Most Accurate National IQs Possible

v3

Version 3 (.1)

Methodology: a total of 5,645 means were gathered during the data collection phase, most drawing from Becker’s dataset, PISA, PIRLS, TIMSS, the working memory meta-analysis, and the harmonised learning outcomes.

These means were classified into three categories: international scholastic tests, IQ samples (from Becker’s dataset + some other sources), and the samples from the working memory meta.

The IQ samples were adjusted for the effect of perceived selectivity (e.g. college students or white collar workers would be highly selected samples, janitors would be negatively selected) and location (urban, rural, national, etc). All samples were age-normed prior to the conduction of statistical analysis.

The memory samples were adjusted for the age of testing, year, perceived selectivity, and the type of test (e.g. forward span vs backward span). The standard deviations for these variables also had to be adjusted for these factors. IQs were then calculated using these means and standard deviations. The quality of these averages was nothing short of awful, despite my best efforts; according to the model, some of the samples had average IQs as high as 150 and as low as 40.

Then, all of the samples were combined into one file and subject to an anchoring process which controls all of the samples for bias at the category level. For example, if the PISA math test from 2008 gives scores 20 points higher in comparison to what would be expected from the set of tested countries, the scores were adjusted downwards by 20 points. The reference group (IQ = 100) is white Britons.

I then grouped the tests into six different categories: PIRLS, TIMSS, memory, PISA, IQ, and all other tests (other). I gave weights of 3 to the PIRLS, TIMSS, PISA, and ‘other’ averages, a weight of 2 to the IQ averages, and a weight of 1 to the memory averages. No manual revisions were made and I refused to impute the IQs of Turkmenistan, Guyana, Suriname, and French Guinea with proxies like performance in the IMO.

I also calculated the standard errors of these estimates by taking the standard deviation of the sample means and dividing it by the square root of the number of them¹. Countries with only one sample had their standard errors estimated based on the observed relationship between number of samples and standard error:

On average, each country had a standard error of 1.53. This is a 41% improvement from the prior dataset (V2) which had an estimated average standard error of 2.58.

Countries that were missing in this version of the dataset were either left blank or had their value imputed from the previous version².

National IQ FAQ

• Do IQ tests have a regional bias?

I reviewed the literature in this preprint and couldn’t come to a definitive conclusion. There is one study (Wicherts’, I believe) that found that GPA and IQ were less correlated in Africa than they were in Europe, but that could also be an artefact of worse grading. Most of the tools traditionally used to assess bias (differential item testing, comparing g-loadings across groups) find little evidence of bias, but I personally don’t think these methods work well.

When it comes to bias testing, the golden standard is taking variables that are assumed to be racially unbiased (e.g. education, income), regressing those onto IQ between groups, and observing if the slopes/intercepts differ to a practical/statistical extent. To my knowledge, there has been no attempt to do this across countries.

Speaking clearly, I think the largest issue with comparing cognitive tests across countries are education and effort. Education causes people’s IQ scores to increase because they become more familiar with standardised testing and improve their skills in maths/reading, but don’t actually get more intelligent. As such, more educated countries should have higher IQ scores that are not reflective of superior general intelligence. Some people have theories that IQ tests are biased by effort across countries, but I don’t see much evidence for this; empirical attempts to assess the question have found the opposite (check the appendix of said paper).

Alternatively, one could argue that there are factors that shrink the observed averages, like using between group standard deviations and imperfectly g-loaded tests. Overall, I would guess that the differences between countries are slightly inflated.

• Is the average IQ in Sub-Saharan Africa really 70?

More or less. The score is reflective of their ability to take cognitive tests. An IQ of 70 is commonly used as a cutoff for intellectual disability in the United States, but I should note that this is not a hard cutoff and there is general agreement that IQ alone cannot be used to diagnose intellectual disability. I suspect that, controlling for education and measurement invariance, the true IQ of Sub-Saharan Africa is closer to 75.

• Is the average IQ in China really 101?

The average is inflated by several points because Eastern provinces and urban areas are oversampled. A nationally representative sample of China would probably score somewhere between 95 and 100.

• How much does biased collection play a role in the differences across countries?

Almost none. Lynn made some mistakes in the collection of the data but he was making an honest effort from what can be inferred.

About half of the data come from international scholastic assessments that take (roughly) representative samples of students and compare them in terms of their ability. Despite the samples being massive and the data collection being conducted by an independent body, the scores on these tests correlate reasonably well with the ones found in IQ tests:

from here

• Are any of these values based on geographical imputations?

No.

• Do you have any concerns with the estimates of specific countries?

I think that China probably has an IQ in the upper 90s, Mynamar in the mid 80s, Kazakhstan in the upper 80s, and North Korea in the low 90s.

I think the North Korean IQ is deflated by sampling refugees who live in South Korea and I think the Chinese IQ is inflated because Eastern provinces higher in intelligence are disproportionately sampled. I don’t have any specific comments on the others but they seem to be at odds with their levels of development and cultural prominence.

V3 Appendix:

Average IQ by region:

 region                          mean_IQ
   <chr>                             <dbl>
 1 Eastern Asia                     99.823
 2 Western Europe                   99.794
 3 Northern Europe                  98.801
 4 Australia and New Zealand        98.721
 5 Northern America                 96.111
 6 Eastern Europe                   96.004
 7 Southern Europe                  91.588
 8 South-eastern Asia               88.311
 9 Polynesia                        86.841
10 Central Asia                     85.608
11 Western Asia                     83.863
12 Micronesia                       82.047
13 Latin America and the Caribbean  82.007
14 Southern Asia                    78.556
15 Melanesia                        78.513
16 Northern Africa                  78.503
17 Sub-Saharan Africa               69.308

Values:

    alpha3                     name     NIQ    se
1      HKG      Hong Kong SAR China 105.689 0.846
2      SGP                Singapore 105.317 0.923
3      JPN                    Japan 104.964 0.556
4      TWN                   Taiwan 104.870 1.073
5      KOR              South Korea 103.270 0.915
6      LIE            Liechtenstein 102.050 0.734
7      EST                  Estonia 101.553 0.315
8      FIN                  Finland 101.335 0.838
9      CHN                    China 101.286 1.503
10     MAC          Macao SAR China 101.206 0.610
11     CHE              Switzerland 101.055 0.669
12     AUT                  Austria 100.735 0.905
13     NLD              Netherlands 100.724 0.654
14     CAN                   Canada 100.106 0.817
15     IRL                  Ireland  99.923 0.637
16     HUN                  Hungary  99.791 0.422
17     SWE                   Sweden  99.778 0.348
18     AUS                Australia  99.619 0.703
19     LUX               Luxembourg  99.479 0.594
20     RUS                   Russia  99.432 0.431
21     GBR           United Kingdom  99.406 0.584
22     CZE                  Czechia  99.396 0.283
23     DEU                  Germany  99.152 0.724
24     DNK                  Denmark  99.025 0.414
25     VNM                  Vietnam  98.904 1.706
26     SVN                 Slovenia  98.841 0.513
27     USA            United States  98.770 0.465
28     POL                   Poland  98.398 0.611
29     BEL                  Belgium  98.380 0.641
30     BLR                  Belarus  98.035 1.918
31     SVK                 Slovakia  97.884 0.381
32     NZL              New Zealand  97.823 0.546
33     NOR                   Norway  97.804 0.472
34     LVA                   Latvia  97.423 0.407
35     SCO                 Scotland  97.308 0.384
36     LTU                Lithuania  97.123 0.358
37     FRA                   France  96.774 0.609
38     HRV                  Croatia  96.464 0.696
39     ISL                  Iceland  96.140 0.531
40     ITA                    Italy  95.478 0.471
41     PRT                 Portugal  95.223 0.691
42     MMR          Myanmar (Burma)  95.079 3.362
43     ISR                   Israel  94.981 0.776
44     ESP                    Spain  94.733 0.902
45     CYP                   Cyprus  93.698 0.559
46     BGR                 Bulgaria  93.345 0.864
47     GRC                   Greece  92.931 0.551
48     KAZ               Kazakhstan  92.868 1.177
49     BMU                  Bermuda  92.844 1.624
50     GRL                Greenland  92.725 2.491
51     SRB                   Serbia  92.354 1.147
52     MLT                    Malta  92.233 0.595
53     MYS                 Malaysia  91.854 0.788
54     MNG                 Mongolia  91.801 3.429
55     UKR                  Ukraine  91.572 0.613
56     MDA                  Moldova  91.524 0.760
57     BRB                 Barbados  91.365 2.785
58     ARM                  Armenia  91.097 1.132
59     TUR                   Turkey  90.913 0.544
60     ROU                  Romania  90.664 0.534
61     ALB                  Albania  90.510 1.015
62     BRN                   Brunei  90.060 0.920
63     WSM                    Samoa  90.000 2.493
64     CHL                    Chile  89.460 0.625
65     TCA   Turks & Caicos Islands  89.400 2.494
66     PLW                    Palau  89.287    NA
67     THA                 Thailand  89.181 0.949
68     URY                  Uruguay  89.088 0.464
69     COK             Cook Islands  89.000 2.494
70     BIH     Bosnia & Herzegovina  88.339 1.122
71     CRI               Costa Rica  88.190 1.048
72     MEX                   Mexico  88.046 0.839
73     MNE               Montenegro  88.000 0.351
74     PRI              Puerto Rico  87.930 1.849
75     SUR                 Suriname  87.843    NA
76     TJK               Tajikistan  87.710 2.495
77     TTO        Trinidad & Tobago  87.523 0.547
78     LKA                Sri Lanka  87.375 2.541
79     GEO                  Georgia  87.199 0.526
80     ARE     United Arab Emirates  86.866 0.526
81     BHR                  Bahrain  86.700 0.727
82     AZE               Azerbaijan  86.512 1.111
83     ARG                Argentina  86.172 0.897
84     VIR      U.S. Virgin Islands  86.100 3.695
85     VEN                Venezuela  85.802 3.771
86     PRK              North Korea  85.500 3.500
87     NCL            New Caledonia  85.000 2.496
88     MUS                Mauritius  84.978 1.716
89     BOL                  Bolivia  84.970 4.784
90     BHS                  Bahamas  84.733 2.557
91     MKD          North Macedonia  84.587 0.705
92     BRA                   Brazil  84.389 0.864
93     COL                 Colombia  84.365 1.157
94     PER                     Peru  84.265 0.694
95     FJI                     Fiji  84.000 2.497
96     MHL         Marshall Islands  83.960 2.497
97     IRN                     Iran  83.751 0.752
98     TUN                  Tunisia  83.486 1.322
99     UZB               Uzbekistan  83.321 2.307
100    JAM                  Jamaica  82.915 2.188
101    QAT                    Qatar  82.755 1.064
102    CUB                     Cuba  82.507 1.489
103    ECU                  Ecuador  82.209 1.688
104    IDN                Indonesia  82.162 0.816
105    OMN                     Oman  82.026 0.845
106    LBN                  Lebanon  81.966 0.940
107    TON                    Tonga  81.522 2.520
108    SAU             Saudi Arabia  81.421 0.787
109    LAO                     Laos  81.350 3.336
110    TKM             Turkmenistan  81.260    NA
111    DOM       Dominican Republic  81.230 1.020
112    SYC               Seychelles  81.198 1.061
113    KIR                 Kiribati  81.180 2.499
114    ANT     Netherlands Antilles  81.035 2.499
115    MNP Northern Mariana Islands  81.000 2.499
116    KSV                   Kosovo  80.954 0.573
117    JOR                   Jordan  80.884 0.815
118    KHM                 Cambodia  80.729 2.817
119    PSE  Palestinian Territories  80.630 0.912
120    GTM                Guatemala  80.366 0.663
121    PAN                   Panama  80.087 0.660
122    IRQ                     Iraq  79.664 3.252
123    PHL              Philippines  79.475 1.841
124    SLV              El Salvador  79.326 1.051
125    HND                 Honduras  79.290 1.433
126    KWT                   Kuwait  79.162 1.305
127    DZA                  Algeria  79.134 0.811
128    SDN                    Sudan  78.713 1.604
129    KGZ               Kyrgyzstan  78.535 2.334
130    IND                    India  78.533 1.428
131    KEN                    Kenya  78.456 1.253
132    SYR                    Syria  78.222 2.132
133    NIC                Nicaragua  77.966 2.501
134    LBY                    Libya  77.812 2.580
135    PRY                 Paraguay  77.764 1.595
136    GAB                    Gabon  77.537 1.970
137    SWZ                 Eswatini  77.407 0.635
138    NPL                    Nepal  77.332 0.303
139    EGY                    Egypt  77.331 1.627
140    TLS              Timor-Leste  77.304 0.702
141    MDV                 Maldives  77.260    NA
142    BWA                 Botswana  76.902 0.657
143    AFG              Afghanistan  76.400 2.502
144    MRT               Mauritania  76.400 2.502
145    BTN                   Bhutan  76.312    NA
146    BGD               Bangladesh  76.188 3.660
147    CYM           Cayman Islands  76.000 2.502
148    GUY                   Guyana  75.573    NA
149    KNA        St. Kitts & Nevis  75.516    NA
150    SLB          Solomon Islands  75.486 2.502
151    ATG        Antigua & Barbuda  75.470    NA
152    BDI                  Burundi  75.312 2.094
153    SOM                  Somalia  75.202 2.503
154    GRD                  Grenada  74.671    NA
155    MAR                  Morocco  74.541 0.781
156    PNG         Papua New Guinea  74.381 4.383
157    TZA                 Tanzania  74.150 2.578
158    VUT                  Vanuatu  73.695 2.504
159    NRU                    Nauru  73.569    NA
160    ZNZ                 Zanzibar  73.537 3.345
161    ETH                 Ethiopia  73.141 3.002
162    ZWE                 Zimbabwe  73.024 1.761
163    BFA             Burkina Faso  72.936 2.310
164    ERI                  Eritrea  72.619 5.112
165    MOZ               Mozambique  72.547 2.391
166    RWA                   Rwanda  72.367 1.998
167    GMB                   Gambia  72.063 1.734
168    CPV               Cape Verde  71.260    NA
169    ZAF             South Africa  71.156 1.485
170    VCT St. Vincent & Grenadines  70.880 2.505
171    PAK                 Pakistan  70.311 3.617
172    UGA                   Uganda  70.099 2.101
173    SEN                  Senegal  69.972 1.108
174    AGO                   Angola  69.817 5.257
175    GNQ        Equatorial Guinea  69.667 2.506
176    MWI                   Malawi  69.269 1.700
177    SSD              South Sudan  68.818 2.062
178    LBR                  Liberia  68.757 1.570
179    NGA                  Nigeria  68.535 1.617
180    BLZ                   Belize  67.914 2.507
181    BEN                    Benin  67.181 1.894
182    LCA                St. Lucia  67.109    NA
183    HTI                    Haiti  66.902 2.008
184    NAM                  Namibia  66.873 1.252
185    LSO                  Lesotho  66.423 0.662
186    CMR                 Cameroon  66.199 1.147
187    DMA                 Dominica  66.040 0.250
188    COG      Congo - Brazzaville  65.653 1.459
189    COD         Congo - Kinshasa  64.951 1.042
190    YEM                    Yemen  64.839 3.115
191    GHA                    Ghana  64.516 1.224
192    GNB            Guinea-Bissau  64.260    NA
193    CAF Central African Republic  64.000 2.510
194    MDG               Madagascar  63.983 2.912
195    ZMB                   Zambia  63.654 1.541
196    SLE             Sierra Leone  62.818 2.718
197    CIV            Côte d’Ivoire  62.755 0.815
198    STP      São Tomé & Príncipe  62.260    NA
199    TGO                     Togo  62.210 1.179
200    GIN                   Guinea  62.094 1.757
201    COM                  Comoros  60.355 2.512
202    MLI                     Mali  60.348 1.919
203    TCD                     Chad  60.046 1.076
204    DJI                 Djibouti  60.000 2.512
205    NER                    Niger  58.249 1.981

(The following statistics do not include the countries that are not present in the data of V3 but were in V2).

Correlation matrix:

Sample mean and estimated standard error:

Correlation between GDP per capita (IMF data, PPP controlled) and IQ:

World IQ: 85.8

Link to dataset (it’s the last file).

Version 2

White British mean/SD set to 500/100.

Methodology: composite of various datasets. Paper will be out in a few months.

Preprint is now out!

World IQ: 85.6 weighted by population

1. Technically this estimates the standard error of the mean of all samples (not weighted subcategories), though I thought that this standard error would be a better estimate of the error in the estimates than the one that would be calculated from the subcategories.

2. Change introduced in V 3.1