ArticlePDF Available

Public Preferences and Reality: Crime Rates among 70 Immigrant Groups in the Netherlands

Authors:
  • Ulster Institute for Social Research

Abstract and Figures

We estimated crime rates among 70 origin-based immigrant groups in the Netherlands for the years 2005-2018. Results indicated that crime rates have overall been falling for each group in the period of study, and in the country as a whole, with about a 50% decline since 2005. Immigrant groups varied widely in crime rates, with East Asian countries being lower and Muslim countries, as well as Dutch (ex-)colonial possessions in the Caribbean, being higher than the natives. We found that national IQ and Muslim percentage of population of the origin countries predicted relative differences in crime rates, r’s = .64 and .45, respectively, in line with previous research both in the Netherlands and in other European countries. Furthermore, we carried out a survey of 200 persons living in the Netherlands to measure their preferences for immigrants for each origin country in terms of getting more or fewer persons from each country. Following Carl (2016), we computed a mean opposition metric for each country. This correlated strongly with actual crime rates we found, r’s = .46 and .57, for population weighted and unweighted results, respectively. The main outliers in the regression were the Dutch (ex-)colonial possessions, and if these are excluded, the correlations increase to .68 and .66, respectively. Regressions with plausible confounders (Muslim percentage, geographical fixed effects) showed that crime rates continued to be a useful predictor of opposition to specific countries. The results were interpreted as being in line with a rational voter preference for less crime-prone immigrants.
Content may be subject to copyright.
MANKIND QUARTERLY 2020 60:3 327-351
327
Public Preferences and Reality: Crime Rates among 70
Immigrant Groups in the Netherlands
Emil O. W. Kirkegaard*
Ulster Institute for Social Research, London, UK
Max de Kuijper
Independent researcher, Netherlands
*Corresponding author: emil@emilkirkegaard.dk
We estimated crime rates among 70 origin-based immigrant
groups in the Netherlands for the years 2005-2018. Results
indicated that crime rates have overall been falling for each group in
the period of study, and in the country as a whole, with about a 50%
decline since 2005. Immigrant groups varied widely in crime rates,
with East Asian countries being lower and Muslim countries, as well
as Dutch (ex-)colonial possessions in the Caribbean, being higher
than the natives. We found that national IQ and Muslim percentage
of population of the origin countries predicted relative differences in
crime rates, r’s = .64 and .45, respectively, in line with previous
research both in the Netherlands and in other European countries.
Furthermore, we carried out a survey of 200 persons living in the
Netherlands to measure their preferences for immigrants for each
origin country in terms of getting more or fewer persons from each
country. Following Carl (2016), we computed a mean opposition
metric for each country. This correlated strongly with actual crime
rates we found, r’s = .46 and .57, for population weighted and
unweighted results, respectively. The main outliers in the regression
were the Dutch (ex-)colonial possessions, and if these are excluded,
the correlations increase to .68 and .66, respectively. Regressions
with plausible confounders (Muslim percentage, geographical fixed
effects) showed that crime rates continued to be a useful predictor
of opposition to specific countries. The results were interpreted as
MANKIND QUARTERLY 2020 60:3
328
being in line with a rational voter preference for less crime-prone
immigrants.
Key Words: Immigration, Crime, Netherlands, Stereotype accuracy
Western European countries are seeing levels of immigration that are
unprecedented in the last hundreds of years. Generally speaking, immigrants as
a whole perform below the level of the natives in terms of income, education,
crime rates, use of benefits etc. (Andersson & Jespersen, 2018; MG, 2015, 2016;
D. Murray, 2017; Roth, 2010; Sanandaji, 2017; Sarrazin, 2012). Causes of such
social performance gaps are heavily debated both in the academic literature and
society in general (Hesson, 2019; Pickering & Ham, 2015; Salmi et al., 2015).
Unfortunately, determining the causes of performance gaps is difficult for a
number of reasons. First, the case-level data for studying immigrant crime are not
generally available for researchers since these depend on government data
protected by privacy regulations, or only available by application limited to
university employed researchers which requires an arduous process. Thus, most
data published for public use are aggregated statistics. Second, most such
published government aggregated statistics and studies do not distinguish
between immigrant groups by country of origin, which would allow for using origin
country characteristics as predictors (Borjas, 2016; Hamilton & Hummer, 2011).
For instance, many reports group together all immigrants into categories such as
EU-origin or Western immigrants, whose definitions can both change over time
(as EU membership changes) and between reports. However, in recent years, a
number of datasets that disaggregate by country of origin have become available
for many countries.
A number of studies have been done using these national origin datasets.
Most of them use national IQ as a predictor. The idea with this is that immigrant
populations are at least roughly representative of their origin countries in terms of
intelligence, and thus one can approximate the average intelligence levels of
immigrant groups by using (the estimate of) their home country average. This
method is more dubious when the data include second and later generation
immigrants, as these are generally found to have better cognitive scores than first
generation immigrants (Rindermann & Thompson, 2016; Robie et al., 2017; but
see Kirkegaard, 2013), suggesting environmental causation of between country
gaps and biasing the estimates of between group gaps in the host country. In the
same way, some studies have used the Muslim % of the home country as a best
guess for the prevalence of this religious affiliation among the immigrants and
their descendants. With regards to predictive analysis using country of origin
variables, Kirkegaard (2017) presented a preliminary meta-analysis of results
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
329
from studies covering six host countries (Denmark, Norway, Sweden, Finland,
Germany, and the Netherlands), all of which are located in Northern Europe. He
found that national IQ predicted individual outcomes (e.g. crime rate) or a general
social outcome or status composite (SES/S factor) with correlations of .40 to .62,
with a mean of .51 (n = 16). For Muslim %, the values ranged from .18 to .76, with
a mean of .55. The summary of results is shown in Figure 1. Other researchers
analyzing data for the US have also found national IQs to be useful in analysis of
immigrant data (Jones & Schneider, 2010; Vinogradov & Kolvereid, 2010;
Whitaker, 2018).
Figure 1. Summary of meta-analysis of country of origin predictive studies. From
Kirkegaard (2017).
Decades of research in sociology, criminology and differential psychology
show that higher intelligence causes better outcomes in general, whether these
relate to education, income, health, unemployment or criminality. Evidence for
this claim comes from many research designs including prospective studies,
which rule out reverse causation, sibling studies, which rule out any confounder
MANKIND QUARTERLY 2020 60:3
330
that differs between families (Frisell et al., 2012; Hegelund et al., 2018, 2019;
Herrnstein & Murray, 1994; C. Murray, 2002), and GWASs that allow for
functional analysis of genetic causes of intelligence and outcomes (Hill et al.,
2019). In line with expectations, it is well established that immigrant populations
in Western countries in general have below native levels of average intelligence
(Kirkegaard, 2019b; Rindermann & Thompson, 2016; Robie et al., 2017) and that
these are related to their origin countries’ level of ability. Putting these facts
together, it was predicted that national IQs would predict variation in crime rates
among immigrant groups.
The use of Muslim% as a predictor in the literature is not founded upon an
equally strong theoretical or empirical basis, but is included because many
commentators have noted that Muslim groups in particular tend to perform poorly
(but see Kuran, 2018; Rindermann, 2018, sec. 4.4.3; Sarrazin, 2018). In fact,
origin country Muslim% often predicts social outcomes better than does national
IQ, and sometimes is a stronger predictor in multiple regression, though because
of the collinearity and limited sample sizes, these regressions give very imprecise
estimates (Kirkegaard & Fuerst, 2014).
The reason for this predictive validity is unclear. First, it could be because
Islam teaches Muslims to act in a hostile manner towards non-Muslims, which
would result in a prediction of high outgroup crime rates but not e.g. high
unemployment rates (unless these are interpreted as economic aggression) or
crimes against other Muslims (unless from another variant of Islam, Shia vs.
Sunni vs. Alevism). Second, it could be because Muslim faith is correlated with
other traits that are causal for general social performance, such as interpersonal
trust, work ethic, mental well-being, or clannishness (Carl, 2017; Schulz et al.,
2019). The question is difficult to examine without the use of individual-level
datasets that also contain measures of potential confounders, especially
intelligence, and mediators, such as educational attainment. The authors have
been looking for such datasets for a number of years without luck. At present, the
authors do not advance any particular model for why the relationship exists, but
we include the predictor because of its potential causality and evident predictive
validity, and hope that future studies might clarify its role in the nomological
network.
The reasons for the present study were two-fold. First, we are only aware of
one published study that examined immigrant groups in the Netherlands grouped
by country of origin (Kirkegaard, 2015). This study however relied on old and
limited data from a Dutch language report that examined data from 2002 (Blom
et al., 2005), whose reliability was questionable. Thus, there was a need to further
investigate immigrant crime in the Netherlands. The previous study found that
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
331
crime rates among the 57 origins studied were highly predictable from national IQ
(r = -.80) and moderately by Muslim % (r = .34). Second, the British intelligence
researcher and sociologist Noah Carl was recently the target of a harassment
campaign by left-wing activists and eventually fired from his job as a research
fellow at the University of Cambridge (Lehmann, 2019). One of the complaints
against him was that he had investigated the relationship between immigrant
crime rates by country of origin and public preferences regarding further
immigration from the same countries. Carl (2016) found that the crime rates
correlated r = .69 with net opposition (defined below) to immigration from the
countries, which suggests that public stereotypes about immigrant groups are
both fairly accurate and taken into account when forming immigration political
preferences. However, no published replication currently exists of his finding, so
we additionally sought to replicate it for the Netherlands.
Data
We used publicly available data about registered suspects among persons
living in the Netherlands (https://opendata.cbs.nl/statline/#/CBS/nl/dataset/
81959NED/table?ts=1569478057921). These data are compiled and published
by CBS (Centraal Bureau voor de Statistiek, Central Bureau of Statistics), which
is the official government body that publishes statistics for the Netherlands. These
were divided into groups by country of origin. Country of origin was defined as
including both first and second generation immigrants, i.e. persons who
themselves were born elsewhere, or whose parents were born elsewhere
(exception being if someone is born elsewhere but both parents are Dutch, for
instance, as part of longer foreign stays or medical tourism). Using these, we
calculated per capita suspect rates using population counts from
https://opendata.cbs.nl/statline/#/CBS/nl/dataset/37325/table?ts=156622104863
9 . The population count data were limited to persons aged 12 to 45, which is the
range who commits most recorded crimes. We used all available years of data,
spanning 2005 to 2018. In total, we have data from 70 immigrant countries as
well as the natives. Thus, we have a total of 994 country-years of data, which is
based on 8.3 million person-years of data. The new estimates were strongly
correlated with those used in the previous study: the correlation with the best
estimate from the previous study was .96. Furthermore, we downloaded data split
by generation, so as to investigate any potential confounding by distribution of
immigrant generation. Finally, we found another variable concerning arrests that
ran from 2005 to 2014. Since it had less data, we used it only for robustness
testing.
MANKIND QUARTERLY 2020 60:3
332
For country of origin variables, we used national IQs from Lynn and
Vanhanen’s 2012 dataset (Lynn & Vanhanen, 2012). This dataset has been
extensively used by other researchers for both country level analyses and
immigrant studies. Recently, German political scientist David Becker undertook
an independent analytic replication of the estimates by obtaining copies of each
original source and redoing all the calculations previously done by Richard Lynn
and colleagues. His work is presented in Lynn and Becker (2019) and is
continuously updated at http://viewoniq.org (currently in version 1.3.3). Many
other sets of estimates have been produced by others, most importantly Heiner
Rindermann (2018). However, we used the 2012 dataset because, as of writing,
it is more comprehensive than the 2019 recalculation (see discussion in
Kirkegaard, 2019c). For estimates of the proportion of Muslims in each group, we
used estimates from Pew Research (Pew Research Center, 2011). Countries with
missing data were imputed based on neighboring or component countries as
done in previous research (Kirkegaard & Becker, 2017).
We were unable to find a published survey with information about which
countries of origin Dutch people prefer and dislike as immigrants in their country.
For this reason, we sought to do our own small survey. Sample size was not
crucial for this because the differences between countries were expected to be
large, and we were only interested in estimating the mean preference for each
country. Using Prolific (https://www.prolific.co/) (Palan & Schitter, 2018), we
polled approximately 200 persons living in the Netherlands with regards to their
preferences for immigration from 68 immigrant countries we had crime data from.
We skipped two countries that no longer exist (Soviet Union and Czechoslovakia)
as well as the Netherlands itself, which isn’t an immigrant origin (and thus one
cannot have immigration preferences for it).
For each country of origin, subjects were asked “Thinking about people who
want to come and live in the Netherlands from different countries, to what extent
should people from the following countries be allowed to come and live in the
Netherlands?” with the available options of “none”, “fewer”, “same”, and “more”
(all in Dutch). This is the same format as used by YouGov when collecting the
data that Noah Carl used (Smith, 2016). Finally, we re-weighed the results by
party vote in the last election because our survey was tilted towards immigrant
friendly voters (e.g. Greens got 36% of the votes in our survey but 9.1% in the
2017 general election). This re-weighting did not affect the relative differences
much (r = .95 before and after). Prolific keeps track of whether their users provide
good data, and removes persons who provide poor data (i.e. click through
surveys very fast/at random). To ensure our participants were paying attention to
our somewhat tedious survey (they had to answer ~70 similar questions in a row,
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
333
one for each origin), we included two attention checks in the country list that asked
participants to select a particular response, and excluded participants who failed
these (n = 20, ~10%). This exclusion did not alter results noticeably.
Results
The differences in crime (suspect) rate by country of origin were large. The
lowest rate was seen for Northeast Asian countries, with Japan having a relative
rate of 0.21 to that of Dutch natives, while Netherlands Antilles had a relative rate
of 3.81. Thus, the relative difference between the most and least criminal groups
was about a factor of 17. Figure 2 shows a world map with the estimated crime
rates.
Figure 2. World map of relative crime (suspect) rates among immigrant groups
in the Netherlands by origin country (native = 1). Averaged from 2005 to 2018.
Grey indicates no data (few immigrants in the Netherlands).
Crime rates are generally falling in Western countries despite the influx of
above average crime rate immigrants, and indeed have been generally falling for
centuries (Pinker, 2012). This outcome is due to the fact that the crime rate among
the natives is falling sufficiently fast to offset the increase from immigration, so
that the net effect is negative (decreasing). Figures 3 and 4 show the timeline of
crime rates and relative rate versions for the 10 largest groups.
MANKIND QUARTERLY 2020 60:3
334
Figure 3. Timeline of crime (suspect) rates in the Netherlands by origin group,
10 largest groups.
Figure 4. Timeline of relative crime (suspect) rates in the Netherlands by origin
group, 10 largest groups. The Netherlands = 1.
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
335
In the period of study, the crime rate fell by 51% in the population as a whole,
and by an average of 50% within each origin group from 2005 to 2018. Still, the
relative group differences stayed approximately the same over the period of
study, r = .93 between 2005 and 2018 crime rates. This can also clearly be seen
in Figure 4 (above). The reason for the uptick in crime in 2010 is not known.
Figures 5 and 6 show the scatterplots for the two national-level predictors and the
crime rates.
The two predictors are correlated in the present sample (r = -.42, but only -
.27 worldwide), and thus their individual effect size is likely overestimated from
the bivariate analyses. For this reason, we fit a regression model with both
predictors. Adjusted R² was strong at .41 (i.e. model R = .64). In the weighted
model, the effect of IQ was stronger than that of Muslim origin: βIQ = -0.61 (SE =
0.12, p<.0001), βMuslim% = 0.16 (SE = 0.10, p = .12). In the unweighted model,
Muslim% did a little better (model adj. R² = .45, βIQ = -0.57 with p<.0001,
βMuslim% = 0.20 with p = .04).
Figure 5. Scatterplot of national IQ of origin country and relative crime rate
among origin groups in the Netherlands, 2005-2018. Weighted by the square root
of population size.
MANKIND QUARTERLY 2020 60:3
336
Figure 6. Scatterplot of Muslim percentage of origin country and relative crime
rate among origin groups in the Netherlands, 2005-2018. Weighted by the square
root of population size.
Figure 7 shows the scatterplot of crime rates and mean opposition. The
results replicate the general result found by Carl (2016). Suriname and
Netherlands Antilles stand out as strong outliers with large populations in the
Netherlands. Suriname is a former Dutch colonial possession in the north of
South America (gained independence in 1975), and Netherlands Antilles is a
current Dutch colonial possession consisting of several small islands north of
Suriname. A third ex-colonial possession is Indonesia, which has a low relative
crime rate (0.92, below natives) and faces low net opposition. However, even this
country still has a negative residual (standardized residuals are -1.96, -2.17, and
-0.08 for Suriname, Netherlands Antilles, and Indonesia, respectively). Thus, it
appears that people living in the Netherlands are willing to grant persons from
these countries relatively more or easier entrance to the Netherlands. If the
colonial origins are excluded, the correlations increase to .68 and .66, weighted
and unweighted, respectively. Another pattern that stands out is the dual clusters
of countries in the left. The bottom cluster consists of other traditional west
European countries (before the fall of USSR), while the upper one appears to be
a remainder category of countries that are more culturally distant but about
equally low in crime rates.
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
337
Figure 7. Scatterplot of relative crime rate and net opposition to origin groups.
Weighted by the square root of population size. Unweighted r = .57.
We can combine these insights into a series of regression models, the results
of which are summarized in Table 1. Colonial is a dummy variable for whether the
origin country is Suriname, Netherlands Antilles, or Indonesia. We included
regional dummies, one based on continents and one based on macroregions.
The macroregions were copied from a previous study and based on UN
classifications (Kirkegaard, 2019a). Maps with the classification schemes are
given in the appendix. Because of the central role of Islam/Muslims in public
debate, we included this variable in the regressions. We did not include national
IQs as these have at best a minor role in public debates about immigration in
Europe.
MANKIND QUARTERLY 2020 60:3
338
Table 1. Model results for predicting public net opposition to immigration from
particular origin countries. N = 68. Numerical variables are standardized, value in
parentheses = standard error. * = p<.01, ** = p<.005, *** = p<.001.
Predictor/Model
1
2
3
4
5
Intercept
-0.07
(0.114)
0.08
(0.107)
-0.01
(0.083)
-0.37*
(0.136)
-0.88***
(0.129)
Suspect rate
0.44***
(0.105)
0.55***
(0.097)
0.33***
(0.081)
0.37***
(0.099)
0.21
(0.089)
Colonial
-1.32***
(0.308)
-1.20***
(0.236)
-1.20***
(0.248)
-1.12***
(0.233)
Muslim
0.50***
(0.074)
0.27*
(0.094)
0.35***
(0.096)
UN_continent=Europe
0
(ref)
UN_continent=Africa
0.66
(0.269)
UN_continent=Americas
0.18
(0.249)
UN_continent=Asia
0.83***
(0.222)
UN_continent=Oceania
-0.71
(0.514)
UN_macroregion=N & W Europe +
offshoots
0
(ref)
UN_macroregion=Caribbean
0.89
(0.394)
UN_macroregion=Latin America
0.85***
(0.223)
UN_macroregion=Africa
1.55***
(0.245)
UN_macroregion=Eastern Asia
0.95***
(0.256)
UN_macroregion=Eastern Europe
1.71***
(0.225)
UN_macroregion=MENA
1.18***
(0.291)
UN_macroregion=South-Eastern Asia
1.11***
(0.240)
UN_macroregion=Southern Asia
1.47
(0.276)
UN_macroregion=Southern Europe
0.54
(0.221)
R
2
adj.
0.198
0.365
0.628
0.697
0.841
N
68
68
68
68
68
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
339
First, we see that in all models, higher crime rate predicts more opposition
(all betas are positive). As expected, adding the colonial dummy increased the
beta for crime rate (β = 0.44→0.55, models 1→2). Second, Muslim% was also a
predictor beyond the crime rates (β = 0.27 to 0.50, models 3-5). Adding controls
for continents decreased the effect size for Muslim% = 0.50→0.27, model
3→4) but not crime rate (β = 0.33→0.37). Adding macroregions decreased the
effect size of both crime rate = 0.33→0.21, model 3→5) and Muslim
percentage (β = 0.50→0.35). Not too much can be made of some of the smaller
changes because the standard errors are fairly large given the sample size of 68.
As a robustness test, we examined the correlations using the crime rates
computed for each immigrant generation. Table 2 shows the correlations among
the variables.
Table 2. Correlations among main variables. 1st and 2nd refer to immigrant
generations. Prop 2nd = proportion of population who is second generation
immigrant. Weighted correlations below the diagonal.
Suspect
rate
st
Suspect
rate 2
nd
Muslim IQ
Net
opposition
Prop
2
nd
Suspect rate
0.96 0.90 0.44 -0.66 0.57 -0.08
Suspect rate
1st 0.95 0.80 0.39 -0.66 0.58 -0.24
Suspect rate
2nd 0.91 0.76 0.43 -0.59 0.58 -0.09
Muslim
0.45
0.43
-0.42
0.66
-0.07
IQ
-0.64 -0.56 -0.55 -0.52 -0.71 0.30
Net
opposition 0.46 0.41 0.53 0.68 -0.63 -0.48
Prop 2
nd
-0.02
-0.02
0.10
0.18
-0.43
The results broken down by immigrant generation were quite similar,
sometimes a bit weaker. This does not necessarily imply a confound because the
correlations by generation are based on smaller samples which produce noisier
estimates for each origin group, and thus weaker correlations.
We conducted several other robustness tests. First, we compared results
using Rindermann’s national IQ estimates. These were calculated independently
from Lynn and Becker’s calculations, and give more weight to the scholastic
ability (school) tests such as PISA. The results however were mostly the same.
Second, we computed a raw version of the net opposition metric, without
weighting to control for over-sampling of left-wing voters. This however correlated
.96 with the weighted version and did not produce any notable differences. Third,
MANKIND QUARTERLY 2020 60:3
340
we examined the main scatterplots for evidence of nonlinearity. However, the
results indicated a general lack of nonlinearity in that the error bars of the
smoothed fit (LOESS) almost always overlapped the linear fit. Fourth, by reviewer
request, we fit models containing an interaction term between colonial status and
crime rate. These models produced similar results for our main variable of interest
(crime rate) and varying estimates for colonial heritage and its interaction with
crime rate, but with large standard errors which limit their interpretation. Fifth, we
ran the regressions without the imputed Muslim%. This did not result in any
notable differences. Sixth, we ran analyses with our alternative crime measure,
arrest rates. However, these correlated .995, so the results were practically
identical. Full output from these robustness tests can be found in the
supplementary materials.
Discussion
We found that differences in crime rates between origin-based immigration
groups were large, up to a factor of 17 between the most criminal (Netherlands
Antilles) and the least (Japan). The relative differences in crime rates were well-
predicted by the origin characterstics of national IQ and Muslim percentage as
well as their combination, as found in many previous studies (Kirkegaard, 2017)
as well as the previous Dutch study (Kirkegaard, 2015). Population-wide, the
crime rate has been decreasing in the Netherlands and many other countries for
hundreds of years (Pinker, 2012). This decrease still seems to be happening in
northern and western European countries despite growing immigrant populations
with above average crime rates. This seeming contradiction is explained in part
by the aging of the native populations combined with the lower crime rates of the
elderly, which more than offsets the increase from the rising numbers of younger
and more crime-prone immigrants. However, some categories of crime show
some recent upticks in relation to the migrant wave of 2015-2017, in particular
rapes and other violent crimes (Pallesen, 2018; Sanandaji, 2017).
We furthermore conducted a replication study of Carl (2016), who studied
the relationship between crime rates and immigrant preferences. Specifically,
Carl reasoned that sensible voters would base their immigration preferences
partly on variables such as crime rates, and to the extent voters are aware of real
group differences in crime rates, their preferences will be correlated with the real
crime rates. He found this to be the case (r = .69) for the United Kingdom using
two surveys with an overlapping set of 23 countries of origin. In the present study,
we had access to a much larger set of 68 countries. We also find that our
estimated crime rates are moderately strongly related to immigrant preferences,
population weighted r = .46 and unweighted r = .57. The pattern was weaker than
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
341
the UK results, mainly due to the Dutch colonial (ex-)possessions. The public
seems to be willing to grant these some leeway with allowing immigration despite
the high crime rates of some of them, perhaps as a sort of reparation payment.
We furthermore ran a number of regressions to see if the predictive power of
crime rate was merely due to some obvious confounding factor. However, we
found that it retains much validity in the face of plausible confounders. A diagram
of our conceptual causal model is shown in Figure 8.
Figure 8. Conceptual causal model of immigrant crime rates, perceptions and
preferences.
In the model, origin populations are assumed to vary in traits such as
intelligence, time-preference, testosterone level and so on. These traits are then
MANKIND QUARTERLY 2020 60:3
342
brought with the immigrants when they migrate to a new country (spatial
transferability), modified by selection effects. After this, the different immigrant
groups in the destination country also vary in their traits, and this gives rise to
variations in crime rates, which are also influenced by contextual factors such as
age and sex composition, duration of stay, cultural conflicts and so on. These real
differences in crime rates then give rise to stereotypes, understood as subjective
perceptions about group differences (Jussim, 2018). The perceptions are also
caused by media reports and various proxies. Few people memorize detailed
reports of crime rates, so they rely on proxies such as geographical location of
countries or their wealth levels. Finally, these perceptions cause people to modify
their preferences for immigrants from specific countries, which is also affected by
special relationships between the countries (e.g. colonial history). The current
study did not measure all the relevant variables, but focused on three of the
variables in the main (middle column) pathway, and showed that they were
related as expected from the model. Furthermore, the addition of detailed
geographical proxies (model 5 in Table 1) reduced the relationship between crime
rates and preferences, in line with the model (since people rely on proxies in part).
Limitations
First, we only investigated one broad outcome, crime, whereas many others
are possible (e.g. use of social welfare) and plausibly affect public perception as
well. However, according to survey data from the UK cited by Carl (2016), violent
crime proneness is the most important variable people consider.
Second, related to the first, crime rates are difficult to estimate empirically
since these are rare events (being the suspect of a crime) and thus unstable in
small populations. Furthermore, insofar as the goal is to estimate criminal
propensity of a population, one will need to adjust for the age and sex distributions
of the population which are generally considered exogenous variables. It was not
possible to do so completely in our data. We used population data for age 12-45
since this is the main age span where people commit crimes, but a prior study of
Danish and German data showed that more detailed adjustments matter little for
relative differences in crime rates between immigrant populations (Kirkegaard &
Becker, 2017). The very strong correlation of our crime rates with those from the
prior study (r = .96) suggests age and sex confounding is not a large problem in
this case either.
Third, our survey of Dutch persons to estimate the immigration preferences
was smaller than typical public surveys owing to cost limitations, and furthermore
was not politically representative. We adjusted for the political representation
using the party vote in the last election, but this did not seem to affect results
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
343
much. It is possible it was unrepresentative in other ways we did not study. We
did not have age, sex, education information about the subjects, so we were
unable to calculate representativeness in terms of these.
Fourth, our use of national origin country-level data assumes that the
immigrant groups are representative of their origin populations, and that we have
reliable estimates of the origin country’s characteristics themselves. The
estimates of national IQs, based on Richard Lynn’s work, have in particular been
questioned (Hunt & Sternberg, 2006; Wicherts et al., 2010). Recently, however,
David Becker independently redid all the data extraction from the original sources
as well as every calculation (Flynn effect adjustment, age adjustments, quality
weights etc.). His work is described in a recent book coauthored with Lynn (Lynn
& Becker, 2019). Generally speaking, the estimates can be considered quite
reliable for many countries, but not all, and much work remains to be done
examining questionable aspects of measurement invariance (Dutton et al., 2018;
Kirkegaard, 2019c).
Aside from the question of the country estimates, it is well known that
immigrants tend to be self-selected on traits evident inside their countries of origin
(non-random emigration) (Aksoy & Poutvaara, 2019; Connor, 2019; Hamilton &
Hummer, 2011; Knudsen, 2019) and in their choice of destination country, and by
the need to obtain legal rights to live in the destination country (non-random
immigration). It is possible to account for some such selectivity factors, for
instance using the Brain Drain dataset (https://www.iab.de/en/daten/iab-brain-
drain-data.aspx), but it requires a more complex approach than used here and is
left for future research (see for example Fuerst & Kirkegaard, 2014). It should be
noted, however, that selection would probably have to be unrealistically strong to
overcome the national differences (see also Pesta et al. (2019) with regards to
potential immigrants and GRE testing scores). Furthermore, selection that is
similar across sending countries (e.g. everybody sends elites, say, the top 5%)
does not alter conclusions from the kinds of analyses in this study since they don’t
affect relative differences between groups, only the intercept. This conclusion
depends on the additional assumption of equal variances across sending
countries, which is known not to be entirely true (Kirkegaard & Tranberg, 2015;
Meisenberg, 2008; Rindermann, 2018, Chapter 8). However, generally speaking,
the point stands that for selection to notably affect the results, it must be in the
form of differential selection, either from the sending countries themselves (e.g.
country A sends higher class people and country B sends lower class people), or
among incoming immigrants in the host country (e.g. a host country decides to
accept higher class people from country A but lower class from country B).
MANKIND QUARTERLY 2020 60:3
344
Supplementary materials
R analysis code and full dataset are available at https://osf.io/46pwz/ and at
https://rpubs.com/EmilOWK/Dutch_immigrant_crime_2019.
Acknowledgements
We wish to thank an anonymous sponsor without whom this research would
not be possible. The sponsor had no influence on the study design or publication.
References
Aksoy, C.G. & Poutvaara, P. (2019). Refugees’ self-selection into Europe: Who migrates
where? (SSRN Scholarly Paper ID 3373580). Social Science Research Network.
Andersson, M. & Jespersen, N. (2018). Eksperimentet, der slog fejl. Gads forlag.
https://gad.dk/eksperimentet
Blom, M., Oudhof, J., Bijl, R.V. & Bakker, B.F. (2005). Verdacht van criminaliteit:
Allochtonen en autochtonen nader bekeken. Cahiers. http://www.narcis.nl/publication/
RecordID/oai:tudelft.nl:uuid%3A099b3cc7-2616-4050-992b-28ac8d4c6c4b
Borjas, G.J. (2016). We Wanted Workers: Unraveling the Immigration Narrative, 1st
edition. W.W. Norton & Co.
Carl, N. (2016). Net opposition to immigrants of different nationalities correlates strongly
with their arrest rates in the UK. Open Quantitative Sociology & Political Science.
https://openpsych.net/paper/48
Carl, N. (2017). Ethnicity and electoral fraud in Britain. Electoral Studies 50: 128-136.
Connor, D.S. (2019). The cream of the crop? Geography, networks, and Irish migrant
selection in the age of mass migration. Journal of Economic History 79: 139-175.
Dutton, E., Becker, D., Osman, H.A., Bakhiet, S.F., Essa, Y.A.S., Ali, H.A.A., Alqafari,
S.M., Hamdi, A.H.M.N. & Alfaleh, A.S.H. (2018). The Raven’s test performance of South
Sudanese samples: A validation of criticisms of the utility of Raven’s among Sub-Saharan
Africans. Personality and Individual Differences 128: 122-126.
Frisell, T., Pawitan, Y. & Långström, N. (2012). Is the association between general
cognitive ability and violent crime caused by family-level confounders? PLoS ONE 7(7):
e41783.
Fuerst, J. & Kirkegaard, E.O.W. (2014). Do national IQs predict U.S. immigrant cognitive
ability and outcomes? An analysis of the National Longitudinal Survey of Freshmen. Open
Differential Psychology. https://openpsych.net/paper/20
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
345
Hamilton, T.G. & Hummer, R.A. (2011). Immigration and the health of U.S. black adults:
Does country of origin matter? Social Science & Medicine 73: 1551-1560.
Hegelund, E.R., Flensborg-Madsen, T., Dammeyer, J. & Mortensen, E.L. (2018). Low IQ
as a predictor of unsuccessful educational and occupational achievement: A register-
based study of 1,098,742 men in Denmark 1968-2016. Intelligence 71: 46-53.
Hegelund, E.R., Flensborg-Madsen, T., Dammeyer, J., Mortensen, L.H. & Mortensen,
E.L. (2019). The influence of familial factors on the association between IQ and
educational and occupational achievement: A sibling approach. Personality and Individual
Differences 149: 100-107.
Herrnstein, R.J. & Murray, C.A. (1994). The Bell Curve: Intelligence and Class Structure
in American Life, 1st ed. New York: Simon & Schuster.
Hesson, L. (2019). What happened last night in Sweden? Analysis of Western news
media portrayal of crime in Sweden. http://muep.mau.se/handle/2043/29044
Hill, W.D., Davies, N.M., Ritchie, S.J., Skene, N.G., Bryois, J., Bell, S., Angelantonio, E.D.,
…, & Deary, I.J. (2019). Genome-wide analysis identifies molecular systems and 149
genetic loci associated with income. Nature Communications 10(1): 1-16.
Hunt, E. & Sternberg, R.J. (2006). Sorry, wrong numbers: An analysis of a study of a
correlation between skin color and IQ. Intelligence 34: 131-137.
Jones, G. & Schneider, W.J. (2010). IQ in the production function: Evidence from
immigrant earnings. Economic Inquiry 48: 743-755.
Jussim, L. (2018). The Accuracy of Demographic Stereotypes. PsyArXiv
https://doi.org/10.31234/osf.io/beaq3
Kirkegaard, E.O.W. (2013). Predicting immigrant IQ from their countries of origin, and
Lynn’s national IQs: A case study from Denmark. Mankind Quarterly 54: 151-167.
Kirkegaard, E.O.W. (2015). Crime among Dutch immigrant groups is predictable from
country-level variables. Open Differential Psychology.
Kirkegaard, E.O.W. (2017, May 12). Differential immigrant group performance: A matter
of intelligence? Abstract: London Conference on Intelligence 2017.
https://www.youtube.com/watch?v=jMSRZQ1usDc
Kirkegaard, E.O.W. (2019a). Is national mental sport ability a sign of intelligence? An
analysis of the top players of 12 mental sports. Mankind Quarterly 59: 296-311.
Kirkegaard, E.O.W. (2019b). First names, cognitive ability and social status in Denmark.
Mankind Quarterly 59: 521-531.
Kirkegaard, E.O.W. (2019c). Solid numbers, missed opportunities: Review of The
Intelligence of Nations. Evolutionary Behavioral Sciences. https://doi.org/
10.1037/ebs0000181
MANKIND QUARTERLY 2020 60:3
346
Kirkegaard, E.O.W. & Becker, D. (2017). Immigrant crime in Germany 2012-2015. Open
Quantitative Sociology & Political Science, 1(1). https://doi.org/10.26775/
OQSPS.2017.02.11
Kirkegaard, E.O.W. & Fuerst, J. (2014). Educational attainment, income, use of social
benefits, crime rate and the general socioeconomic factor among 70 immigrant groups in
Denmark. Open Differential Psychology. https://openpsych.net/paper/21
Kirkegaard, E.O.W. & Tranberg, B. (2015). Increasing inequality in general intelligence
and socioeconomic status as a result of immigration in Denmark 1980-2014 |.
https://openpsych.net/paper/33
Knudsen, A.S.B. (2019). Those who stayed: Individualism, self-selection and cultural
change during the age of mass migration (SSRN Scholarly Paper ID 3321790). Social
Science Research Network. https://papers.ssm.com/abstract=3321790
Kuran, T. (2018). Islam and economic performance: Historical and contemporary links.
Journal of Economic Literature 56: 1292-1359.
Lehmann, C. (2019, May 2). Cambridge capitulates to the mob and fires a young scholar.
Quillette https://quillette.com/2019/05/02/cambridge-capitulates-to-the-mob-and-fires-a-
young-scholar/
Lynn, R. & Becker, D. (2019). The Intelligence of Nations. Ulster Institute for Social
Research.
Lynn, R. & Vanhanen, T. (2012). Intelligence: A Unifying Construct for the Social
Sciences. Ulster Institute for Social Research.
Meisenberg, G. (2008). How does racial diversity raise income inequality? Journal of
Social Political and Economic Studies 33: 3-26.
MG (2015, December 3). Why we culturally profile. Those Who Can See
http://thosewhocansee.blogspot.com/2015/12/why-we-culturally-profile.html
MG (2016, December 13). The diversity tax. Those Who Can See.
http://thosewhocansee.blogspot.com/2016/12/the-diversity-tax.html
Murray, C. (2002). IQ and income inequality in a sample of sibling pairs from advantaged
family backgrounds. American Economic Review 92(2): 339-343.
Murray, D. (2017). The Strange Death of Europe: Immigration, Identity, Islam, 1st ed.
Bloomsbury Continuum.
Palan, S. & Schitter, C. (2018). Prolific.acA subject pool for online experiments. Journal
of Behavioral and Experimental Finance 17: 22-27.
Pallesen, J. (2018, June 10). Crime and immigration in Germany.
http://jsmp.dk/posts/2018-06-11-germanycrime/
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
347
Pesta, B.J., Fuerst, J., Kirkegaard, E.O.W. & Papaleo, B. (2019). Does intelligence
explain national score variance on graduate admissions exams? Intelligence 73: 8-15.
Pew Research Center (2011). Table: Muslim Population by Country. Pew Research
Center. http://pewforum.org/2011/01/27/table-muslim-population-by-country/
Pickering, S. & Ham, J. (eds.) (2015).
Pinker, S. (2012). The Better Angels of Our Nature: Why Violence Has Declined. PenThe
Routledge Handbook on Crime and International Migration, 1st ed. Routledge, Taylor &
Francis Group.guin Books USA.
Rindermann, H. (2018). Cognitive Capitalism: Human Capital and the Wellbeing of
Nations. University Printing House.
Rindermann, H. & Thompson, J. (2016). The cognitive competences of immigrant and
native students across the world: An analysis of gaps, possible causes and impact.
Journal of Biosocial Science 48: 66-93.
Robie, C., Christiansen, N.D., Hausdorf, P.A., Murphy, S.A., Fisher, P.A., Risavy, S.D. &
Keeping, L.M. (2017). International comparison of group differences in general mental
ability for immigrants versus non-immigrants. International Journal of Selection and
Assessment 25: 347-359.
Roth, B.M. (2010). The Perils of Diversity: Immigration and Human Nature, 1st ed.
Washington Summit.
Salmi, V., Kivivuori, J. & Aaltonen, M. (2015). Correlates of immigrant youth crime in
Finland. European Journal of Criminology 12: 681-699.
Sanandaji, T. (2017). Massutmaning: Ekonomisk politik mot utanförskap och antisocialt
beteende. Kuhzad Media.
Sarrazin, T. (2012). Deutschland schafft sich ab: Wie wir unser Land aufs Spiel setzen,
1. Aufl. Dt. Verl.-Anst.
Sarrazin, T. (2018). Feindliche Übernahme: Wie der Islam den Fortschritt behindert und
die Gesellschaft bedroht. FinanzBuch Verlag.
Schulz, J.F., Bahrami-Rad, D., Beauchamp, J.P. & Henrich, J. (2019). The Church,
intensive kinship, and global psychological variation. Science 366(6466), eaau5141.
https://doi.org/10.1126/science.aau5141
Smith, M. (2016, August 26). If voters designed a points-based immigration system... |
YouGov. YouGov. https://yougov.co.uk/topics/politics/articles-reports/2016/08/26/one-in-
five-say-uk-should-not-admit-single-migrant-tu
Vinogradov, E. & Kolvereid, L. (2010). Home country national intelligence and self-
employment rates among immigrants in Norway. Intelligence 38: 151-159.
MANKIND QUARTERLY 2020 60:3
348
Whitaker, S. (2018). Assessing the intellectual ability of asylum seekers. International
Journal of Developmental Disabilities 64: 309-317.
Wicherts, J.M., Dolan, C.V. & van der Maas, H.L.J. (2010). A systematic literature review
of the average IQ of sub-Saharan Africans. Intelligence 38: 1-20.
Appendix
Maps of regional coding. Continents and regions are given by the UN, and
macroregions are an aggregate based on the UN regions. The regions were not
used in this study, but given here for reference.
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
349
Main data table. Average values across years given for population and crime
rates.
Origin Cont. Macroreg. Pop. Sus
rate
Sus
rate
RR
Sus
rate
1st
RR
Sus
rate
2nd
RR
Muslim
IQ Net
opp
Afghanistan
Asia
S. Asia
25016.14
0.04
2.05
2.1
1.37
1
75
-0.06
Algeria
Africa
MENA
4219.5
0.07
3.24
2.73
3.64
0.98
84.2
-0.02
Angola
Africa
Africa
5755.86
0.07
3.52
3.65
2.76
0.01
71
-0.06
Argentina
Americas
Latin
America
2669 0.01 0.65
0.26 1.24 0.02 92.8 -0.58
Australia
Oceania
N & W
Europe +
offshoots
10573 0.02 1.16
0.36 1.53 0.02 99.2 -0.74
Austria
Europe
N & W
Europe +
offshoots
5951.36 0.02 1.11
0.77 1.33 0.06 99 -0.72
Belgium
Europe
N & W
Europe +
offshoots
41977.07 0.02 1.23
0.95 1.42 0.06 99.3 -0.71
Brazil
Americas
Latin
America
11406 0.02 1.15
1.07 1.39 0 85.6 -0.52
Bulgaria
Europe
E. Europe
11523.07
0.03
1.55
1.48
3.01
0.13
93.3
-0.09
Canada
Americas
N & W
Europe +
offshoots
9540.14 0.02 0.88
0.41 1.13 0.03 100.4 -0.68
Cape Verde
Africa
Africa
11495.93
0.06
3.08
2.44
3.52
0
76
-0.34
Chile
Americas
Latin
America
2752.79 0.03 1.65
1.17 2.16 0 89.8 -0.49
China
Asia
E. Asia
36051.57
0.01
0.49
0.48
0.51
0.02
105.8
-0.38
Colombia
Americas
Latin
America
7189.14 0.04 1.83
1.66 2.27 0 83.1 -0.49
MANKIND QUARTERLY 2020 60:3
350
Origin Cont. Macroreg. Pop. Sus
rate
Sus
rate
RR
Sus
rate
1st
RR
Sus
rate
2nd
RR
Muslim
IQ Net
opp
Congo (D. R.)
Africa
Africa
4618.14
0.06
3.07
3.29
2.49
0.01
68
-0.09
Denmark
Europe
N & W
Europe +
offshoots
2965.07 0.01
0.7 0.39 1.1 0.04 97.2 -0.72
Dominican
Republic
Americas Caribbean 6864.57 0.07
3.53
3.25 4.35 0 82 -0.38
Egypt
Africa
MENA
10597.86
0.04
1.94
1.4
2.6
0.95
82.7
-0.2
Ethiopia
Africa
Africa
8016.57
0.04
2.03
1.72
3.04
0.34
68.5
0.01
Finland
Europe
N & W
Europe +
offshoots
2604.14 0.01
0.5 0.16 1.18 0.01 100.9 -0.61
Former
Czechoslovakia
Europe E. Europe 8049.86 0.02
0.96
0.84 1.32 0 98.6
Former Soviet
Union
Europe E. Europe 36559.14 0.03
1.53
1.53 1.61 0.05 96.6
Former
Yugoslavia
Europe S. Europe 43570.57 0.04
2.05
1.84 2.49 0.08 92.33 -0.16
France
Europe
N & W
Europe +
offshoots
21128.21 0.02
0.93
0.68 1.26 0.08 98.1 -0.71
Germany
Europe
N & W
Europe +
offshoots
107339.5 0.02
1.04
0.73 1.27 0.05 98.8 -0.73
Ghana
Africa
Africa
10561
0.05
2.27
1.62
3.42
0.16
69.7
0.01
Guyana
Americas
Latin
America
2099.79 0.05
2.44
1.91 2.82 0.07 81 -0.14
Hong Kong
Asia
E. Asia
9935
0.01
0.55
0.46
0.59
0.01
105.7
-0.49
Hungary
Europe
E. Europe
8932.21
0.02
0.94
0.78
1.34
0
98.1
-0.36
India
Asia
S. Asia
14997.43
0.01
0.64
0.44
1.51
0.15
82.2
-0.23
Indonesia
Asia
S.E. Asia
147590.21
0.02
0.92
0.46
0.97
0.88
85.8
-0.47
Iran
Asia
S. Asia
18965.86
0.04
2.19
2.22
1.98
1
85.6
-0.09
Iraq
Asia
MENA
29432.5
0.05
2.4
2.41
2.35
0.99
87
0
Ireland
Europe
N & W
Europe +
offshoots
4658.93 0.02
1.06
0.58 1.6 0.01 94.9 -0.67
Israel
Asia
MENA
4900
0.02
0.98
0.75
1.31
0.18
94.6
-0.37
Italy
Europe
S. Europe
23803.36
0.02
1.23
0.8
1.63
0.03
96.1
-0.63
Japan
Asia
E. Asia
4322.5
0
0.22
0.14
0.59
0
104.2
-0.67
Lebanon
Asia
MENA
2913.29
0.05
2.3
2.19
2.5
0.6
84.6
-0.15
Malaysia
Asia
S.E. Asia
2840.79
0.01
0.57
0.34
0.81
0.61
91.7
-0.29
Mexico
Americas
Latin
America
2914.29 0.01
0.34
0.32 0.41 0 87.8 -0.38
Morocco
Africa
MENA
199533.64
0.07
3.66
2.49
4.74
1
82.4
-0.01
Netherlands
Antilles
Americas Caribbean 81888.5 0.08
3.81
4.4 2.98 0 87 -0.57
New Zealand
Oceania
N & W
Europe +
offshoots
3428.21 0.02
1.06
0.4 1.44 0.01 98.9 -0.77
Nigeria
Africa
Africa
5698.14
0.04
2.03
1.91
2.28
0.48
71.2
-0.03
Norway
Europe
N & W
Europe +
offshoots
2245.07 0.01
0.55
0.21 1.05 0.03 97.2 -0.66
Pakistan
Asia
S. Asia
11386.14
0.03
1.54
1.26
1.94
0.96
84
0.06
KIRKEGAARD, E.O.W. & DE KUIJPER, M. PUBLIC PREFERENCES AND REALITY
351
Origin Cont. Macroreg. Pop. Sus
rate
Sus
rate
RR
Sus
rate
1st
RR
Sus
rate
2nd
RR
Muslim
IQ Net
opp
Peru
Americas
Latin
America
2908 0.02 0.85
0.61 1.46 0 84.2 -0.37
Philippines
Asia
S.E. Asia
9719.86
0.01
0.74
0.47
1.27
0.05
86.1
-0.35
Poland
Europe
E. Europe
61822.07
0.02
1.22
1.2
1.42
0
96.1
-0.21
Portugal
Europe
S. Europe
12607.5
0.03
1.45
1.42
1.51
0.01
94.4
-0.71
Romania
Europe
E. Europe
10918.64
0.02
1.09
1.03
1.61
0
91
-0.27
Sierra Leone
Africa
Africa
4261.57
0.06
2.95
2.98
2.4
0.72
64
-0.02
Singapore
Asia
S.E. Asia
2429.36
0.01
0.57
0.24
0.87
0.15
107.1
-0.46
Somalia
Africa
Africa
17921.36
0.06
2.95
2.8
4.12
0.99
72
0.08
South Africa
Africa
Africa
10580.64
0.02
0.89
0.69
1.14
0.02
71.6
-0.54
South Korea
Asia
E. Asia
3059.43
0.01
0.51
0.24
1.52
0
104.6
-0.55
Spain
Europe
S. Europe
21226.57
0.02
1.07
0.65
1.52
0.02
96.6
-0.76
Sri Lanka
Asia
S Asia
5509.64
0.03
1.39
1.3
1.69
0.08
79
-0.15
Sudan
Africa
MENA
3998.5
0.04
1.83
1.8
2.23
0.71
77.5
0.05
Suriname
Americas
Latin
America
182732.79
0.06 2.79
2.47 3.01 0.16 89 -0.65
Sweden
Europe
N & W
Europe +
offshoots
3577.36 0.01 0.61
0.45 0.92 0.05 98.6 -0.77
Switzerland
Europe
N & W
Europe +
offshoots
5092.64 0.02 0.74
0.61 0.88 0.06 100.2 -0.72
Syria
Asia
MENA
13836
0.03
1.57
1.5
2.29
0.93
82
-0.12
Thailand
Asia
S.E. Asia
9868.64
0.02
0.9
0.66
1.87
0.06
93.9
-0.43
Netherlands
Europe
N & W
Europe +
offshoots
4826127 0.02 1 0.06 100.4
Tunisia
Africa
MENA
4871.14
0.06
3.1
2.63
3.46
1
85.4
-0.13
Turkey
Asia
MENA
230197
0.04
2.08
1.57
2.55
0.99
89.4
-0.03
UK
Europe
N & W
Europe +
offshoots
39369 0.02 1.03
0.71 1.37 0.05 99.1 -0.67
USA
Americas
N & W
Europe +
offshoots
18612.29 0.02 0.78
0.55 1.21 0.01 97.5 -0.52
Venezuela
Americas
Latin
America
3378.5 0.02 1.21
1.32 1.05 0 83.5 -0.34
Vietnam
Asia
S.E. Asia
11108.5
0.02
1.08
1.14
1
0
91.4
-0.45
Cont.= UN continent; Macroreg.= UN macroregion; Sus. = Suspect; Net opp. = Net
opposition.
... There were 6 biology, 4 math/statistics, 2 economics, 1 history, 2 psychology/psychiatry, 2 linguistics, 2 physics, and 1 geography questions. countries from a recent study of immigrant crime rates (Kirkegaard & de Kuijper, 2020). These estimates should be highly reliable, as they are based on public data published by the government, and thus suitable as criterion data (Jussim, 2012). ...
... Student households) for the 12 provinces of the Netherlands. 6 For immigrant preferences data, we reused data from a prior study that concerned the same 68 origins (Kirkegaard & de Kuijper, 2020). The subjects in this prior dataset overlapped with the current ones to some extent, but a prior study found that sample overlap between subjects asked about preference and crime stereotypes did not affect results . ...
... In terms of immigration opinions, a prior study measured the preferences for the same origin groups in a sample of 200 people living in the Netherlands (Kirkegaard & de Kuijper, 2020), partially overlapping with the present. Figure 18 shows the results. ...
Article
Full-text available
In this pre-registered study, we gathered two online samples totaling 615 subjects. The first sample was nationally representative with regards to age, sex and education, the second was an online convenience sample with mostly younger people. We measured intelligence (vocabulary and science knowledge, 20 items each) using newly constructed Dutch language tests. We measured stereotypes in three domains: 68 national origin-based immigrant crime rates, 54 occupational sex distributions, and 12 provincial incomes. We additionally measured other covariates such as employment status and political voting behaviors. Results showed substantial stereotype accuracy for each domain. Aggregate (average) stereotype Pearson correlation accuracies were strong: immigrant crime .65, occupations .94, and provincial incomes .85. Results of individual accuracies found there was a weak general factor of stereotype accuracy measures, reflecting a general social perception ability. We found that intelligence moderately but robustly predicted more accurate stereotypes across domains as well as general stereotyping ability (r’s .20, .25, .26, .39, β’s 0.17, 0.25, 0.21, 0.37 from the full regression models). Other variables did not have robust effects across all domains, but had some reliable effects for one or two domains. For immigrant crime rates, we also measured the immigration preferences for the same groups, i.e. whether people would like more or fewer people from these groups. We find that actual crime rates predict net opposition at r = .55, i.e., subjects were more hostile to immigration from origins that had higher crime rates. We examined a rational immigration preference path model where actual crime rates→stereotypes of crime rates→immigrant preferences. We found that about 84% of the effect of crime rates was mediated this way, and this result was obtained whether or not one included Muslim% as a covariate in the model. Overall, our results support rational models of social perception and policy preferences for immigration.
... These studies show that immigrant populations usually, but not always, perform below the native population. Furthermore, country of origin characteristics such as Muslim percentage of the population and average intelligence in the country (measured as IQ or as scores on scholastic achievement tests such as PISA) strongly predict outcomes (mean absolute correlation around .60) and plausibly serve as causal variables that country of origin is a proxy of (Kirkegaard & de Kuijper, 2020). ...
... Decreases in slopes from adding spatial lags can indicate a successful control for unmeasured confounding variables, but it can also indicate measurement error in 2 The association of Muslim immigrants with crime is long-standing and of great interest to the public according to survey data. While associations in the USA are weaker, they are quite strong in European countries (Ahlberg, 1996;Carl, 2016;Junger & Polder, 1992;Kirkegaard, 2014b;Kirkegaard & Becker, 2017;Kirkegaard & de Kuijper, 2020;Skardhamar et al., 2014). the predictor. ...
Article
Full-text available
We examined regional inequality in Belgium, both in the 19 communes of Brussels and in the country as a whole (n = 589 communes). We find very strong relationships between Muslim% of the population and a variety of social outcomes such as crime rate, educational attainment, and median income. For the 19 communes of Brussels, we find a correlation of-.94 between Muslim% and a general factor of socioeconomic variables (S factor) based on 22 diverse indicators. The slope for this relationship is-7.52, meaning that a change in S going from 0% to 100% Muslim corresponds to a worsening of overall social well-being by 7.52 (commune-level) standard deviations. For the entire country, we have data for 8 measures of social inequality. Analysis of the indicators shows an S factor which is very similar to the one from the Brussels data only based on the full set of indicators (r's = .98). In the full dataset, the correlation between S and Muslim% is-.52, with a slope of-8.05. Adding covariates for age, population density, and spatial autocorrelation changes this slope to-8.77. Thus, the expected change going from a 0% to 100% Muslim population is-8.77 standard deviations in general social well-being. We discuss our findings in relation to other research on immigration and social inequality, with a focus on the causal influence of intelligence on life outcomes in general.
... These real differences give rise to descriptive stereotypes (i.e., subjective perceptions about group differences). Few individuals are able to memorize detailed reports about the groups, so they rely on proxies such as the geographical location of countries or their wealth levels (see Kirkegaard & de Kuijper, 2020;Kirkegaard & Fedorov, 2021). ...
Article
The study extends the approach of the Stereotype Content Model to ethnic stereotype content beyond intergroup relations within societies by exploring the North-South hypothesis for competence and warmth. This paper claims that the “desperate” (resource-poor and unpredictable) of lower-latitude climate regions and “hopeful” (resource-sufficient and stable) ecology higher-latitude climate regions translate into typical aggregate attributes and are afterward generalized to the status of all their residents. Further, people use this information as a diagnostic for judgments about the economic value or burden of ethnic groups in their society. Based on the data about aggregated means of competence and warmth for 77 ethnic groups in 38 regions, the multivariate models show that ethnic groups from warmer climates and from lower wealth countries are given lower evaluation in both competence and warmth stereotypes. However, ethnic groups from more northerly countries are also given a lower evaluation in warmth. Ethnic stereotypes reflect both features of ethnic groups in countries of origin (e.g., the North-South polarization) and group characteristics carried by ethnic groups in new contexts (i.e., intergroup relations). Thus, reactions to ethnic groups seem to differ partly depending on countries of origin mixed in people’s minds with information about geography, climate, and national wealth in the social perception process. Stereotypes associated with ethnic groups across countries to some extent track the stereotypes associated with the ecologies in which these ethnic groups are assumed to predominantly live. This highlights the importance of the establishment or expansion of policies and programs regarding international inequality.
... With regards to immigrant outcomes specifically, (Kirkegaard & de Kuijper, 2020) advocated the following causal model, which we present here in generalized format in Figure 1. Specifically, populations vary in their mean trait levels, including human capital ones. ...
Article
Full-text available
Immigrants to Western countries typically have worse social outcomes than natives, but country of origin immigrant groups differ widely. We studied school performance in Denmark for 116 immigrant groups measured by the grade point average (GPA) of the 9th grade exam at the end of compulsory schooling. General intelligence is a strong causal factor of school outcomes and life outcomes in general for individuals. We accordingly predicted that country of origin average intelligence (national IQ) will predict immigrant group outcomes. We furthermore included as covariates immigrant generation (first vs. second) as well as the Muslim percentage of country of origin. Results show that GPA in Denmark can be predicted by national IQ r = .47 (n = 81), Muslim percentage r = -.40 (n = 81), and educational selectivity of immigrants entering Denmark r = .35 (n = 71). Regression modeling indicated that each predictor is informative when combined. The final model explained 46.3% of the variance with first generation (binary) β = -0.65, βIQ = 0.29, βMuslim = -0.21, and β education selectivity index = 0.27 (all predictors p < .001, n = 97). Our results are in line with existing research on cognitive stratification and immigration.
Article
Full-text available
Socioeconomic position (SEP) is a multi-dimensional construct reflecting (and influencing) multiple socio-cultural, physical, and environmental factors. In a sample of 286,301 participants from UK Biobank, we identify 30 (29 previously unreported) independent-loci associated with income. Using a method to meta-analyze data from genetically-correlated traits, we identify an additional 120 income-associated loci. These loci show clear evidence of functionality, with transcriptional differences identified across multiple cortical tissues, and links to GABAergic and serotonergic neurotransmission. By combining our genome wide association study on income with data from eQTL studies and chromatin interactions, 24 genes are prioritized for follow up, 18 of which were previously associated with intelligence. We identify intelligence as one of the likely causal, partly-heritable phenotypes that might bridge the gap between molecular genetic inheritance and phenotypic consequence in terms of income differences. These results indicate that, in modern era Great Britain, genetic effects contribute towards some of the observed socioeconomic inequalities. Household income is used as a marker of socioeconomic position, a trait that is associated with better physical and mental health. Here, Hill et al. report a genome-wide association study for household income in the UK and explore its relationship with intelligence in post-GWAS analyses including Mendelian randomization.
Article
Full-text available
It is well established that general intelligence varies in the population and is causal for variation in later life outcomes, in particular for social status and education. We linked IQ-test scores from the Danish draft test (Børge Prien Prøven, BPP) to social status for a list of 265 relatively common names in Denmark (85% male). Intelligence at the level of first name was strongly related to social status, r = .64. Ten names in the dataset were non-western, Muslim names. These names averaged an IQ of 81 (range 76-87) compared with 98 for the western, mostly Danish ones. Nonwestern names were also lower in social status, with a mean SES score of 2.66 standard deviations below that of western names. Mediation analysis showed that 30% of this very large gap can be explained by the IQ gap. Reasons for this relatively low level of mediation are discussed.
Article
Full-text available
Research at the individual level shows strong positive relationships between performance in video games and on intelligence tests. Together with evidence of above average IQs of players of traditional mental sports such as chess, this suggests that national IQs should be strongly related to national performance on mental sports. To investigate this, lists of top players for 12 different electronic sports (e-sports) and traditional mental sports were collected from a variety of sources (total n = 36k). Using a log count approach to control for population size, national cognitive ability/IQ was found to be a predictor (p<.05) of the relative representation of countries among the top players for every game except Go. When an overall mental sports score was calculated using a factor analytic approach, the factor scores correlated r = .79 with Lynn and Vanhanen's (2012) published national IQs. The pattern was somewhat nonlinear such that national IQs below 85 seemed to have no relationship. The games that related most strongly with the general factor of mental sport ability also correlated more strongly with national IQs (r = .94). The relationship was fairly robust to controls for geographical region (coefficient 74% of the original in chosen model specification).
Article
Recent research not only confirms the existence of substantial psychological variation around the globe but also highlights the peculiarity of many Western populations. We propose that part of this variation can be traced back to the action and diffusion of the Western Church, the branch of Christianity that evolved into the Roman Catholic Church. Specifically, we propose that the Western Church’s transformation of European kinship, by promoting small, nuclear households, weak family ties, and residential mobility, fostered greater individualism, less conformity, and more impersonal prosociality. By combining data on 24 psychological outcomes with historical measures of both Church exposure and kinship, we find support for these ideas in a comprehensive array of analyses across countries, among European regions, and among individuals from different cultural backgrounds.
Article
The present register-based study investigated the influence of familial factors on the association of IQ with educational and occupational achievement among young men in Denmark. The study population comprised all men with at least one full brother where both the individual and his brothers were born from 1950 and appeared before a draft board in 1968–1984 and 1987–2015 (N = 364,193 individuals). Intelligence was measured by Børge Priens Prøve at age 18. Educational and occupational achievement were measured by grade point average (GPA) in lower secondary school, time to receiving social benefits at ages 18–30, and gross income at age 30. The statistical analyses comprised two distinct statistical analyses of the investigated associations: A conventional cohort analysis and a within-sibship analysis in which the association under investigation was analysed within siblings while keeping familial factors shared by siblings fixed. The results showed that an appreciable part of the associations of IQ with educational and occupational achievement could be attributed to familial factors shared by siblings. However, only the within sibling association between IQ and GPA in lower secondary school clearly differed from the association observed in the cohort analysis after covariates had been taken into account.