of IQ tests has been massive, and these gains continue during the course of one’s maturation. Thus, cross-sectional data are questionable to use in order to determine the effects of aging on IQ. The gains in question indicate that IQ tests don’t measure intelligence as is generally thought, but they seem to be a correlate with only a weak/ small causal link to intelligence. For this reason, between-groups IQ differences can’t, at this moment in time, be associated with intelligence differences. The environmental factors that may have a strong influence on IQ have not yet been determined. The hypothesis that the gains on the Wechsler-Binet IQ do not seem to represent an increase in intelligence indicates how these gains can be reconciled with the decline in SAT score.