HVGIQ: Cuba

There are several intelligence studies for Cubans, including at least two Raven standardizations, some international assessment data, and a few different test measures for Cuban Americans. The message of all these test results is complicated and leaves us with a few puzzles.

IQ and the Wealth of Nations (2002) and IQ and Global Inequality (2006) both list one intelligence study for Cuba: a large Standard Progressive Matrices standardization (Alonso Garcia, 1973 ). In another anomaly, however, this reference has disappeared from the latest version of the dataset (Lynn & Vanhanen, 2012). Cuba is missing from their list of ‘National IQs’ (p. 391), and the reference is not in the bibliography. But this omission is clearly a mistake, since one table (p. 22) still features a ‘Measured IQ’ score for Cuba—85—the same score paired with this study in the previous books.

Unfortunately, this score seems to be incorrect, as are a number of other things that they report for this study … starting with the citation information, which contains several errors (e.g. misspelling of title; wrong year). The full sample size is larger than reported (1,461 instead of 1,144). The age range is somewhat wider than reported (12-20 instead of 12-18). More importantly, Lynn & Vanhanen (2002) state that these norms were collected in 1971 (p. 202). But this date is not found in the study, and the date reported in one table suggests that the norms were actually collected ten years earlier, in 1961 (p. 69). Since the UK comparison norms are from 1979, L+V add 2 IQ points to their raw score conversion (an 8 year Flynn adjustment), but the 1961 administration requires 5 IQ points (an 18 year adjustment).

Alonso Garcia (1973 ) collected these norms on a representative sample of high school and jr. high school students in the Camaguey province. The average age of the sample was 15.4 years, and the average SPM score was 38.9. On the 1979 UK norms this translates to an IQ of 89.

Alonso Garcia also provided some evidence of test validity by looking at the association of Raven ability with academic performance. The correlation between Raven and marks in Spanish was 0.54, and the correlation between Raven and marks in mathematics was 0.56.

This study is relatively old, but the good news is that there is also a more recent standardization for the SPM in Cuba (Valcarcel et al, 2000 ).

This team collected test data for 589 people, ages 11-68, including high school students, jr. high school students, college students, and working professionals.

The average age of the sample was 26, and the average Raven score was 46. This translates to an IQ score of 91 on the UK norms. The average IQ for the college students was 101, and the average IQ for the professionals was 103. This study also shows a 3 point advantage for females compared with males. This contrasts with Alonso Garcia (1973 ), which shows no sex difference.

An additional advantage of this more recent study is that we can compare it with the data from the 1961 standardization to answer questions about the Flynn Effect in Cuba.

Both studies provide data for different age groups. Alonso Garcia (1973 ) provides SPM norm data for 18 year olds born in 1941, and Valcarcel et al (2000 ) provides SPM norm data for 18 year olds born in 1982. Raven scores increase from 42 to 47 between the 1940 and 1980 birth cohorts, suggesting an increase of 11 IQ points over 40 years. This is a Flynn Effect of 3.6 IQ points per decade, which is similar to the 3 points per decade experienced by the UK and other nations.

People from the 1940s cohort are also tested in both studies (first as 20 year olds, then as 60 year olds), and the raw scores for this cohort are similar in both studies, as expected. The Flynn Effect occurs between different birth cohorts, while scores within birth cohorts remain relatively stable over the adult life span.

A weighted average of the two Raven standardizations gives us an IQ of 90 for Cuba.

International Assessment Tests

Another possible source of IQ data is international assessment tests, like PISA and TIMSS, which more or less capture the same abilities as intelligence tests (Rindermann, 2007). Cuba has not participated in these assessments, however in 1996, UNESCO organized a similar regional assessment for 3rd and 4th grade students in 13 Latin America nations: The First International Comparative Study. This was followed by the Second Regional Comparative and Explanatory Study in 2006.

Cuba participated in these tests and outscored all the other participating nations by 1 to 2 standard deviations!

Some nations have taken these Latin American assessment tests, as well as the TIMSS and/or PISA tests (e.g. Mexico). Altinok & Murseli (2007) used these scores to make a common scale for comparing nations that have not taken the same assessment test. Their international scores can then be converted into deviation IQs, and compared with the UK (the “Greenwich IQ”). This gives Cuba an IQ of 101. In fact this might be an underestimate. Using a similar method, I put the scores from the First Comparative study (Casassus et al, 2000 ) on an international scale, and this suggested a Cuban IQ of 105. Similarly, Chuck put the scores from the more recent Second Comparative Study (Trevino et al, 2010 ) on a common scale and estimated a Cuban IQ of 109!

These estimates are not accordant with the Raven data reported above. One difference is the test type, and some will argue that these achievement tests, which measure math and language skills, etc, are more reflective of input variables like schooling quality, in contrast to intelligence tests which measure more fundamental abilities, and are less radically influenced by external conditions. (At least one book seeks to explain Cuba’s advantage on these tests by reference to policy variables.) A more cynical person, however, might notice that international rankings can serve as a propaganda tool for non-Democratic governments. This invites a certain skepticism when one-party states like Cuba, North Korea, and China are intermediaries to statistics which are unusually impressive or inconsistent with results shown by multiple, independent research teams.

This is not to say that I will exclude these UNESCO studies, or even these Cuban scores, from the dataset.

Cuban-Americans

There are also several different sources of data for ethnic Cubans in the USA. There is a Wechsler intelligence scale standardization for Cubans in Dade County, Florida. This would be the best source of data, but I haven’t obtained this study yet. At least three large US sociological surveys have IQ-type test data for Cuban-Americans:

1) The General Social Survey provides WORDSUM scores for 25 Cubans. The average IQ is 95. The Cubans also identify themselves by race. The 17 Cubans self-identified as white have an IQ of 99. The 8 Cubans self-identified as black or “other” have an IQ of 82.

2) The Add Health dataset has Peabody Picture Vocabulary Test scores for 44 adolescent Cubans (administered 1994-95). The average IQ is 97. This dataset also identifies participants by race, although it shows the opposite pattern. The 29 Cubans identified as white scored 97, but the 16 Cubans identified as black or “other” have an IQ of 100.

3) The National Longitudinal Survey of Youth has Armed Forces Qualification Test data for 94 adult Cubans (administered 1980-81). The average IQ is 93.

A weighted average of the three US surveys gives us an IQ of 94 for Cuban-Americans.

Cuban-Americans have been also tested by the National Assessment of Educational Progress (NAEP), however the exact sample sizes are not reported through the NAEP Data Explorer. I looked at math and reading scores for 4th, 8th, and 12th grade. Cuban Americans do surprisingly bad on this assessment measure. Converted into deviation IQs, these tests suggest white Cuban-Americans have an IQ of 91, while black Cuban-Americans have an IQ of 81 (non-Cuban black Americans have an NAEP IQ of 87). This means there is still a sizable black-white test score gap among Cuban-Americans, but this gap appears slightly smaller among Cubans than among non Cubans. Of course this does not allow us to assert there is BW score gap within Cuba. I do not have data to demonstrate this. (Later posts will discuss available social science statistics for different racial groups in Cuba and other nations).

There appears to be an incongruity between achievement test and intelligence test performance among Cuban citizens and their Cuban-American cousins. Cuban school children did much better on math and reading tests than they did on the Raven’s intelligence test, while Cuban-American school children did worse on math and reading tests than they did on several intelligence tests.

REFERENCES

Valcarcel, E.C., Martinez, E.C., Bouza, C., & Solozabal, T.P. (2000). Algunas características y posibilidades del test de matrices progresivas de Raven. Revista Cubana de Psicología, 17, 95-105.