ArticlePDF Available

The Average IQ of Sub-Saharan Africans: Comments on Wicherts, Dolan, and Van der Maas

Authors:

Abstract and Figures

Wicherts, Dolan, and van der Maas (2009) contend that the average IQ of sub-Saharan Africans is about 80. A critical evaluation of the studies presented by WDM shows that many of these are based on unrepresentative elite samples. We show that studies of 29 acceptably representative samples on tests other than the Progressive Matrices give a sub-Saharan Africa IQ of 69; studies of the most satisfactory representative samples on the Standard Progressive Matrices give an IQ of 66; studies of 23 acceptably representative samples on the Colored Progressive Matrices give an IQ of 71. The international studies of mathematics, science, and reading give a sub-Saharan African IQ of 66. The four data sets can be averaged to give an IQ of 68 as the best reading of the IQ in sub-Saharan Africa.
Content may be subject to copyright.
The average IQ of sub-Saharan Africans: Comments on Wicherts, Dolan, and
van der Maas
Richard Lynn
a,
, Gerhard Meisenberg
b
a
University of Ulster, Coleraine, Northern Ireland, United Kingdom
b
Ross University, Dominica
article info abstract
Article history:
Received 2 July 2009
Received in revised form 22 September 2009
Accepted 25 September 2009
Available online 21 October 2009
Wicherts, Dolan, and van der Maas (2009) contend that the average IQ of sub-Saharan Africans
is about 80. A critical evaluation of the studies presented by WDM shows that many of these are
based on unrepresentative elite samples. We show that studies of 29 acceptably representative
samples on tests other than the Progressive Matrices give a sub-Saharan Africa IQ of 69; studies
of the most satisfactory representative samples on the Standard Progressive Matrices give an IQ
of 66; studies of 23 acceptably representative samples on the Colored Progressive Matrices give
an IQ of 71. The international studies of mathematics, science, and reading give a sub-Saharan
African IQ of 66. The four data sets can be averaged to give an IQ of 68 as the best reading of the
IQ in sub-Saharan Africa.
© 2009 Published by Elsevier Inc.
Keywords:
IQ
School assessments
Math
Science
Reading
Sub-Saharan Africa
Contents
1. Introduction ...................................................... 21
2. IQ studies other than the Progressive Matrices ...................................... 22
3. The Progressive Matrices ................................................ 24
4. Attainment in mathematics, science, and reading ..................................... 25
4.1. Third International Mathematics and Science Study (TIMSS) 19952007 . . . .................... 25
4.2. PIRLS Reading, 2006 ............................................... 26
4.3. 1991/92 Reading Literacy Assessment of the IEA .................................. 26
4.4. IAEP Mathematics Study 1990/91 ......................................... 26
4.5. Second International Science Study (SISS), 1983/84 ................................ 27
4.6. Second International Mathematics Study (SIMS), 1981 ............................... 27
5. Conclusion ....................................................... 28
References .......................................................... 28
1. Introduction
Wicherts, Dolan, and van der Maas (in press) (WDM) and
Wicherts, Dolan, Carlson and van der Maas (in press)
(WDCM) contend that the average IQ of sub-Saharan Africans
is about 80 and that the international studies of achievement
in mathematics, science, and reading corroborate this
conclusion. These estimates are much higher than the
average of 67 proposed by Lynn and Vanhanen (2006).We
consider four data sets on this question consisting of (1) tests
other than the Progressive Matrices; (2) the Standard
Intelligence 38 (2010) 2129
Corresponding author.
E-mail address: lynnr540@aol.com (G. Meisenberg).
0160-2896/$ see front matter © 2009 Published by Elsevier Inc.
doi:10.1016/j.intell.2009.09.009
Contents lists available at ScienceDirect
Intelligence
Progressive Matrices; (3) the Colored Progressive Matrices;
(4) the international studies of mathematics, science, and
reading.
2. IQ studies other than the Progressive Matrices
WDM propose nine inclusion criteria for the acceptability
of studies of the IQ in sub-Saharan Africa, but these do not
include the crucial criterion that the African samples should
be representative of the population. This is a strict criterion
because there are no perfectly representative samples from
sub-Saharan Africa. We therefore have to make judgments on
which studies are sufciently representative to use. We do
not follow WDM in rejecting studies in which (1) Test
administration should not be described as problematic
because this means that the samples lacked the cognitive
ability to understand the instructions and/or the test was
too difcult; (2) sub-Saharan Africans are compared with
matched whites, because in some studies this comparison is
more appropriate.
In this section we summarize studies we consider suf-
ciently representative for inclusion. IQs are corrected for a
Flynn effect of 3 IQ points a decade and calculated in relation to
a British IQ of 100 (sd 15), and deduction of 2 IQ points from
IQs calculated on American norms. These are described as FE/
British IQ adjusted.
Ani and Grantham-McGregor (1998). WDM give an IQ of
73.2 for this Nigerian study of boys tested with the similarities
subtest of the WISC-R, but reject it. Comment: The similarities
subtest is a good measure of reasoning.
Avenant (1988). WDM calculate a WAIS-R IQ of 76 for this
study of South African prison wardens described by Nell (2000,
p.27) as competent men, all in long-standing employment in a
sophisticated environment. They reject the study because the
wording of some items was changed to facilitate understand-
ing. Comment: These changes wouldimprove performance and
should not invalidate the results. Wicherts (2009) has
acknowledged that he has made a mistake in calculating the
IQ and that the correct gure for the IQ is 70.
Badri (1965a). WDM give an IQ of 61.9 for this Sudanese
sample of Negroid boys but reject the study because they
were culturally deprived. Comment: The sample is de-
scribed as representative of the people of this region of
southern Sudan.
Bardet, Moreigne and Sénécal (1960). WDM's IQ of 66.3
for this sample of Senegalese school children is accepted.
Boivin et al. (1993). WDM report an IQ 66.6 for a sample of
healthy children in the Congo but reject this study on the
grounds that it was from an underdeveloped rural area and
the test was adapted. Comment: Most of the Congo consists of
underdeveloped rural areas, so the sample should be
reasonably representative; the authors are condent that
the K-ABC is suitable for use with their sample because they
write this measure was especially intended for use with
nonverbal and non-English speaking children.The battery as
a whole has a good record in terms of construct validity
when adapted for other cultural groups(p. 221).
Boivin, Giordani and Bornefeld (1995). WDM give an IQ of
64.9 for this sample of children in the Congo but reject this
study because the children had intestinal parasites. We are
unable to nd any reference to intestinal parasites in this
paper, and the Boivin et al. (1993) study found that parasitic
and malarial infections had only a marginal (2 points)
adverse effect on the IQ.
Buj (1981). WDM give an IQ of 75.7 for this Ghanaian adult
sample. Comment: This is only marginally acceptable because
the sample came from the capital city and likely has a higher
IQ than the population, and no information was given about
the way the sample was recruited for the study.
Conant, Fasternau, Giordani, Opel and Nseyila (1999). This
study reports a K-ABC IQ of 65 for this sample of children at
school in the Congo. WDM do not list this study. The FE/
British IQ adjusted is 59.
Dent (1937). WDM calculate an IQ of 68 for this sample of
children in South Africa tested with Koh's Blocks but they
reject the study. Comment: Koh's Blocks is similar to the
Wechsler block design subtest which provides a good
measure of full scale IQ (Wechsler, 1974). The sample appears
representative.
Fahmy (1964). WDM report an IQ of 84.3 for this sample of
Nilotic school children in southern Sudan. Comment: Four
tests were administered but WDM only use the one that gave
the highest IQ. The average of the four tests is 74.5, FE/ British
IQ adjusted to 69.
Fahrmeier (1975). WDM calculate an IQ of 72.3 for this
sample of Nigerian children tested with the spatial relations
subtest of the Primary Mental Abilities (PMA), but reject the
study. Comment: The spatial relations subtest is acceptable as
a measure of IQ and the sample appears to be representative.
Ferron (1965). This study reports IQs for seven samples of
children in Nigeria and Sierra Leone that included three
grammar school samples (IQs 91, 95, and 81), two samples
taking the entrance exam to grammar schools (IQs given as
80+ and 70+), and two samples from two primary schools.
WDM average the seven samples to give an IQ of around 77.
Comment: The rst ve samples should be excluded because
grammar school students are selected for higher IQs and the
imprecise IQs given for those taking the entrance exams to
grammar schools are unusable. Ferron's two samples from
primary schools are acceptable. The IQs of these were 74 and
66, averaged to 70.
Fick (1929). This study reports a non-verbal AAB IQ of 65
for children at school in South Africa in relation to an IQ of 100
of the white children. WDM reject this study on the grounds
that the test used the concept of mental age and the native
does not grow up with pictures and diagrammatic represen-
tations of things. Comment: Since the black children were at
school, they should have had adequate experience of pictures
and diagrammatic representations. The AAB gave accurate
results in the United States where blacks in the World War I
military draft obtained an IQ of 83 compared with 100 for
whites. This difference has been conrmed in numerous
subsequent studies including the military drafts in World
War II and the Korean War (Loehlin, Lindzey & Spuhler,
1975), and up to the present day, showing that the test gives
valid results (Rushton & Jensen, 2005).
Giordani, Boivin, Opel, Dia Nseyila and Lauer (1996).
WDM do not list this study of 130 healthy and apparently
representative children in the Congo who obtained a K-ABC
IQ of 65.
Haward and Roland (1954). WDM do not list this study of
a typical cross-sectionof 30 adult Nigerians tested with the
22 R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
DAM. Special weighting was given to compensate for the
penalization of lack of clothingand scores were compared
with a matched British sample (n= 100) in relation to which
the Nigerians obtained an IQ of 67.
Holding et al. (2004). WDM calculate a K-ABC IQ of 63 for
this sample of Kenyan children but reject the study on the
grounds that the K-ABC was adapted and children had had
malaria. Comment: The alterations made in the adaptation of
the test were trivial and consisted of replacing less familiar
with more familiar pictures. The effect of these alterations
would have been to improve the performance of the Kenyan
children. The authors note similar procedural modications
made by other investigators have yielded enhancements in
test performance relative to unmodied items(p.252). The
authors state that their results demonstrate the utility of a
cross cultural adaptation of the K-ABC and of other tests not
originally designed for the children of the region(p.252) and
that all children in our region have malaria(p.249). Hence,
it is considered that the sample can be regarded as
representative.
Hunkin (1950). WDM calculate a DAM IQ of 74.5 for this
large (n=2300) urban South African sample of 613 year
olds. Comment: As the sample is urban it likely gives an
inated IQ. Nevertheless, the sample is considered sufciently
representative to be acceptable.
Klein, Pohl and Ndagijimana (2007). WDM calculate an
IQ of 69.6 for this sample of sub-Saharan African adults in
Belgium. Accepted.
Lloyd and Pidgeon (1961). WDM assign this sample of
Zulu children an IQ of 88.7. Comment: The study reports a
Zulu IQ of 86.75 and an IQ of a white South African
comparison sample as 103.2. WDM calculate the Zulu IQ in
relation to British norms on the assumption that the sample is
representative. But the authors state it was not possible to
make a random selection either of children or schools(p.
146). To overcome this problem, considerable care was
taken to choose schools that would yield samples of children
representative of the total population for each racial group.
Hence it is better to calculate the Zulu IQ in relation to the IQ
of the matched white sample. In relation to the white South
African IQ of 100, the Zulu IQ is 83.5. The average sd for the
white and the Zulu samples is 9.35. Adjustment for this
reduces the Zulu IQ to 74.
Lynn and Owen (1994) give an IQ of 68 for blacks in a
South African sample in relation to 100 for whites aged 16.5
and 15 years, respectively, calculated from the South African
Junior Aptitude Test (JAT). WDM give an IQ for this sample
but do not list it in their Table 2. They reject it on the grounds
that the test has been shown to be biased for blacks because a
test of invariance fails. Comment: This is the only study of
sub-Saharan Africans that has been tested for measurement
invariance, so it is unreasonable to reject it on these grounds.
On 9 out of the 10 subtests (except memory) the blacks
obtained IQs below 70. Four of the subtests are verbal and
could have been biased against blacks. However, their IQs are
the same on the verbal and non-verbal subtests. The lower
reliability of the blacks' responses suggests that the main
reason for the invariance failure is that many blacks were
guessing the answers to many of the questions. The results
are likely biased in favor of blacks because they were aged
16.6 years while the whites were aged 15.0 years, and many
of the less able blacks leave school before this age. Finally, the
same samples were tested with the Progressive Matrices on
which WDCM (2009) calculate that the blacks obtained an IQ
of 68. This identical result to that obtained on the JAT
conrms the validity of this test for the blacks in this study.
Minde and Kantor (1976). WDM calculate an IQ of 83.9 for
this Ugandan sample of children. The IQ obtained in this study
is so much higher than that in any of the other 29 studies
listed in this section that it could arguably be rejected as an
outlier.
Murdoch et al. (1994). WDM do not consider this study
that administered 6 subtests of the WISC-R to a sample of
13 year old adolescents at high schools in Johannesburg. The
mean IQ was 79, FE/British IQ adjusted= 70.
Nissen, Machover and Kinder (1935). WDM give an AAB
IQ of 63 for this sample in Guinea but reject the sample on the
grounds that the children were handicapped in the mazes test
by inexperience in manipulating a pencil, and on the
Manikin and Feature Prole test because the subjects
appeared utterly bewildered. Comment: The sample's
performance was the same on these two tests as on the
other tests. WDM also state that the sample consisted of
unschooled test takers, but the authors state that all the
children had attended school and considered the sample as
representative.
Richter, Griesel and Wortley (1989). WDM calculate an IQ
of 74.5 for this sample of South African school children.
Accepted.
Skuy, Schutte, Fridjhon and O'Carroll (2001). This study
reports on two samples of South African Soweto high school
students, both described as representative of the Soweto
high school population(p. 1415). The rst sample took six
subtests from the WISC-R, the DAM, and the WCST (Wiscon-
sin Card Sorting Test). WDM use the WISC-R and the DAM to
give an IQ of 71.6. Comment: WDM did not use the WCST
because We did not consider scores on tests that are not
meant to measure g, such as the (WCST), because this test is
not an IQ test. The WCST manual states that the test was
developed as a measure of abstract reasoning ability
(Kongs, Thompson, Iverson, & Heaton, 2000, p. 1), so the
WCST should be accepted as a measure of intelligence. The IQ
of this sample on the WCST (scored for errors) is 62. The
average of the three tests taken by this sample is an IQ of 68.
Skuy et al.'s (2001) second sample took the full WISC-R
and the WCST. On this occasion WDM accept the WCST as a
measure of intelligence (contrary to their rejection of it in the
paragraphs immediately above and below) and combine it
with the WISC-R performance scale to give an IQ of 74.3.
Sternberg et al. (2002). WDM accept the IQ of 72 given in
Lynn (2006) for this sample of children in Tanzania tested
with the WCST, but they reject this study on the grounds that
the WCST is not a measure of g. Comment: The WCST is a
good measure of gand the sample appears representative.
Sternberg et al. (2001). WDM accept the IQ of 64 given in
Lynn (2006) for this sample of Kenyan children tested with
the MillHill vocabulary test but reject the study because
vocabulary does not measure g. Comment: Vocabulary is an
excellent measure of g; for instance, vocabulary has the
highest correlation (0.74) of all the Wechsler subtests with
the full scale WISC-R (Wechsler, 1974, p.47). The sample
appears representative.
23R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
Zindi (1994). WDM calculate an IQ of 71.6 for this
Zimbabwean sample of high school children. Comment:
Wicherts (2009) has acknowledged that the correct gure is
70.7.
This section has summarized the results of 29 studies of
acceptably representative samples. The median IQ is 69.
In this section we summarize studies we consider insuf-
ciently representative for inclusion.
Ashem and Janes (1978). WDM calculate an IQ of 88.8 for
this sample of 4 year old Nigerian children. Comment: The
sample consisted of well nourished higher socio-economic
children(1Q=109), adequately nourished mainly middle
class children (IQ = 91.4), and poorly nourished rural
children (IQ= 79.6). WDM average the three results, but
the combined sample cannot be accepted as representative.
Furthermore, IQs for 4 year olds are unsatisfactory because
many American studies have shown that the black-white IQ
gap is relatively small at the age of 4 years but increases as
children grow older (Jensen, 1974).
Badri (1965b). WDM give an IQ of 73.1 for this Sudanese
sample. Comment: The population of central and northern
Sudan is predominantly North African Caucasoid (Cavalli-
Sforza, Menozzi, & Piazza, 1994, p.169).
Bakare (1972). WDM calculate an IQ of 83.1 for this
sample of Nigerian upper-class and lower-class children.
Comment: Fathers in the upper class homes were senior civil
servants or university administrators, lecturers or professors
(p. 356). Their children cannot be averaged with those of
lower-class children to give a representative sample.
Claassen, Krynauw, Paterson and Mathe (2001).WDM
calculate an IQ of 83.1 for this sample of English-speaking
black South Africans, in relation to 100 for white South
Africans. Comment: This was a well educated sample, and
English-speaking black South Africans have a higher IQ than
native language speakers (e.g. Shuttleworth Edwards et al.,
2004), so the sample cannot be accepted as representative.
Khaleefa,Abdelwahid,AbdulradiandLynn(2008).
These Sudanese samples are predominantly North African
Caucasoids.
Ohuche and Ohuche (1973). WDM calculate an IQ of 91.3
for this sample of children at an experimental school in Sierra
Leone. Comment: The sample is unrepresentative; the ages of
the children are unknown; all children repeating the year
(i.e. those with low IQs) were excluded; the IQ of 56 year
olds=69.5; the IQ of 712 year olds = 94.2, a discrepancy
indicating serious problems with the data; there was no
correlation between IQs and tests of English, math and social
science in grades 47, showing IQs have no validity for these
ages. The study is so unsatisfactory it has to be rejected.
Shuttleworth Edwards et al. (2004). WDM give an average
IQ of 94 on the US WAIS-III for this South African sample of
40 educated adults. Comment: The sample consists of those
who have been educated to age 18. The authors write it is not
representative of the population(p. 915), so it has to be
rejected.
Wilson, Mundy-Castle and Sibanda (1991). WDM calcu-
late an IQ of 86.2 for this Zimbabwean sample. Comment: This
sample attended a primary school with whites in a middle
class neighborhood and cannot be accepted as representative.
Other studies not listed above are rejected as unacceptable
for the reasons given by WDM.
3. The Progressive Matrices
Studies of the Progressive Matrices also need to be con-
sidered to arrive at an estimate of the IQ in sub-Saharan
Africa. For the Standard Progressive Matrices, Wicherts et al.
(in press) (WDCM) give results for 40 studies, for which the
median IQ is 78, Flynn effect corrected to 77, and reduced
further to 76 to adjust for around 20% of Africans who do not
attend school and are credited with an IQ of 71.
A number of these studies have to be rejected as based on
clearly unrepresentative samples. These include ve samples
of university students; Crawford Nutt's (1976) sample of high
school students (IQ 84) in math classes admission to which
is dependent on the degree of excellence of the pupil's
performance in the lower classes(p. 202) and described as a
select segment of the population(p. 204), and who were
coached on how to do the test; a sample of psychiatric
patients (IQ 86) because these had to pay fees, would have
been higher SES and are not a representative sample
(Morakinyo, 1985); a selected sample of technical college
students (IQ 79) in Zambia (MacArthur, Irvine, & Brimble,
1964); and a sample of Madagascans in France (IQ 82)
(Raveau, Elster, & Lecoutre, 1976) because these are not pure
Negroids but a mixed race people with substantial South East
Asian ancestry.
We are not able to discuss these 40 studies in detail.
Instead, we consider the two most satisfactory studies. Owen
(1992) compared representative samples of 1093 blacks (age
16.5 years) and 1056 whites (age 15 years). Owen's analyses
show that at the item level the test is not biased against blacks
in terms of any of the standard statistical criteria for the
detection of bias. There are no British norms for 16.5 year
olds, but the 1979 norms for 15.5 year olds can be used. The
score is at the 2nd percentile of these = 69 IQ. Deduct 2 for
FE= 67.
The second of the most satisfactory studies is Vass's
(1992) data for a representative sample of Xhosa-speaking
South African secondary school students (n=711, mean
age=19.3 years). The mean score was 32.9 (sd =9.72), well
below the British 5th percentile. Extrapolating the British
norms downwards, the Xhosa scored at about the rst
percentile of the British norms (IQ 65). The average of these
two studies is an IQ of 66.
WDCM summarize 16 studies of the Colored Progressive
Matrices for which they give a median IQ of 78. The highest
IQ of these samples is 96 for children at a private fee paying
school in Nigeria and is evidently an elite sample that should
be rejected. WDCM only count results for children aged 511
on the grounds that the CPM is too easy for those aged over
11 years and the African IQ is reduced by ceiling effects. This
cannot be accepted because several studies have shown that
Africans above the age of 11 obtain no-where near the full
score. For instance, Knoetze, Bass and Steele (2005) report a
standardi zation of the CPM in South Afric a on a sample of 379
school students aged 7 to 18 and found that none of them
obtained the full score. Heyneman and Jamison (1980)
report a mean score of 23.1 out of a possible total of 36 on the
CPM for 13 year olds in Uganda. The sd is 3.2, suggesting that
none of the sample can have achieved near the maximum
score. In Fahrmeier's (1975) sample of 1213 year olds in
Nigeria the mean score was 15.1. In Heady's (2003) sample of
24 R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
918 year olds in Ghana, the highest score was 24.3 obtained
by the 17 year olds in school. All these scores are well below
the maximum of 36 and show no ceiling effect.
WDCM's exclusion of results on samples aged over 11
inates the African IQ. For instance, WDCM give an IQ of 73
for 172 children aged 711 in Knoetze, Bass and Steele's
(2005) study, but for the complete sample (n=379) in-
cluding older children the IQ is 71. WDCM's elimination
strategy excludes eight CPM studies of adolescents listed in
Lynn (2006). The IQs derived from these studies range
from 62 to 69. If we add these to the 15 acceptable studies
reported by WDCM, there is a total of 23 studies with a
median IQ of 71.
4. Attainment in mathematics, science, and reading
We agree with WDM that international school assess-
ments in mathematics, science, and reading are measures of
attained intellectual competence and can be adopted as
proxies for IQs. Country level correlations between school
assessment results and IQs are around 0.92 (Lynn, Meisen-
berg, Mikk, & Williams, 2007). However, we contend that
WDM have used an inappropriate method for calculating IQs
of sub-Saharan Africans from the international school assess-
ments. Because the more recent assessment programs are
methodologically the most advanced, they will be discussed
rst.
4.1. Third International Mathematics and Science Study (TIMSS)
19952007
The TIMSS assessments were organized by the IEA
(International Association for the Evaluation of Educational
Achievement) in 1995, 1999, 2003 and 2007. Three African
countries participated in the 8th-grade assessments: Bots-
wana and Ghana in 2003 and 2007, and South Africa in 1995,
1999 and 2003. Ghana has a measured IQ of 71, and South
Africa of 72. Botswana has no measured IQ. L&V offer an
estimate of 70.
TIMSSresultsareavailableathttp://timss.bc.edu/
timss2003.html and http://nces.ed.gov/timss/tables07.asp,
scaled to a meanof 500 and individual-level standard deviation
of 100 for thecountries participating in the 1995study (see also
Martin,Mullis, Gonzales, & Chrostowski,2004; Martin, Mullis,&
Foy, 2008; Mullis, Martin, Gonzales, & Chrostowski, 2004;
Mullis, Martin, & Foy, 2008).The TIMSS Technical Reports show
within-country standard deviations of about 85 (e.g., Gonzalez,
Galia, Arora, Erberber, & Diaconu, 2004).
We have averaged the mathematics and science scores of
8th grade students separately for each of the four assess-
ments, and applied a minor trend adjustment based on the
means for the 18 countries participating in all four assess-
ments. We have averaged these scores to produce a single
TIMSS score for each country. The scores for the three African
countries (Botswana, Ghana, South Africa) are shown in
Table 1. Three non-African countries scored lower than
Botswana: Qatar (306.9), Morocco (350.3), and the Philip-
pines (350.8). The correlation of the TIMSS score with
measured IQ is 0.88 for the 57 countries for which both
measures are available.
The TIMSS score was converted into a school achieve-
ment IQ(SAIQ) by three alternative methods:
1. Following the method of WDM, the SAIQ was calculated by
linear regression with TIMSS score and IQ for the 55 non-
African countries having both measures:
IQ =43:04 + 0:109 × TIMSS score:
Table 1 shows the results for the three African countries,
the four non-African countries with the lowest IQs (IQ 7883:
Egypt, Lebanon, Qatar, Syria), the four highest-scoring
countries (Hong Kong, Japan, South Korea, Taiwan), and the
United Kingdom. Ghana and South Africa have slightly higher
SAIQs than expected from their L&V IQs, and the IQ of
Botswana is a remarkable 81.4.
However, linear regression produces biased results be-
cause it systematically reduces the standard deviation in
proportion to the correlation between the variables, as shown
at the bottom of Table 1. It raises the scores for all low-scoring
countries, and reduces those of high-scoring countries.
Therefore we consider linear regression unsuitable for the
rescaling of school achievement results into the IQ metric.
2. The results were rescaled directly, converting the score of
the United Kingdom (503.3) to 100 and the within-country
standard deviation of 85 to 15. This is our preferred
procedure for comparisons of school achievement between
countries. Now the between-country standard deviation is
higher for school achievement than for IQ (10.4 versus 7.3
in the 55 non-African countries). This shows that school
assessments have more cultural biasthan IQ tests. One
explanation that we have advanced before (Lynn et al.,
2007) is that poor quality of schooling in low-scoring
countries depresses school achievement to a greater extent
than it depresses IQ.
With direct-scoring the SAIQs of Ghana and South Africa
are substantially lower than expected from their IQs, whereas
Botswana's is higher than its estimated IQ.
3. The TIMSS scores of the 55 non-African countries that
have measured IQs were rescaled to equal mean and
standard deviation with their IQs. This is our preferred
procedure for scaling school achievement to the IQ metric
because it controls for the handicap that low-scoring
countries in general have on school assessments relative
Table 1
Countries grouped by their average L&V IQ (IQ).
Country L&V IQ TIMSS
score
SAIQ
regression
SAIQ
direct
SAIQ
equalized
IQ105 106 569.0 104.0 109.6 105.0
United Kingdom 100 510.1 98.0 100.0 98.3
Non-African
IQ 7883 81 390.6 84.7 78.9 83.4
Botswana (70) 361.1 81.4 73.7 79.7
Ghana 71 284.3 72.9 60.2 70.2
South Africa 72 275.3 71.9 58.6 69.1
55 non-African
countries
94.5±
7.3
479.5±
58.6
94.6±
6.5
94.6±
10.3
94.5±
7.3
The school achievement IQ (SAIQ) is calculated by linear regression of the
TIMSS scores on the L&V IQs, direct transformation to the 100/15 scale, and
scaling to equal means and standard deviations.
25R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
to their IQ. Table 1 shows that the SAIQs for Ghana and
South Africa are now only marginally lower than their IQs.
However, Botswana's SAIQ of 79.7 is still remarkably high.
A similar exercise for calculating IQs of African countries
from TIMSS and PIRLS data has been published by Rinder-
mann, Sailer and Thompson (2009). They estimate IQs of
73.93 for Botswana, 61.25 for Ghana, and 63.28 for South
Africa. These IQs are a little lower than our estimates cal-
culated by direct transformation scaling.
To put African TIMSS scores in perspective, in 2003, when
the three African countries participated, the average score of
Botswana was at the 5th percentile of the English distribution
in mathematics and well below the 5th percentile in science.
Ghana and South Africa scored a bit below the 25th percentile
of the Botswana distribution in mathematics and near the
15th percentile in science (Martin et al., 2004; Mullis et al.,
2004). Age at testing was generally higher for the African than
the non-African samples. For example, on the 2003 assess-
ment the average age was 14.4 years (range: 13.815.2) for
the non-African samples, 15.1 years for Botswana and South
Africa, and 15.5 years for Ghana. However, age at testing has
only a marginal effect on TIMSS performance.
4.2. PIRLS Reading, 2006
South Africa was the only African country participating in
this assessment of reading literacy in 4th graders in 39
countries. Results are reported on a 500/100 scale. The
within-country standard deviation is about 80 (Kennedy &
Trong, 2007). The highest score (Russia) was 565. The United
Kingdom scored 535, and South Africa obtained the lowest
score with 302. Morocco was second-lowest, with a score of
323. The correlation between PIRLS score and IQ is 0.83 for
the 32 non-African countries that have both measures.
The SAIQ of South Africa is 79.2 with linear regression,
56.3 with direct transformation, and 75.7 with equalization of
mean and standard deviation. The reason for the large
discrepancy between the scores obtained by direct transfor-
mation and by equalization is that for the 32 non-African
countries that have both IQ and PIRLS score, the between-
country standard deviation is far greater for the direct-
transformed SAIQ (12.5) than for IQ (6.7). It appears that 4th-
grade reading is even more culturally biasedthan 8th-grade
mathematics and science.
For the 38 non-African countries the average age at testing
was 10.3 years. For South Africa the average age was
11.9 years, and children were tested in grade 5 rather than
grade 4 (Mullis et al., 2007). Linear regression predicting
PIRLS score with IQ and age showed that each year adds 39.2
points to the PIRLS raw score (95% condence interval: 6.7
71.7 points). Age-correction reduces the South African score
(with equalized means and standard deviations) from 75.7 to
69.1.
4.3. 1991/92 Reading Literacy Assessment of the IEA
This study of 9 and 14 year olds pioneered the methods
that were used later in TIMSS and PIRLS. 30 countries
participated in the assessment of 14 year olds, including
Nigeria, Zimbabwe and Botswana. Grading was done with a
Rasch model, and the results are reported on a 500/100 scale
(Elley, 1992). The within-country standard deviation is about
80. The correlation with IQ is 0.73 for the 25 non-African
countries for which measured IQs are available.
The top score (560) was obtained by Finland, the United
States scored 535, and the three African countries obtained
the lowest scores: 401 for Nigeria, 372 for Zimbabwe, and 330
for Botswana. The lowest non-African scores were 417 for
Venezuela and 430 for the Philippines. Table 2 shows the
results for the three African countries and the average for the
two lowest-scoring non-African countries. A minor correction
for age at testing is included in Table 2. Direct scaling
produces results that are somewhat higher than expected for
Nigeria and Zimbabwe, and lower than expected for Bots-
wana and the two non-African countries. With equalization of
mean and standard deviation, however, SAIQs of Nigeria and
Zimbabwe are up to 9.5 points higher than expected based on
their IQs, although Botswana scores somewhat lower than
expected.
4.4. IAEP Mathematics Study 1990/91
This study assessed mathematics in 13 year olds. 19
countries participated, with Mozambique as the only African
country. The average percent correct is published by Lapointe
(1992). The correlation between the scores of the 18 non-
African countries and their IQs is 0.86.
Direct transformation of raw scores into SAIQ is not
possible because within-country standard deviations are not
published. Based on the 18 non-African countries, the SAIQ of
Mozambique is 85.6 with linear regression and 82.2 with
equalization of means and standard deviations.
This looks impressive, but the raw scores explain the
result. China scored highest, with 80.2% correct answers.
British children got 60.6% correct, and Americans 55.3%.
Mozambique got 28.3%, which was the lowest score. The
second-lowest country, Brazil, scored 34.7%. Multiple-choice
exams usually have 4 or 5 answer choices. Therefore children
who know nothing will score 20% or 25%. Even with
equalization of mean and standard deviation, a raw score of
20% translates into an SAIQ of 77.1. To match their IQ of 64,
Mozambiquan children would have to get 6% correct with
equalization of means and standard deviations, and 17.5%
with linear regression!
Table 2
Results of the three participating African countries in the 1991/92 IEA
Reading Literacy Assessment.
Country IQ Reading
score
SAIQ
regression
SAIQ
direct
SAIQ
equalized
Philippines,
Venezuela
85 423.5 86.5 77.1 82.6
Nigeria 69 401 83.4 72.3 78.5
Zimbabwe 66 372 79.7 66.5 73.6
Botswana (70) 330 75.8 60.2 68.3
African average 68.3 367.7 79.6 66.3 73.5
Venezuela and the Philippines are the lowest-scoring non-African countries
both on IQand reading literacy.SAIQ, school achievement IQs corrected for age.
26 R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
The 1990/91 mathematics assessment discriminated well
between China and the United States, but the test was far too
difcult for children in low-scoring countries. Scaling to an IQ
equivalent is not possible with this raw score distribution.
4.5. Second International Science Study (SISS), 1983/84
This study covered the ages of 10 and 14 years. 23 countries
participated at age 14, including Ghana, Nigeria and Zimbabwe.
The results are published as percent of the questionsanswered
correctly (Keeves, 1992). The top score (Hungary) was 70.7%,
England scored 55.9%, and the United States 54.6%. The lowest
scores were 39.7% (Philippines), 42.2% (Nigeria), 42.8% (Zim-
babwe), and 46.7% (Ghana). The correlation with IQ is 0.44 for
the 20 non-African countries.
With direct transformation of scores, based on the averaged
English and US standard deviations (16.2), the SAIQ is 91.5 for
Ghana, 87.3 for Nigeria and 87.9 for Zimbabwe. With
equalization of means and standard deviations it is 86.2, 81.3
and 82.0, respectively. Thistime the between-country standard
deviation is higher for IQ than for school achievement (6.5
versus 5.5, N= 20 non-African countries).
The average age at testing in the 20 non-African countries
ranged from 14.2 years (England) to 17.1 years (Papua New
Guinea), with an average of 15.1 years. Average age at testing
was 16.1 years in Ghana and Zimbabwe, and 16.2 years in
Nigeria. Nigeria also participated in the assessment of 10-year-
olds. This time Nigeria obtained the lowest score (35.1%),
followed by the Philippines (42.3%).
4.6. Second International Mathematics Study (SIMS), 1981
17 countries participated in this assessment of 13 year
olds, including Nigeria and Swaziland. The raw scores are
published in Medrich and Grifth (1992), separately for
arithmetic (46 questions), algebra (30 questions), geometry
(39 questions), measurement (24 questions), and descriptive
statistics (18 questions).
Scores on the content areas were averaged, weighted by
the number of questions. For the 14 non-African countries
with measured IQ, SIMS score correlated 0.50 with IQ. SAIQ by
linear regression is 93.8 for Nigeria and 93.1 for Swaziland,
and by equalization of means and standard deviations it is
88.5 and 87.1, respectively. Direct transformation of the SIMS
score into the IQ metric is not possible because the within-
country standard deviations are not reported.
Nigerians would have to obtain a raw score of 40.7% to
match their IQ of 69 when linear regression is used, and 4.3%
when means and standard deviations are equalized. As in the
1991 IAEP Mathematics study, the reason is that the test was
too difcult. The top-scoring country, Japan, achieved 62.1%,
the United Kingdom 47.6%, Nigeria 33.6%, and Swaziland
31.5%. It is apparent that the African children knew very little.
Age is another confounding factor in this assessment. The
average age for the 15 non-African countries was 13.8 years
(range 13.0 to 14.1 years), but 15.1 years in Swaziland and
16.1 years in Nigeria.
To summarize, only the more recent studies (IEA Reading
Literacy 1991/92, PIRLS 2006, TIMSS), that are graded with
state-of-the-art methods of item response theory (IRT) that
model student prociency as a latent variable (Foy et al.,
2008), can be scaled easily to the IQ metric. The sub-Saharan
African countries perform poorly in these studies. Direct-
transformed SAIQs in these three assessments range from
56.3 (South Africa, PIRLS 2006) to 73.7 (Botswana, TIMSS),
with an average of 63.5 when taking account of the multiple
participation of countries in TIMSS.
However, all three of the IRT-scaled assessments have a
higher between-country standard deviation, relative to
within-country standard deviation, than IQ. When we control
for this general handicap of low-scoring countries on school
assessments relative to IQ tests by equalizing means and
standard deviations, the school achievement IQs range from
68.3 (Botswana, reading literacy 1991/92) to 79.7 (Botswana,
TIMSS 2003/2007 average).
Using this scaling method and counting each TIMSS assess-
ment separately, the average African school achievement
translates into an SAIQ of 72.4. This is an overestimate of
theAfrican IQ for three reasons. First, participation in
international school assessments is generally restricted to the
economically and educationally more advanced African
countries, such as South Africa and Botswana. IQ and school
performance are likely higher in these countries than in the less
developed African countries. For example, according to the
Human Development Report of the United Nations, in 2002 the
youth literacy rate was 88.6% in Nigeria, whichhas data for both
IQ and school achievement, but only 19.8% in neighboring
Niger, which has no cognitive test data of any kind.
The average of the L&V IQs of these countries, weighted for
their frequency of participation in the school assessments, is
70.5, which is slightly higher than the overall African average.
This includes the estimated IQ of 70 for Botswana, which is
possibly an underestimate considering the advanced state of
the country's economy and educational system.
Second, the TIMSS and PIRLS samples do not represent the
whole population but only those in sc hools. Writing of the early
years of the twenty-rst century, Hanushek and Woessman
(2007, p. 51) note that In West and Central Africa, 59% of each
cohort do not even complete grade 5, and 44% never enroll in
school in the rst place.Those who are in secondary schools
will have above average IQs because typically they have secured
entry bypassing entrance examinations, their parents havepaid
school fees, they have done well and not dropped out by the age
of 15, and because schooling has increased their IQs. Thus,
Garden (1987) writes of the Second International Mathematics
Study: It should be noted that in Swaziland in 1980 19.9% of
1217 year olds were in school.(p. 32). The average age of
children tested in Swaziland was 15.6 years. Schools with high
pass rates on external exams were somewhat overrepresented,
and the author concludes that “… upward bias in achievement
with respect to the population is indicated.(p. 99).
In Nigeria, the assessment was limited to students in state-
owned secondary grammar schools, which prepare for the
West African Sch ool Certicate Examination. Students in trade
schools, technical and other vocational and pre-vocational
institutionswere excluded (Garden, 1987, p. 28). In Nigeria
the assessment was limited to the southern states, which have
50% of the total population but approximately 90% of the
enrolment in secondary grammar/commercial schools. There
are no United Nations and World Bank data about secondary
enrolment in Nigeria at that time, but Garden (1987) writes:
The enrolment rate is low inNigeria and since mathematics is
27R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
compulsory for all students in Nigerian secondary schools it is
apparent that the enrolment rate is much higher for boys than
for girls.(p. 97). 72.8% of the Nigerian sample was male.
The same was observed in the Second International Science
Study. Here, the percentage of children still in school in the older
cohort was reported as 6% in Ghana and 30% in Zimbabwe, with
data for Nigeria unavailable (Keeves, 1992, p. 54). In the
Philippines, whose students scored even lower than the
Africans, 60% were still in school. The good performance of the
African children on this test (relative to the Philippines) is
explained at least in part by the fact that the less capable African
children were no longer in school at the age of testing (about
16 years). At the 10-year age level, however (actually 12.1 years
in Nigeria, compared to 11.1 years in the Philippines and
10.7 years in the remaining 15 countries), 92% of the Nigerian
children were still in school, and they got the lowest score.
The magnitude of the ination of IQs calculated from
TIMSS studies of secondary school students can be estimated
from two studies. Ferron (1965) administered the Leone test
devised by an African for African childrento two primary
school samples (average IQ 70) and a selected secondary
school sample (IQ 81) in Zaria (Nigeria), suggesting that the
secondary school sample inates the African IQ by 11 IQ
points. Heady (2003) found in samples of 1218 year olds in
Ghana that those in school had an IQ 2 points higher than
those not in school. The average of the two results is a 6.5
advantage for those in secondary school. The adoption of this
gure reduces the sub-Saharan African IQ estimated from
TIMSS studies from 72.4 to 65.9, rounded to 66.
A third caveat is the presence of oor effects. Because of
severe oor effects, the older assessments cannot be
translated into the IQ metric. These tests were designed to
discriminate between developed countries. Even for the
otherwise modernIEA Reading Literacy assessment of
1991/92, the report states: “… Botswana, Nigeria and
Zimbabweeach had large numbers of students below the
chance level mark of 25 percent.(Elley, 1992, p. 26, italics
in original). Therefore it is likely that oor effects contributed
to the relatively good performance of the Africans in this
assessment (Table 2). We must not forget that the actual
performance of the African samples (indexed by the direct-
transformed scores) is well below the 2nd percentile of the
British distribution in most cases.
An example for an extreme oor effect is the IAEP
Mathematics assessment of 1991, which produced a score
of 28% for Mozambique. The result is remarkable also because
the sample was drawn from the cities of Beira and Maputo,
omitting children from small towns and rural areas. Generally
in Africa, children in urban and more developed areas
perform better in school than those from backward rural
areas. This has been demonstrated abundantly in the recent
(20012004) SACMEQ assessment of reading and mathemat-
ics in 14 countries of South and East Africa (http://www.
sacmeq.org/indicators.htm). In Mozambique, children in
Maputo City scored 24.5 scaled points (equivalent to about
4 IQ points) higher than the Mozambiquan average.
5. Conclusion
The three IQ data sets show that studies of acceptably
representative samples on tests other than the Progressive
Matrices give a sub-Saharan Africa IQ of 69; studies of the
most satisfactory representative samples on the Standard
Progressive Matrices and on the Colored Progressive Matrices
give IQs of 66 and 71. These results are corroborated by the
international studies of math, science, and reading that give
an IQ of 72.4, adjusted down to 66 because these studies are
based mainly on high school samples in the more advanced
African countries. The average of the four data sets gives an IQ
of 68 and should be regarded as the best reading of the IQ in
sub-Saharan Africa.
References
Ani, C. C., & Grantham-McGregor, S. (1998). Family and personal character-
istics of aggressive Nigerian boys. Journal of Adolescent Health,23,
311317.
Ashem, B., & Janes, M. D. (1978). Deleterious effects of chronic undernutri-
tion on cognitive abilities. Journal of Child Psychology and Psychiatry,19,
2331.
Avenant, T. J. (1988). The establishment of an individual intelligence scale for
adult South Africans. Report on an exploratory study conducted with the
WAIS-R on a Sample of Blacks. (No. P-91) Pretoria, South Africa: Human
Sciences Research Council.
Badri, M. B. (1965). The use of nger drawing in measuring the Goodenough
quotient of culturally deprived Sudanese children. Journal of Psychology,
59, 333334.
Badri, M. B. (1965). Inuence of modernization on Goodenough quotients of
Sudanese children. Perceptual and Motor Skills,20, 931932.
Bakare, C. G. M. (1972). Social-class differences in the performance of
Nigerian children on the Draw-a-Man test. In L. J. Cronbach, & P. J. Drenth
(Eds.), Mental tests and cultural adaptation (pp. 355363). The Hague,
The Netherlands: Mouton.
Bardet, C., Moreigne, F., & Sénécal, J. (1960). Application de test de
Goodenough à des écoliers africains de 7 à 14 ans. Enfance, 199208.
Boivin, M. J., Giordani, B., & Bornefeld, B. (1995). Use of the Tactual
Performance Test for cognitive ability testing with African children.
Neuropsychology,9, 409417.
Boivin, M. J., Giordani, B., Ndanga, K., Maky, M. M., et al. (1993). Effects of
treatment for intestinal parasites and malaria on the cognitive abilities of
schoolchildren in Zaire, Africa. Health Psychology,12, 220226.
Buj, V. (1981). Average IQ values in various European countries. Personality
and Individual Differences,2, 168169.
Cavalli-Sforza, L. L., Menozzi, P., & Piazza, A. (1994). The history and
geography of human genes. Princeton, NJ: Princeton University Press.
Claassen, N. C. W., Krynauw, A. H., Paterson, H., & Mathe, M. W. G. (2001). A
standardisation of the WAIS-III for English-speaking South Africans.
Pretoria, South Africa: Human Sciences Research Council.
Conant, L. L., Fasternau, P. S., Giordani, B., Opel, B., & Nseyila, D. D. (1999).
Modality specicity of memory span tasks among Zairian children.
Journal of Clinical and Experimental Neuropsychology,21, 375384.
Crawford Nutt, D. H. (1976). Are Black scores on Raven's Standard
Progressive Matrices an artifact of method of test presentation? Psy-
chologia Africana,16, 201206.
Dent, G. R. (1937). The educability of the Bantu. In E. G. Malherbe (Ed.),
Educational adaptations in a changing society Capetown, South Africa:
Juta & Co.
Elley, W. B. (1992). How in the world do students read? IEA study of reading
literacy. International Association for the Evaluation of Educational
Achievement, Report No. ISBN-92-9121-002-3 Accessed 15 June 2009 at:
http://eric.ed.gov/ERICDocs/data/ericdocs2sql/content/storage_01/
0000019b/80/13/07/42.pdf
Fahmy, M. (1964). Initial exploring of the intelligence of Shilluk children:
Studies in the Southern Sudan. Vita Humana,7, 164177.
Fahrmeier, E. D. (1975). The effect of school attendance on intellectual
development in northern Nigeria. Child Development,46, 281285.
Ferron, O. (1965). The test performance of colouredchildren. Educational
Research,8,4257.
Fick, M. L. (1929). Intelligence test results of poor white, native (Zulu),
coloured, and Indian school children and the educational and social
implications. South African Journal of Science,16, 904920.
Foy, P., Galia, J., & Li, I. (2008). Scaling the data from the TIMSS 2007
mathematics and science assessments. In J. F. Olson, M. O. Martin, & I. V. S.
Mullis (Eds.), TIMSS 2007 Technical Report (pp. 225279). Chestnut Hill
(MA): TIMSS & PIRLS International Study Center Accessed 9 June 2009 at:
http://timss.bc.edu/TIMSS2007/techreport.html
28 R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
Garden, R. A. (1987). Second IEA Mathematics Study. Sampling report. Report No.
CS-87-333 Washington DC: Center for Education Statistics Accessed 9 June
2009 at: http://eric.ed.gov/ERICDocs/data/ericdocs2sql/content_storage_01/
0000019b/80/1b/f1/4f.pdf
Giordani, B., Boivin, M. J., Opel, B., Dia Nseyila, D. N., & Lauer, R. E. (1996). Use
of the K-ABC with children in Zaire, Africa: An evaluation of the
sequential-simultaneous processing distinction within an intercultural
context. International Journal of Disability, Development and Education,
43,524.
Gonzalez, E. J., Galia, J., Arora, A., Erberber, E., & Diaconu, D. (2004). Reporting
student achievement in mathematics and science. In S. J. Chrostowski
(Ed.), TIMSS 2003 Technical Report (pp. 274307). Chestnut Hill (MA):
TIMSS & PIRLS International Study Center Accessed 9 June 2009 at:
http://timss.bc.edu/timss2003i/technicalD.html
Hanushek, E. A., & Woessman, L. (2007). The role of school improvements in
economic development. Munich: Institut für Wirtschaftsforschung Cesco
Working Paper No 1911.
Haward, L. C., & Roland, W. A. (1954). Some inter-cultural differences on the
Draw-a-Man test: Goodenough scores. Man,54,8688.
Heady, C. (2003). The effect of child labor on learning achievement. World
Development,31, 385398.
Heyneman, S. P., & Jamison, D. T. (1980). Student learning in Uganda. Com-
parative Education Review,24, 207220.
Holding,P.A.,Taylor,H.G.,Kazungu,S.D.,Mkala,T.,Gona,J.,Mwamuye,B.,
et al. (2004). Assessing cognitive outcomes in a rural African
population: Development of a neuropsychological battery in Kili
District, Kenya. Journal of the International Neuropsychological Society,
10, 246260.
Hunkin, V. (1950). Validation of the Goodenough Draw-a-Man Test for
African children. Journal for Social Research,1,5263.
Jensen, A. R. (1974). Cumulative decit: A testable hypothesis. Developmental
Psychology,10, 9961019.
Keeves, J. P. (1992). Learning science in a changing world. Cross-national
studies of science achievement: 1970 to 1984. International Association
for the Evaluation of Educational Achievement Report No. ISBN-92-9121-
001-3 Accessed 14 June 2009 at: http ://eric.ed.gov/ERICDocs/data/
ericdocs2sql/content_storage_01/0000019b/80/15/2b/8d.pdf
Kennedy,A. M., & Trong, K. L. (2007).Reporting studentachievementin reading.
In M. O. Martin, I. V. S. Mullis, & A. M. Kennedy (Eds.),PIRLS 2006 Technical
Report (pp. 173193). Chestnut Hill (MA): TIMSS & PIRLS International
Study Center Accessed 13 June 2009 at: http://timss.bc.edu/pirls2006/
tech_rpt.html
Khaleefa, O., Abdelwahid, S. B., Abdulradi, F., & Lynn, R. (2008). The increase
of intelligence in Sudan 19642006. Personality and Individual Differ-
ences,44, 412413.
Klein, O., Pohl, S., & Ndagijimana, C. (2007). The Inuence of intergroup
comparisons on Africans' intelligence test performance in a job selection
context. The Journal of Psychology,141, 453467.
Knoetze, J., Bass, N., & Steele, G. (2005). The Raven's Colored Progressive
Matrices: Pilot norms for Xhosa speaking primary school learners in part-
urban Eastern Cape. South African Journal of Psychology,35,175194.
Kongs, S. K., Thompson, L. L., Iverson, L., & Heaton, R. K. (2000). WCST-64
Manual. Odessa, FL: Psychological Assessment Resources.
Lapointe, A. E. (1992). Learning mathematics.Educational Testing Service
Report No. ETS-22-CAEP-01; ISBN-0-88685-120-3. Princeton (NJ):
ETS Accessed 16 June 2009 at: http://eric.ed.gov/ERICDocs/data/
ericdocs2sql/content_storage_01/0000019b/80/24/21/ac.pdf
Lloyd, F., & Pidgeon, D. A. (1961). An investigation into the effects of coaching
on nonverbal test material with European, Indian and African children.
British Journal of Psychology,31, 145151.
Loehlin, J. C., Lindzey, G., & Spuhler, J. N. (1975). Race differences in
intelligence. San Francisco, CA: Freeman.
Lynn, R. (2006). Race differences in intelligence: An evolutionary analysis.
Augusta, GA: Washington Summit Publishers.
Lynn, R., & Owen, K. (1994). Spearman's hypothesis and test score differences
between Whites, Indians, and Blacks in South Africa. Journal of General
Psychology,121,2736.
Lynn, R., & Vanhanen, T. (2006). IQ and global inequality. Augusta (GA):
Washington Summit.
Lynn, R., Meisenberg, G., Mikk, J., & Williams, A. (2007). National IQs predict
differences in scholastic achievement in 67 countries. Journal of Biosocial
Science,39, 861874.
MacArthur, R. S., Irvine, S. H., & Brimble, A. R. (1964). The Northern Rhodesia
Mental Ability Survey. Lusaka: Rhodes Livingstone Institute.
Martin, M. O., Mullis, I. V. S., Gonzales, E. J., & Chrostowski, S. J. (2004). TIMSS
2003 International Science Report Chestnut Hill (MA): IEA Accessed 23
Jan. 2007 at: http://timss.bc.edu/timss2003i/intl_reports.html
Martin, M. O., Mullis, I. V. S., & Foy, P. (2008). TIMSS 2007 Science Report
Chestnut Hill (MA): TIMSS & PIRLS International Study Center Accessed 9
June 2009 at: http://timss.bc.edu/TIMSS2007/intl_reports.html
Medrich, E. A., & Grifth, J. E. (1992). International mathematics and science
assessment: What have we learned? National Center for Education Statistics
Research and Development Report. Washington DC: U.S. Department of
Education.
Minde, K., & Kantor, S. (1976). Instructing Ugandan primary schoolchildren
in the execution of an intelligencetest: A controlled evaluation. Journal
of Cross Cultural Psychology,7, 209222.
Morakinyo, O. (1985). The brain-fag syndrome in Nigeria: Cognitive decits in
an illness associated with study. British Journal of Psychiatry,146,209210.
Mullis, I. V. S., Martin, M. O., Gonzales, E. J., & Chrostowski, S. J. (2004). TIMSS
2003 International Mathematics Report Chestnut Hill (MA): IEA Accessed
23 Jan. 2007 at: http://timss.bc.edu/timss2003i/intl_reports.html
Mullis,I.V.S.,Martin,M.O.,Kennedy,A.M.,&Foy,P.(2007).PIRLS2006
International Report Chestnut Hill (MA): TIMSS & PIRLS International
Study Center Accessed 6 June 2009 at: http://pirls.bc.edu/PDF/
PIRLS2006_international_report.pdf
Mullis, I. V. S., Martin, M. O., & Foy, P. (2008). TIMSS 2007 International
Mathematics Report Chestnut Hill (MA): TIMSS & PIRLS International
Study Center Accessed 9 June 2009 at: http://timss.bc.edu/TIMSS2007/
intl_reports.html
Murdoch, B. D., Fleming, J. H., Skuy, M., Painter, A., Schmidt, L., & Schutte, E. M.
(1994). Neuropsychological performance of a group of selected asympto-
maticblack South African children. Pretoria, South Africa: Human Sciences
Research Council.
Nell, V. (2000). Cross-cultural neuropsychological assessment: Theory and
practice. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Nissen, H. W., Machover, S., & Kinder, E. F. (1935). A study of performance
tests given to a group of native African negro children. British Journal of
Psychology,25, 308355.
Ohuche, N. M., & Ohuche, R. O. (1973). The Draw-A-Man Test as a predictor of
academic achievement. West African Journal of Educational and Vocational
Measurement,1,2027.
Owen, K. (1992). The suitability of Raven's Progressive Matrices for various
groups in South Africa. Personality and Individual Differences,13,149159.
Raveau, F. H. M., Elster, E., & Lecoutre, J. P. (1976). Migration et acculturation
différentiale. International Review of Applied Psychology,25, 145163.
Richter, L. M., Griesel, R. D., & Wortley, M. E. (1989). The Draw-a-Man test: A
50-year perspective on drawings done by black South African children.
South African Journal of Psychology,19,15.
Rindermann, H., Sailer,M., & Thompson,J. (2009). The impactof smart fractions,
cognitiveability of politiciansand average competence of peopleson social
development. Talent Development & Excellence,1,325.
Rushton,J. P., & Jensen, A. R. (2005). Thirtyyears of research on race differences
in cognitive ability. Psychology, Public Policy, and Law,11,235
294.
Shuttleworth Edwards,A. B., Kemp, R. D., Rust, A.L., Muirhead, J. G. L., Hartman,
N. P., & Radloff, S. E. (2004).Cross-culturaleffects on IQ test performance: A
review and preliminary normative indications on WAIS-III test perfor-
mance. Journal of Clinical and Experimental Neuropsychology,26,903920.
Skuy, M., Schutte, E., Fridjhon, P., & O'Carroll, S. (2001). Suitability of published
neuropsychologicaltest norms for urbanAfrican secondaryschool students
in South Africa. Personality and Individual Differences,30,14131425.
Sternberg,R. J., Nokes,C., Geissler, P. W.,Prince, R., Okatcha,F., Bundy, D. A.,et al.
(2001). The relationship between academic and practical intelligence: A
case study in Kenya. Intelligence,29,401418.
Sternberg, R. J., Grigorenko, E. L., Ngorosho, D., Tantufuye, E., Mbise, A., Nokes,
C., et al. (2002). Assessing intellectual potential in rural Tanzanian school
children. Intelligence,30, 141162.
Vass, V. (1992). Standardisation of Raven's Standard Progressive Matrices for
Secondary School African pupils in the Grahamstown Region. Unpub-
lished Master's Thesis, Rhodes University, Grahamstown.
Wechsler, D. (1974). WISC-R Manual. New York, NY: The Psychological
Corporation.
Wicherts, J.M. (2009). Personal communication.
Wicherts, J. M., Dolan, C. V., & van der Maas, H. L. J. (in press). A systematic
literature review of the average IQ of sub-Saharan Africans. Intelligence.
doi:10.1016/j.intell.2009.05.002.
Wicherts, J. M., Dolan, C. V., Carlson, J. S., & van der Maas, H. L. J. (in press).
Raven's test performance of sub-Saharan Africans: Average performance,
psychometric properties, and the Flynn Effect.
Wilson, D., Mundy-Castle, A., & Sibanda, P. (1991). Cognitive effects of LOGO
and computer-aided instruction among Black and White Zimbabwean
primary school girls. Journal of Social Psychology,131, 107116.
Zindi, F. (1994). Differences in psychometric performance. The Psychologist,
7, 549552.
29R. Lynn, G. Meisenberg / Intelligence 38 (2010) 2129
... Such scholars have repeatedly concluded that the average IQ of Sub-Saharan Africans is range that would be considered mildly mentally retarded (Lynn & Meisenberg, 2010;Rushton & Jensen, 2010). While the authors of such research claimed to be scientifically objective, there was evidence reported by others that the IQ researchers may have ignored many variables that would be confounding to their conclusions (Flynn, 2014;Gould, 1981;Steele & Aronson, 1995). ...
... IQs that indicate cognitive ability equivalent to being mildly mentally retarded (Lynn & Meisenberg, 2010;Rindermann, 2013;Rushton, & Jensen, 2010;Wicherts et al., 2010). ...
Full-text available
Thesis
The specific problem addressed is that implicit bias among scholars involved in studying or researching Sub-Saharan Africa can result in mindsets that are harmful and lead to a perpetuation of misinformation and mistreatment. The purpose of this quantitative study, using a descriptive design, was to investigate the levels of intercultural competency development among scholars from around the globe involved in studying Sub-Saharan Africa. The level of intercultural competence development was evaluated using the Intercultural Development Inventory (IDI, v. 3; Hammer, 2011b). Data were divided according to ethnicity and geographic base (Sub-Saharan African or not Sub-Saharan African). A series of t-tests were completed to make comparisons of scores from the IDI. Statistically significant differences were found on comparisons of ethnicity on all three scores. No statistically significant differences were found with comparisons of geographic base. The differences based on ethnicity were contrary to what was expected by the researcher. Scholars who identified as not Sub-Saharan African scored as more interculturally competent then those identifying as Sub-Saharan African. The possible impact of world view in how the research instrument was constructed is discussed as well as its potential impact on the outcomes.
... Unfortunately, as a caveat, the concern of whether the IQ compilations from Lynn et al. represent 'best practice' for estimating 'intelligence' is still open to debate in scholarly circles. Though some researchers continue to interpret underlying outcomes as reflecting biological disparities in intelligence (Kodila-Tedika & Asongu, 2015a, b), others are of the view that IQ scores may be affected by continental specificity (Lynn & Meisenberg, 2010). ...
... Given that Africa is in a comparatively volatile position, compared with other regions of the world, it is relevant to highlight that there are some concerns about Lynn et al.'s IQ measurement for the continent. According to Lynn and Meisenberg (2010), for a more representative African sample the average IQ is low (61), but high (80) for an elitist sample. Rindermann (2013) has suggested that controlling some continent-specific factors such as higher schooling age and lower school enrolment could lead to an average IQ of about 75. ...
Full-text available
Article
The purpose of this study is to assess the nexus between intelligence (or human capital) and statistical capacity in developing countries. The line of inquiry is motivated essentially by: (i) the scarce literature devoted to elucidating poor statistics in developing countries and (ii) an evolving stream of literature on knowledge economy. We have established a positive association between intelligence quotient (IQ) and statistical capacity. The relationship is: (i) consistent with the employment of alternative specifications based on varying conditioning information sets and (ii) robust to the control of outliers. Policy implications are discussed.
... The critics suggest that scores from Sub-Saharan countries are implausibly low, do not use representative samples, and may be unduly deflated by cultural factors, poor nutrition, and poor education. Lynn and Meisenberg (2010a) have responded to these critiques defending the quality of the national IQ scores. Lynn and Meisenberg argued many of the studies Wicherts et al. used to show higher IQ for Sub-Saharan Africans are unrepresentative because they are based on university students. ...
Full-text available
Article
Since Lynn and Vanhanen's book IQ and the Wealth of Nations (2002), many publications have evidenced a relationship between national IQ and national prosperity. The strongest statistical case for this lies in Jones and Schneider's (2006) use of Bayesian model averaging to run thousands of regressions on GDP growth (1960-1996), using different combinations of explanatory variables. This generated a weighted average over many regressions to create estimates robust to the problem of model uncertainty. We replicate and extend Jones and Schneider's work with many new robustness tests, including new variables, different time periods, different priors and different estimates of average national intelligence. We find national IQ to be the "best predictor" of economic growth, with a higher average coefficient and average posterior inclusion probability than all other tested variables (over 67) in every test run. Our best estimates find a one point increase in IQ is associated with a 7.8% increase in GDP per capita, above Jones and Schneider's estimate of 6.1%. We tested the causality of national IQs using three different instrumental variables: cranial capacity, ancestry-adjusted UV radiation, and 19 th-century numeracy scores. We found little evidence for reverse causation, with only ancestry-adjusted UV radiation passing the Wu-Hausman test (p < .05) when the logarithm of GDP per capita in 1960 was used as the only control variable.
... Olazaran et al., 1996;Shuttleworth-Edwards et al., 2004). In Africa, research from Libya (Al-Shahomee and Lynn, 2010;Bakhiet and Lynn, 2015), Egypt (Abdel-Khalek, 1988), Ghana (Anum, 2014), Kenya (Costenbader and Ngari, 2001), and South Africa (Knoetze et al., 2005;Lynn and Meisenberg, 2010a) have established that test scores and IQ equivalents are significantly lower than expected, sometimes by more than 1 standard deviation. These findings have been the basis for the debate of whether lower scores on intelligence tests among sub-Saharan African children are due to genetics or to the environment within with children grow (Rindermann, 2013). ...
Full-text available
Article
Current literature shows an association between intelligence and socio-cultural or socio-economic factors. The available evidence supports a stronger effect of exogenous factors on measures of crystalized intelligence than on fluid intelligence. Despite this, the sources of variability in fluid and crystalized intelligence have not been explored adequately in intelligence research. The purpose of this study was to compare performance on tests that measure fluid and crystallized intelligence among children selected from public and private schools in Ghana. We tested the assumption that socio-economic status (SES) will have a stronger effect on tests that measure crystallized intelligence than on fluid intelligence. We selected 185 children between 6 and 12 years from private and public schools, and used inclusion in a private or public school as a proxy for SES. We administered the Raven’s Coloured Progressive Matrices (RCPM), a fluid intelligence test, the KABC II story completion subtest as a measure of inductive reasoning and crystallized intelligence and the Kilifi Naming Test, a verbal ability measure designed to minimize the effect of school on vocabulary. The results showed age-related improvement in scores on all three tests with effect sizes ranging from 0.42 to 0.52. We also found significant effect for type of school on all the tests with effect sizes ranging from 0.37 to 0.66. The results also showed an increasing disparity in performance on the tests favoring children selected from private schools. These suggest that fluid and crystalized intelligence are affected by socioeconomic factors. The results also showed that SES factors tend to affect crystallized ability more than it affects fluid ability. The results are discussed in the context of differences in socioeconomic resources available to children such as quality of education in low- and middle-income countries.
... According to Wicherts and colleagues, their systematic review adequately circumvented these difficulties, allowing them to conclude an average IQ of 82 was not the product of either publication or sampling biases. Lynn and Meisenberg (2010) questioned the corrections suggested by Wicherts and colleagues (2010), contending that the authors' systematic review relied on unrepresentative elite samples, overestimating the average IQ scores for African countries. Wicherts, Dolan, Carlson, and van der Maas (2010b, a) recomputed the African IQ scores after removing these elite samples. ...
Full-text available
Article
After many waves of out-migration from Africa, different human populations evolved within a great diversity of physical and community ecologies. These ambient ecologies should have at least partially determined the selective pressures that shaped the evolution and geographical distribution of human cognitive abilities across different parts of the world. Three different ecological hypotheses have been advanced to explain human global variation in intelligence: (1) cold winters theory (Lynn, 1991), (2) parasite stress theory (Eppig, Fincher, & Thornhill, 2010), and (3) life history theory (Rushton, 1999, 2000). To examine and summarize the relations among these and other ecological parameters, we divided a sample of 98 national polities for which we had sufficient information into zoogeographical regions (Wallace, 1876; Holt et al., 2013). We selected only those regions for this analysis that were still inhabited mostly by the aboriginal populations that were present there prior to the fifteenth century AD. We found that these zoogeographical regions explained 71.4% of the variance among national polities in our best measure of human cognitive ability, and also more concisely encapsulated the preponderance of the more specific information contained within the sampled set of continuous ecological parameters.
... PISA scores and scores on IQ tests are so highly correlated at the country level that we can only conclude that they measure the same construct. Between-country differences are larger on PISA than on IQ tests, though, suggesting that IQ tests are less "culturally biased" than PISA tests (Lynn & Meisenberg, 2010). The PISA results are used to calculate intelligence differences between countries, together with results from other school assessment programs and from IQ tests (Lynn & Becker, 2019). ...
Full-text available
Article
Saini’s book assails “race science”, which is presented as the root cause of racism in modern societies. The content of this science is either not described at all or is grossly misrepresented. The book is nevertheless valuable because it reveals the cognitive structure of one type of anti-science ideology that is influential in modern societies. This ideology turns Enlightenment philosophy on its head, claiming that science and reason create rather than challenge prejudice.
... To može biti vezano i uz razdoblje potrebno za automatizaciju temeljnih matematičkih vještina i očekivanja u pogledu kurikuluma, odnosno pokazatelje neautomatiziranosti procesa. Razlog kasnijem donošenju dijagnoza može biti i u manjku nacionalnih testiranja iz matematike i "referentnih podataka", kao i nedostatnom broju logopeda koji provode procjenu teškoća u području matematike, ali i nedostatnom općem interesu za uspjeh učenika u matematici -zbog uvjerenja da su "samo pojedinci nadareni za matematiku (Lynn i Meisenberg, 2010) i da je matematika općenito teška za usvajanje svima". ...
Article
U istraživanju poznavanja osnovnih aritmetičkih operacija sudjelovalo je 145 osnovnoškolaca bez teškoća u matematici, polaznika trećeg do osmog razreda, s ciljem dobivanja uvida u razinu ukupne ovladanosti temeljnom matematikom u svakom od razreda. U radu se razmatra uspješnost rješavanja pojedinih aritmetičkih zadataka u domeni zbrajanja, oduzimanja, množenja, dijeljenja te poznavanja pravila izvođenja računskih radnji prema porastu obrazovne dobi. Rezultati su pokazali da se skupina učenika trećih razreda statistički značajno razlikuje od učenika na ostalim obrazovnim razinama. Stoga je moguće pretpostaviti da se upravo u trećem razredu mijenjaju strategije rješavanja zadataka kod temeljnih matematičkih operacija, što je u skladu s drugim sličnim istraživanjima. Čini se da strategije prizivanja podataka zamjenjuju strategije temeljene na brojenju. Nema statistički značajnih razlika u uspješnosti rješavanja zadataka osnovnih aritmetičkih operacija učenika od četvrtog do osmog razreda. Stoga je moguće pretpostaviti da su razine temeljnih matematičkih kompetencija ujednačene s aspekta točnosti za ove obrazovne razine. Kako je ispitivanje provedeno na relativno malom uzorku u odnosu na razrede, generalizacija rezultata je ograničena, a oni mogu biti korisni za oblikovanje smjernica u izradi ispitnog materijala za prepoznavanje teškoća u matematici i određivanje pravca intervencija kod navedenih teškoća ili diskalkulije.
... This approach yields an average IQ of 80 for this sample, which is significantly higher than that found in the LV data. Lynn and Meisenberg (2010b) and Lynn and Vanhanen (2012) provide counter arguments, and Jones and Potrafke (2014) note that the sampling bias inherent in Wicherts, et al. approach could just as easily yield an over-estimate of the average human capital level. 3 Rindermann's (2013) recent analysis of African cognitive measurement is worth noting in this regard. ...
Full-text available
Article
It has been shown that country-level IQ and aggregated performance by school-age children on international assessment tests in math and science are by-in-large capturing analogous indicators of the cognitive human capital. We expand that analysis by comparing country-level IQ to the World Economic Forum’s Human Capital Index (HCI). This index, comprised of several dozen separate indicators, accounts for inputs and outcomes to measure human capital, across age profiles and gender. Two outcomes are of note. First, there is a positive, significant correlation between IQ and the vast majority of the component indicators in the HCI across all age cohorts. Second, because the HCI’s interpretation of educational attainment extends beyond formal education by including indicators such as on-the-job learning and other work-related skills, our finding that IQ is positively correlated with these measures suggests a deeper connection between national average IQ and the fundamental factors of what constitutes the cognitive side of human capital development.
Full-text available
Article
Many intellectuals enthusiastically denounce those who argue that genes play some role in cognitive differences between human populations. However, such proposals are perfectly reasonable and are, in fact, consistent with the Darwinian research tradition in which most modern social scientists profess to operate. We argue that population-based cognitive differences are congruent with our best understanding of the world because there are strong reasons to believe that different environments and niches selected for different physical and psychological traits, including general cognitive ability. Like most hereditarians (those who believe it likely that genes contribute to differences in psychological traits among human populations), we do not believe there is decisive evidence about the causes of differences in cognitive ability. But we will argue that a partial genetic hypothesis is most consistent with the Darwinian research tradition.
Article
The Relational Abilities Index (RAI) has shown considerable utility as a functional proxy measurement of intellectual performance by providing a metric of an important skill set known as relational skills, which are proposed to underlie much of what we conceive of as intellectual behavior. The Relational Abilities Index+ (RAI+) assesses performance across an extended range of relational skills (Same/Opposite, More/Less, Same/Different, Before/After, and Analogy), and has been designed to provide a more comprehensive and nuanced assessment of relational skills. The current study aims to investigate the validity and utility of the RAI+ by assessing its degree of correlation with well-established assessments of intelligence (WASI), numeracy (WAIS: Arithmetic), and educational attainment (WIAT-T-II). Results indicate that the RAI+ displays considerable efficacy in predicting intellectual performance and numeracy, but not educational attainment.