Average IQ of Haiti

65-70

TL;DR

After a search, I found a large sample (1034) of students who took the Raven’s and scored an IQ of 66, a sample of 150 rural children who scored an IQ of 98, and a sample of 15 Haitian-Americans who scored a 74 on the Weschler, but a 98 on the PPVT. On the early-grade learning assessments (EGRA), two Haitian samples scored an IQ of 66 and 70.

There are various ways this data can be aggregated:

• Arithmetic mean of all 6 studies: 78.6

• Median of all 6 studies: 70

• Weighted mean of all 6 studies (assuming sample sizes of 150 for the EGRA studies): 69.7

Given the fact that it is very unlikely that rural Haitian children have an IQ of 98, and that the sample of 15 Haitian-Americans should be ignored, the average IQ of Haiti is probably about 65-70.

Human Varieties IQ data

Jason Malloy has already done a review of Haiti’s national IQ. The first study they present is a sample of 1034 students from 54 schools which participated in a US aid program. They took the Raven’s coloured progressive matrices and scored the equivalent of a 69 on British norms:

The first reference is a 200+ page government report, examining the cognitive impact of an ongoing US aid program, which provides food for children in Haitian schools (Cotten, 1985). The sampling is excellent: 1,034 students from 54 schools (p. 2). The average sample age is 10 (p. 26). The ability test is the Raven’s Coloured Progressive Matrices; the scores are reported as the survey child’s raw score divided by the standardized age-level norm (p. 48). For 518 children participating in the feeding program, the reported score is 56 (Annex G). The average score for 10 year olds in the norm sample was 27, so this suggests a Raven score of 15 in the Haitian sample. This would correspond to an IQ score of 69. However, this test was also administered 16 years later than the UK comparison group. If IQ scores were now 5 points higher in the UK due to the Flynn Effect, then the scores in the Haitian group would actually correspond to an IQ of 64. Scores were only fractionally lower for the 515 children not participating in the feeding program (Annex G). So this study suggests an IQ of 64 for 1,033 Haitian school children in 1981.

I checked the text and was able to verify most of these claims. Visually speaking, the reported mean of 56 seems defensible (see 5th column from the right).

The ages of the children as well as number of schools and children were reported accurately.

The surveys for measuring impact were done in 27 program and 27 non-program (control) schools in the Department of the West, which includes metropolitan Port-au-Prince. The total impact survey sample included 54 schools and 1,034 children

[…]

The Flynn Effect correction here is fine; IQ scores increased by .17 points per year in Britain during this time period, implying that the score of 69 must be corrected to a 66 rather than the 64 that Jason gave.

Questioning the quality of the data after seeing this result would not be out of line. But the scores the children were assigned do have predictive validity, as they correlate at .4 with parental SES. In the United States, the correlation between parental SES and IQ is about .45. This of course does not rule out temporal/national bias in tests, but it does imply that the rank order of the scores the children are given is accurate.

One of the most consistently significant relationships, regardless of the level of aggregation, was that between cognitive performance and the home environment. Table IV C shows the correlations and significance levels of these relationships for a variety of categories. Taking all 54 cases, i.e., program and nonprogram schools combined, the correlation (r = 0.40) was highly significant (a = 0.001).

Jason brings up a second study that finds an average IQ of 98. Rather implausible, given the circumstances, which makes me think that they used national norms. Unfortunately, I cannot read French, so I am out of my depth here.

The second reference is a French language report from a Haitian research institute (de Ronceray & Petit-Frere, 1975). This study was performed at roughly the same time as the USAID study reported above and uses the same test (RCPM), however the results are anomalous and contradictory. 150 rural children, age 6, were selected from three towns in the Leogane region. Their parents were farmers who never attended school. The sample was divided into three groups based on school instructional language (French vs Creole vs flexible). Test data was collected in 1974 and reported for 133 children, yielding an average Raven score of 15 (p. 14). This corresponds to an IQ of 98. This IQ is implausible; so look through the study and let me know what you think.

He also brought up another French study that found two vastly different IQ scores depending on the test, though that is to be expected, as only 15 people were tested. And they’re immigrants living in Boston, not natives.

The third reference, a French language book chapter, describes an investigation into the school problems of 15 Haitian immigrants in Boston, Massachusetts (Douyon, 1982). The average age was 13, and the average stay in the US was 2 years. Only three of the children had parents with professional or semi-professional occupations. Children were tested with Wechsler intelligence scales and the Peabody Picture Vocabulary Test. Full-scale Wechsler IQ was 74 with a SD of 15.5. However, the PPVT yielded an IQ of 98 (SD 15.9).

Unfortunately, when I tried to find studies on the average IQ in Haiti, I found two studies that didn’t report the average score in their data or studies that Jason already found.

Scholastic evidence

From the harmonized learning outcomes dataset (where the British mean would be roughly 500, and global SD is 100), I find two entries for Haiti. In both of these samples, Haitians were assessed using an early grade reading assessment, which tests young children in their reading ability.

An Early Grade Reading Assessment (EGRA) is a test students take that can measure their skill at both pre-reading and reading subtasks. Usually, EGRAs are given to students in Kindergarten through primary school. EGRAs test children’s skill at different subtasks they need to learn, such as letter names and letter sounds, to be able to read fluently. The test is typically administered by a teacher one-on-one with a student, out loud. Watch the videos below to see EGRA in action in Nepal.

The first sample scored at 270.8 (IQ of 65.6) and the second scored at 298.46 (IQ of 70).