Cross-Cultural Research
http://ccr.sagepub.com/content/43/1/3
The online version of this article can be found at:
DOI: 10.1177/1069397108326290
2009 43: 3 originally published online 29 October 2008
Cross-Cultural Research
Michael Minkov
Rates Better Than Socioeconomic Inequality
Risk-Taking Reproductive Competition Explains National Murder
Published by:
http://www.sagepublications.com
On behalf of:
Society for Cross-Cultural Research
can be found at:
Cross-Cultural Research
Additional services and information for
http://ccr.sagepub.com/cgi/alerts
http://ccr.sagepub.com/subscriptions
http://www.sagepub.com/journalsReprints.nav
http://www.sagepub.com/journalsPermissions.nav
http://ccr.sagepub.com/content/43/1/3.refs.html
Risk-Taking Reproductive
Competition Explains
National Murder Rates
Better Than Socioeconomic
Inequality
Michael Minkov
Sofia University
This article presents evidence that national murder rates are not well explained
as a function of socioeconomic inequality; risk-taking reproductive competition
(RTRC) provides a better explanation. RTRC is a single nation-level dimen-
sion, measurable through national road death tolls, adolescent fertility rates,
and Gini coefficients (thus, socioeconomic inequality is just one facet of it).
RTRC explains 50% of the variance in national differences in murder rates—
far more than Gini coefficients alone. RTRC correlates highly with national val-
ues that explicitly reflect full approval of interpersonal competition as well
as perceptions of full life control (being free to act as one pleases). Respondents
from societies that score higher on RTRC tend to perceive their fellow citizens
as more emotional, impulsive, and lacking deliberation. The author dis-
cusses the origins of the national differences in RTRC and he proposes that
high-scoring countries do not have a long history of intensive agriculture.
Keywords:
murder rates; risk-taking; reproductive competition; socioeco-
nomic inequality
N
ational murder rates fluctuate because of various nation-specific
factors. Nevertheless, country rankings from different years are relatively
stable and there emerge some clear geographic trends. A 1998 longitudinal
study for the World Bank (Fajnzylber, Lederman, & Loayza, 1998) concluded
that the highest incidence of intentional homicides was found in “Latin
American countries.” From 1970 to 1974 and from 1985 to 1989, the average
rate in that part of the world was about 8 murders per 100,000 inhabitants.
Cross-Cultural Research
Volume 43 Number 1
February 2009 3-29
© 2009 Sage Publications
10.1177/1069397108326290
http://ccr.sagepub.com
hosted at
http://online.sagepub.com
3
Author's Note: I am very grateful to the anonymous reviewers who made very valuable comments
and helped me considerably improve my paper.
Then, the rate rose to about 12 between 1990 and 1994. According to the same
study, the second highest rate was found in “sub-Saharan Africa.” There, rates
rose from just more than 4 murders per 100,000 inhabitants between 1970 and
1974 to 6 between 1975 and 1979, fell to 2 from 1985 to 1989, and soared to
about 9 between 1990 and 1994. In “Europe and Central Asia,” murder rates
fluctuated between 2 and 4 during the whole period. In “Asia,” they fluctuated
between 2 and 3. The lowest rates were found in “the Middle East and North
Africa”: about 2 throughout the whole period.
The relative stability of these national differences suggests that there exist
durable determinants of homicide rates that consistently produce stronger
effects in some countries than in others. The search for such determinants
has generated various theories, ranging from quite plausible (for the most
recent review, see Barber, 2007) to exotic and highly controversial.
1
The single most preferred explanation in the literature on national differ-
ences in violent crime rates, especially murder, focuses on differences in
socioeconomic inequality (Avison & Loring, 1986; Barber, 2007; Braithwaite
& Braithwaite, 1980; Fajnzylber, Lederman, & Loayza, 2002; Krohn, 1976;
Lim, Bond, & Bond, 2005; Messner, Raffalovich, & Shrok, 2002; Wilkinson,
Kawachi, & Kennedy, 1998; Wilson, Daly, & Pound, 2002). It has been
demonstrated statistically, and beyond any doubt, that all over the world, soci-
eties with greater inequality, such as those of Latin America and parts of sub-
Saharan Africa, have higher homicide rates and other types of violence. The
correlation between Gini coefficients (the most commonly used measure of
socioeconomic inequality) and homicide rates is robust across nations.
The mechanism through which socioeconomic inequality supposedly
generates violent crime is explained in terms of strain theory (Agnew, 1999;
Lim et al., 2005). Strain theory was first proposed by Merton (1938), later
developed by Agnew (1992; 1999), Agnew and White (1992), and other
authors. In a nutshell, this theory postulates that individuals whose aspira-
tions and opportunities are not properly balanced may experience psycho-
logical strain and that will push some of them toward criminal behavior.
Anger and frustration seem to be important elements in this theory. Thus,
from this theoretical perspective, it seems plausible to assume that greater
socioeconomic inequality can generate bitterness and envy on the part of
some members of the less privileged social classes. As a result, some of
those individuals will resort to violent crime, including homicide, in a mis-
guided effort to redress the perceived social injustice.
This theory is not without its detractors. Neumayer (2003) believes
that the positive effect of inequality on homicide rates found in many stud-
ies may be spurious, whereas Butchart and Engstrom (2002) argue that
4
Cross-Cultural Research
redistributing wealth in societies with high economic inequality, without
increasing per capita gross domestic product (GDP), would reduce homicide
rates less than redistributions linked with overall economic development.
Besides, the socioeconomic interpretation of homicide rates does not
consider the possibility of a reverse causation. If criminal violence can be
a result of socioeconomic inequality, it is equally conceivable that it con-
tributes to it. Drug barons and warlords in poor countries amass untold for-
tunes precisely by means of violence and murder. It is these individuals
who are responsible for a high proportion of the homicides in their coun-
tries. The cause-and-effect relationship from violence to affluence and
status is easy to discern in some societies that have not reached statehood,
for instance those of foragers (Wilson et al., 2002).
There are serious problems with the strain-from-socioeconomic-inequality
theory as a universal explanation of group differences in homicide. Some of
the most crucial evidence against it comes from studies of preliterate soci-
eties, characterized by extremely egalitarian cultures. Despite their egalitar-
ianism, some of them have extremely high murder rates. As Wilson and
colleagues (2002) put it, homicide rates in hunter-gatherers’ societies gener-
ally dwarf those of modern nation-states (p. 395). Lee (1979) carried out a
meticulous longitudinal study of intentional homicide among 1,500 !Kung
tribesmen at a time when they had had very few contacts with outsiders and
no social hierarchy or inequality at all. Lee reported an annual rate of 29.3
homicides per 100,000 people (p. 398), which is similar to the high rates of
the northern Latin American countries nowadays (UN Development
Program [UNDP], 2007/2008; UN Office on Drugs and Crime, 2004).
Faurie and Raymond (2004) reported exorbitant murder rates—hundreds per
100,000 people—for several preliterate societies, such as the Yanomamo of
the Amazon basin and some tribes in Papua New Guinea, who are also char-
acterized by very insignificant social hierarchy and inequality.
There is further evidence that compromises the strain-from-socioeco-
nomic-inequality theory. National homicide rates correlate positively with
other variables that cannot be explained in terms of socioeconomic frustra-
tion. The following examples illustrate this point.
Annual national homicide rates per 100,000 inhabitants are available
from the UN Office on Drugs and Crime (2004) and the UNDP (2007/2008).
The data sets in these two sources are for different years in the period
between 2000 and 2004 but they are strongly correlated and load .94 each
on a single factor. Therefore, they can be safely used to calculate an average
murder index for the period from 2000 to 2004 for 120 countries, presented
in Table 1. The country ranking corresponds closely to the longitudinal
Minkov / Risk-Taking Reproductive Competition
5
regional ranking in the World Bank study (Fajnzylber et al., 1998), which
attests to its reliability. The UN Office on Drugs and Crime (2008) indicates
that its data are collected from national authorities and that murder rates
seem to be relatively safe for international comparisons, more so than many
other crimes, because definitions of homicide and reporting patterns are
more or less similar across the world.
Annual national road death tolls per 100,000 inhabitants are available
from the Peden et al. (2004). The data are for the years from 1995 to 2001.
The murder index and the road death tolls correlate at .54 (p
< .000, n = 67).
This correlation is slightly higher than the one between the murder index
and the latest available Gini coefficients (UNDP, 2007/2008): .49 (p
< .000,
n
= 93). Controlling for the Gini coefficients reduces the correlation
between the murder index and the road death tolls to .47 (p
< .000) but does
not make it disappear. Adding other control variables—Kaufmann-Kraay
rule of law index (International Bank for Reconstruction and Development
/ The World Bank, 2007), education index (UNDP, 2006), and GDP per
person at purchasing power parity (PPP) in 1999 (UN Statistics Division,
2007, International Bank for Reconstruction and Development / The World
Bank, 2007)—and controlling for them together with the Gini coefficients,
still leaves a significant correlation of .43 (p
< .001) between the murder
index and road death tolls.
Naturally, a high road death toll cannot cause high intentional homicide
rates. There can be no cause-and-effect relationship between the two vari-
ables. But there must be a common factor behind the two and it is inde-
pendent of national differences in socioeconomic inequality, wealth,
education, or rule of law as conceptualized by the analysts of Western
development banks.
This evidence does not directly compromise the strain theory. One could
argue that even if reckless driving has nothing to do with socioeconomic
factors, it is because of strain, just like murder. But strain theory would lose
much of its appeal if the socioeconomic element were lost. If strain is not
caused by socioeconomic inequality, what is its origin?
There is additional evidence against the strain-from-socioeconomic-
inequality theory. The UN Office on Drugs and Crime (2004) provides non-
intentional homicide rates for 2000. It is obvious that this variable is
different from the road death tolls because the correlation between the two
is weak and insignificant. The nonintentional homicide rates correlate with
the murder index at .69 (p
< .000, n = 40). After controlling for Gini coef-
ficients, GDP per person at PPP, and education, there still remains a partial
6
Cross-Cultural Research
Minkov / Risk-Taking Reproductive Competition
7
Table 1
Murder Index for 120 Countries (from 2000 to 2004)
Rank
Country or Countries
Annual Number of Homicides per 100,000 Population
1
Colombia
62.72
2
Swaziland
51.11
3
Lesotho
50.70
4
South Africa
49.45
5
Jamaica
34.05
6
Venezuela
33.17
7
El Salvador
31.50
8
Guatemala
25.49
9
Russia
19.85
10
Ecuador
18.30
11
Kazakhstan
16.80
12
Bahamas
15.90
13
Guyana
13.80
14-15
Nicaragua, Mongolia
12.80
16
Paraguay
12.33
17
Suriname
10.30
18
Lithuania
9.71
19
Latvia
9.32
20
Belarus
9.22
21-22
Papua New Guinea
9.08
23
Estonia
8.63
24
Thailand
8.49
25
Argentina
8.34
26
Kyrgyzstan
8.20
27
Ukraine
8.16
28
Rwanda
8.00
29
Zambia
7.89
30
Zimbabwe
7.82
31
Mexico
7.71
32
Tajikistan
7.60
33-34
Barbados, Tanzania
7.50
35-36
Uganda, Seychelles
7.40
37
Moldova
7.42
38
Sri Lanka
6.70
39
Namibia
6.30
40
Costa Rica
6.20
41
Oman
6.00
42
Philippines
5.95
43-44
Lebanon, Albania
5.70
45
Peru
5.50
46
Georgia
5.48
47
United States
5.13
(continued)
8
Cross-Cultural Research
Table 1 (continued)
Rank
Country or Countries
Annual Number of Homicides per 100,000 Population
48
Uruguay
5.11
49
Ivory Coast
4.09
50-51
Egypt, Palestine
4.00
52
Turkey
3.80
53
India
3.71
54
Poland
3.61
55
Bulgaria
3.59
56
Nepal
3.40
57
Armenia
2.92
58
Finland
2.83
59
Bolivia
2.80
60
Dominica
2.77
61
Azerbaijan
2.61
62
Israel
2.60
63
Slovakia
2.48
64
Romania
2.45
65
Sweden
2.40
66
Malaysia
2.36
67
Mauritius
2.35
68
Republic of Macedonia
2.31
69
Portugal
2.13
70
South Korea
2.11
71
China
2.10
72
Hungary
2.08
73
Czech Republic
1.95
74
Switzerland
1.93
75
United Kingdom
1.86
76-77
Malta, Croatia
1.80
78
Canada
1.75
79-80
Cyprus, Fiji
1.70
81
France
1.69
82
Slovenia
1.66
83
Chile
1.63
84-85
Nigeria, Belgium
1.50
86
Australia
1.44
87-88
Algeria, Brunei
1.40
89
Iceland
1.39
90
Maldives
1.30
91
Italy
1.25
92
New Zealand
1.24
93
Spain
1.23
94
Tunisia
1.19
95
Syria
1.10
(continued)
correlation of .58 (p
< .000) between intentional and nonintentional homi-
cide rates. Again, this suggests that the two variables have something in
common. And, in the case of nonintentional homicide, it is hard to speak of
unfulfilled goals as the cause of the fatalities.
If the strain-from-socioeconomic-inequality theory is not a good expla-
nation of national murder rates and their correlates, what is? An interesting
perspective is suggested by some evolutionists. From the viewpoint of evo-
lution, violence—and particularly murder—can be viewed as a byproduct
of mating competition (Barber, 2006; Buss & Duntley, 2004; Duntley &
Buss, 2004). These authors admit that evolution can breed adaptive mech-
anisms such as altruism. But evolution is also a competitive process and,
from an evolutionary viewpoint, violence can represent a “fitness contest”
(Duntley & Buss, 2004, p. 106), despite all the potential costs that the vio-
lent party may incur. Greater fitness results in better mating and reproduc-
tion opportunities. Thus, evolution seems to have followed two opposite
paths simultaneously: “cooperative and benefit-bestowing adaptations” and
“adaptations in humans whose proper function is to inflict costs on com-
petitors” (Buss & Duntley, 2004, p. 119). Using the logic of this theory,
Minkov / Risk-Taking Reproductive Competition
9
Table 1 (continued)
Rank
Country or Countries
Annual Number of Homicides per 100,000 Population
96
Germany
1.09
97-98
Netherlands, Indonesia
1.08
99
Bahrain
1.00
100
Singapore
.92
101-103
Jordan, Ireland,
.90
Luxembourg
104-105
Denmark, Norway
.95
106
Austria
.80
107
Greece
.78
108
United Arab Emirates
.60
109
Hong Kong
.58
110
Saudi Arabia
.51
111-114
Botswana, Madagascar,
.50
Morocco, Japan
115
Qatar
.49
116-117
Sudan, Pakistan
.30
118-119
Myanmar, Tonga
.20
120
Kuwait
.10
Sources: UN Development Program (2007/2008); UN Office on Drugs and Crime (2004).
murder can occur for the purpose of the direct elimination of a sexual rival.
Alternatively, it can also be motivated by a desire for power and higher
social and economic status. In turn, power and status will increase a man’s
reproductive opportunities.
There is significant evidence to support the reproductive competition
theory as an explanation of violence and murder. Various anthropologists
(Haviland, 1990; Oswalt, 1986) discuss a link between the polygyny that
characterizes many horticulturalist and hunting-gathering societies (tradi-
tionally typical of sub-Saharan Africa and what is now Latin America) and
aggression: Polygynous men who compete for women tend to fight.
Marlowe (2003) presents evidence that internal warfare (violence within a
society) and assault frequency are positively correlated with polygyny
rates. Barber (2006) has also found evidence that the high violent crime
rates in the Americas nowadays can be attributed to mating competition. He
indicates that compared to other world regions, the Americas have not only
higher murder rates, but also higher rape rates.
Schmitt and the International Sexuality Description Project (2004) car-
ried out a 53-nation study of “mate poaching”: romantically attracting
someone who is already in a relationship. They presented their results after
aggregating them at the regional level, not as national scores. Still, the
aggregated results are interesting. In all regions, men were more likely to
report poaching attempts than women. The highest percentages of men who
reported such attempts were in Latin America, followed by Eastern Europe.
The lowest percentages were in East and Southeast Asia and in the Middle
East. This suggests a positive region-level correlation between mate poach-
ing and the murder index. The correlation does not imply that the main
cause of murder is sexual jealousy. Rather, it suggests cultural differences
in reproductive competition that can take on different behavioral forms,
ranging from mob wars to mate poaching.
Furthermore, the murder index correlates with adolescent fertility rates
(UN Statistics Division, 2008) at .55 Spearman (p
< .000, n = 118).
Adolescent fertility clearly indicates reproductive competition; that is what
reduces the age at which it is socially acceptable for women to have sex and
children.
From the viewpoint of reproductive competition theory, socioeconomic
inequality can also be viewed as a result of a fitness contest. The idea that
richer individuals are somehow fitter may sound socially indigestible. But
it is well known that more affluent men have better mating opportunities in
most societies (Fisher, 1992; White, 1988) and better chances of survival.
Even in strictly egalitarian societies, women tend to prefer men who are
10
Cross-Cultural Research
capable of providing more resources. Thus, although the Hadza share game
meat evenly across households, women prefer better hunters as mates,
probably because hunting success could indicate a man’s health, vigor, and
intelligence (Marlowe, 2004). In a Darwinian sense, men who have the
potential to provide more resources are winners in a survival contest.
Stronger competition inevitably results in a greater distance between the
winners and the losers; hence the greater socioeconomic polarization of
societies where such competition is more prevalent. It appears then that
instead of speaking of a cause-and-effect direction from socioeconomic
equality to high murder rates, it is more logical to accept that the two are
caused by a common factor: strong competition for resources, which ulti-
mately amounts to reproductive competition.
But what is the relationship between murder rates and death tolls? As
Wilson and colleagues (2002) put it, the human male psyche has evolved to
be risk accepting in competitive situations, and more so than the female
psyche. In societies where the competition is fiercer, risk taking will be
more prevalent. High road death tolls indicate aggressive driving and seri-
ous risk taking in a competition for road dominance. This also explains the
significant correlations between road death tolls and the other competition-
related variables: murder, r
= .54 (p < .000, n = 67); Gini, r = .53 (p < .000,
n
= 65); and adolescent fertility, r = .53 (p < .000, n = 65).
The high nation-level correlation between murder rates and involuntary
manslaughter can be explained in the same way. Societies with fiercer com-
petition between males generate more risk-taking behaviors that can
involve recklessness.
HIV rates (UNDP, 2006) tell the same story. They correlate with the
murder index at .39 (p
< .000, n = .98). Nowhere in the world are HIV rates
higher than in sub-Saharan Africa. This peculiar phenomenon has been
studied extensively and the results are clear. Despite the complexity of the
HIV pandemic, its development is attributable, among other things, to risk-
taking sexual networking. Not all politicians and other public figures like
this fact but it has been firmly documented and its denial cannot help con-
tain the pandemic in any way. Caldwell (2000; 2002), who studied exten-
sively the social mechanisms of HIV in the 1990s in association with
African scholars, strongly emphasizes the finding that the sub-Saharan HIV
pandemic is largely a result of heterosexual contacts with multiple parallel
partners. The sexual networking phenomenon in sub-Saharan Africa has
been well studied (Orubuloye, Caldwell, & Caldwell, 1992; 1997) and the
results do not seem to be challenged by serious African scholars. The con-
scious risk taking that this phenomenon involves is also well documented.
Minkov / Risk-Taking Reproductive Competition
11
According to Caldwell (2002), during the 1990s many African interviewees
were knowledgeable about the cause of AIDS but did not consider the
prospect of death a deterrent and did not intend to desist from their sexual
networking practices. Many told the interviewers that they would probably
be dead within 10 years even without AIDS, therefore the disease was not
considered frightful. Moore and Oppong (2006) reached a similar conclu-
sion after interviewing 151 HIV carriers in Togo. They found that despite
the awareness that HIV-positive people may infect their sexual partners, the
former deliberately ignore the risk because other considerations—such as a
desire to have a child—may be more important. The researchers concluded
that condom access is insufficient to change risky sexual behavior. In a
study of attitudes toward HIV and death, Awusabo-Asare, Abane, Badasu,
and Anarfi (1999) reported that some young Ghanaians “think that people
are going to die anyway and it may not matter much what they die from” (p.
125). Summarizing findings from a large-scale research project that tar-
geted 10- to 25-year-olds in Uganda, Kenya, Tanzania, and Ethiopia, car-
ried out from 1995 to 2000, Amuyunzu-Nyamongo et al. (1999) reported
the following reactions to the prospect of contracting HIV/AIDS expressed
by their young interviewees: “AIDS came for the people”; “I am not a tree
to be used for furniture”; and “Everybody will die anyway” (p. 8).
There is another line of research that seeks to link murder rates to cultural
values and societal norms. McAlister (2006) studied the acceptability of
homicide, for instance as an act of revenge, in 19 cultures and found evi-
dence, albeit cautiously expressed, that this variable predicts national homi-
cide rates. Minkov (2007) discusses a cultural dimension, extracted from
World Values Survey (WVS; 2006) data, called “indulgence versus restraint.”
It distinguishes societies whose members report a higher perception of per-
sonal life control (a greater feeling that one can act in accordance with one’s
wishes and indulge one’s desires) versus a perception of lower life control
because of severe social restrictions. Minkov (2007) demonstrates that more
restrictive societies tend to have lower rates of some types of violent crime
than indulgent societies, although the effect of indulgence is not strong.
Hypotheses
This study will attempt to determine whether the risk-taking reproduc-
tive competition theory, as an explanation of murder rates across modern
nations, can find statistical support. This means that several hypotheses
need to be validated.
12
Cross-Cultural Research
Hypothesis 1: Risk-taking reproductive competition (RTRC) is a meaningful
nation-level construct. Nation-level variables that suggest risk taking and—
from an evolutionary viewpoint—mating competition are intercorrelated and
form a strong single factor.
Hypothesis 2: The competition-related nature of the RTRC factor can be con-
firmed through significant nation-level correlations with explicitly stated val-
ues and national character descriptions that indicate various types of
competition–orientation.
Hypothesis 3: The RTRC factor is highly correlated with national murder rates.
It provides a better explanation of those rates than any other variables
that can conceivably affect murder rates, such as GDP per person at PPP, and
education.
Hypothesis 4: As a composite measure, the RTRC factor is a better predictor of
national murder rates than socioeconomic equality alone.
Method
In their chapter on cross-cultural research methodology in Project
GLOBE’s main publication, Hanges and Dickson (2004) outline two
main approaches: theory driven and empirical. Those researchers, like
many others, prefer the former to the latter.
2
This study starts from Duntley
and Buss’s (2004) reproductive competition theory and Wilson and col-
leagues’ (2002) theory about the link between risk taking and competition.
However, if a RTRC factor is found, it is logical to continue with an empir-
ical exploration. The factor should be correlated with some 500 societal
variables from all possible cross-cultural databases and its nomological
network (any high and meaningful correlates) should be reported in an
effort to elucidate the nature of this factor, in addition to what the theory
suggests.
The following is a list of the main variables (other than murder rates)
used in this study and their sources. It contains all variables that were found
to relate statistically and conceptually to murder rates and their main cor-
relates, Variables that yield weak correlations with murder rates have in
principle been left out.
3
Nation-Level Societal Variables
Unless indicated otherwise, the data for these variables are for various
years in the 2000 to 2006 period, usually the latest available. The following
main data sources were used:
Minkov / Risk-Taking Reproductive Competition
13
adolescent fertility rates: UN Statistics Division (2008)
education index: UNDP (2006)
GDP per capita at PPP in 2003: UN Statistics Division, 2007; International Bank
for Reconstruction and Development / The World Bank, 2007
Gini coefficients: UNDP (2007/2008)
HIV rates (percentages of the population carrying the HIV virus): UNDP (2006)
nonintentional homicide rates: UN Office on Drugs and Crime (2004)
rape rates: UN Office on Drugs and Crime (2004)
road death tolls: Peden et al. (2004)
youth rates (percentage of the population aged between 0 and 14): UN Statistics
Division (2007)
Cultural Values, Beliefs, Norms, and Work Goals
Cultural values, beliefs, and norms: the latest wave of the WVS (2006),
a nationally representative study across some 80 countries. Most data in the
latest wave are for the period from 1998 to 2001.
Work goals: Hofstede (2001). The study was done around 1970 in some
50 national subsidiaries of the IBM corporation, involving about 110,000
employees.
Descriptions of National Character (DNC)
Scores for the 30 facets of the Big Five personality traits, based on
respondents’ descriptions of the average person in their own society (a total
number of 49 societies on all continents) are reported in McCrae,
Terracciano, Realo, and Allik (in press). These authors, as well as
Terracciano et al. (2005) state that these DNC do not correlate well with
mean national personality traits from self-reports; consequently, they may
be illusory. Nevertheless, Heine, Buchtel, and Norenzayan (2008) argue
that it is self-reports that have a lower predictive power at the national level,
whereas DNC meaningfully predict external societal variables. Among
other things, the present study will contribute to this debate. If DNC predict
societal variables, they cannot be completely illusory.
Results
The murder index correlates positively with the following societal vari-
ables that suggest risk taking or competition, or both (**p
< .01):
14
Cross-Cultural Research
nonintentional homicide
.69** (n
= 40)
road deaths
.54** (n
= 67)
adolescent fertility
.35**, 55** Spearman (n
= 118)
Gini coefficients
.49** (n
= 93)
rape
.48** (n
= 74)
HIV
.39** (n
= 98)
All of these variables are significantly intercorrelated except road deaths
and nonintentional homicide, rape and adolescent fertility, and rape and
road deaths.
It was already explained why these variables, except rape, reflect risk
taking or reproductive competition, or both. The case of rape is also clear,
but this variable poses methodological problems because of cultural differ-
ences in definition and reporting patterns. This is arguably an unreliable
variable that is best left out of the analysis.
The exorbitant HIV rates in many African countries also create statisti-
cal problems. This variable has an enormous standard deviation and tends
to dominate whatever factor it underpins. Therefore, it was also left out of
the factor matrix.
Nonintentional homicide was also left out because of the relatively small
number of countries for which data are available.
The three remaining variables formed a strong single factor with an
eigenvalue of 2.27, explaining 75.5% of the variance. The factor loadings
after varimax rotation are as follows:
adolescent fertility
.91
Gini coefficients
.89
road deaths
.80
The factor scores, multiplied by 100, are presented in Table 2.
Unfortunately, there are no sub-Saharan countries in Table 2 because none
of them is represented on all three variables. There is no doubt however that
most of sub-Saharan Africa would occupy a high position in that ranking
because the countries of that region have the highest adolescent fertility rates
in the world, very high social inequality, and very high road death tolls.
This validates Hypothesis 1. The RTRC factor is very strong, and it is
indeed underpinned by competition-related variables at least one of which
clearly indicates high risk taking as well. Is there now any additional proof
that the RTRC factor captures aspects of competition?
A number of authors (Hanges & Dickson, 2004; Hofstede, 2001; Hui &
Triandis, 1989) indicate that different nations have different response styles
Minkov / Risk-Taking Reproductive Competition
15
16
Cross-Cultural Research
Table 2
Risk-Taking Reproductive Competition Index for 65 Countries
Rank
Country or Countries
Factor Scores
× 100
1
Dominican Republic
309
2
El Salvador
276
3
Brazil
229
4
Colombia
194
5
Venezuela
180
6
Nicaragua
177
7
Panama
172
8
Ecuador
163
9.
Costa Rica
145
10
Peru
126
11
Chile
105
12
Argentina
81
13
Mexico
79
14
Thailand
73
15
Uruguay
59
16
United States
38
17-18
China, Russia
35
19
Latvia
21
20
Lithuania
3
21
Trinidad and Tobago
1
22
Romania
−10
23
Moldova
−14
24
Estonia
−15
25
Uzbekistan
−16
26
New Zealand
−17
27
Greece
−19
28
Turkmenistan
−23
29-30
Georgia, South Korea
−24
31
Portugal
−28
32
Egypt
−31
33
Kyrgyzstan
−34
34-35
Azerbaijan, Bulgaria
−39
36
Poland
−40
37-38
Spain, Belarus
−43
39
Republic of Macedonia
−48
40
Italy
−49
41-42
Ireland, Belgium
−52
43
Israel
−54
44-45
Australia, United Kingdom
−55
46-47
Armenia, Singapore
−56
48
Ukraine
−58
49
Albania
−61
(continued)
when presented with questionnaire items scored on a Likert-type scale.
Latin Americans tend to choose positive extremes when they rate goals and
values, whereas Asians demonstrate moderation and prefer the middle. For
example, Hofstede’s (2001) IBM questionnaire included 14 seemingly
diverse work goals, such as importance of achievement, pay, free time,
good working conditions, good relationships, and so forth. Despite the
diversity of the goals, Latin Americans, especially those of the northern
parts of Latin America, had the highest scores on almost all items, followed
by West Africans. The lowest scorers were West European nations, East
Asians, and Pakistanis. Eastern Europe was not represented.
Hofstede (2001, appendix 3) calculated a work goal importance index
for each nation in his sample by adding up its scores for all 14 goals. After
multiplying the index by
−1 because it was reversely scored, its correlation
with the RTRC index in Table 1 is .79 (p
< .000, n = 31). The strong corre-
lation suggests that nations with a higher RTRC index have a stronger goal
orientation across a wide range of domains. Hofstede (2001) did not pro-
vide a detailed explanation of the response style phenomenon in his
research but mentioned briefly that extreme response style may be because
of lower education and an inability to distinguish between goals. However,
it is clear from the nationally representative studies of the Pew Research
Center (2007) that Latin Americans do not always exhibit extreme response
style. When they are asked to make judgments about the role of various
Minkov / Risk-Taking Reproductive Competition
17
Table 2 (continued)
Rank
Country or Countries
Factor Scores
× 100
50
France
−62
51
Tajikistan
−64
52
Canada
−66
53
Slovakia
−69
54-55
Croatia, Hungary
−73
56
Slovenia
−74
57
Austria
−80
58
Germany
−95
59
Netherlands
−99
60
Czech Republic
−105
61
Finland
−106
62
Denmark
−111
63
Norway
−112
64
Japan
−125
65
Sweden
−131
institutions in their countries, their answers are as balanced as those of West
Europeans. Thus, the reason that Latin Americans give high ratings to sim-
ply formulated work goals is probably not to be sought in their education
level; it is not because they do not know what they are talking about that
they exhibit extreme response style when they discuss their values. On the
contrary, the competitive nature of Latin American culture makes its
members keenly aware of the importance of most work goals because com-
petition means importance of achievement.
Interestingly, Hofstede’s (2001) work goal importance index correlates
with the murder index at .50 (p
< .002, n = .38).
The RTRC index is significantly correlated with two WVS (2006) items
that unequivocally reflect approval of competition: E035, full agreement
that large income differences are a good thing, r
= .59 (p < .000, n = 52);
and E039, full agreement that interpersonal competition is a good thing,
r
= .55 (p < .000, n = 53). The only other WVS values that correlate highly
with RTRC reflect pride and religiousness. Minkov (2007, 2008) demon-
strates a high ecological correlation between these variables and approval
of competition: At the level of nations, pride goes together with self-stabil-
ity (religiousness), a sense of superiority, and a willingness to demonstrate
superiority through competition.
Most importantly, RTRC correlates highly with WVS (2006) item E173:
percentage of people who state that they are in full control of their own
lives.
4
Across 55 countries, the correlation is .63 (p
< .000), indicating that
countries with a higher RTRC index have higher percentages of people who
believe they can act as they please, without feeling various social restric-
tions. This is a logical finding in the light of the reproductive competition
theory because competition requires a degree of freedom.
These findings validate Hypothesis 2. It is quite clear that the RTRC
index reflects approval of interpersonal competition and a sense of personal
freedom. However, it is worthwhile looking at yet another source of infor-
mation: DNC of the average citizen of particular countries, made by their
fellow countrymen and women.
Across 28 common cases, RTRC correlates significantly (*p
< .05; **
p
< .01), and in excess of +.40, with the following DNC facets reported in
McCrae and colleagues (in press):
fantasy
.62**
feelings
.48**
warmth
.46*
impulsivity
.44*
18
Cross-Cultural Research
positive emotions
.42*
order
−.66**
deliberation
−.55**
dutifulness
−.49**
These DNC facets form two factors; however a single-factor solution is
also possible as all facets except feelings have high loadings (.76 and
above) on the single factor. A nearly identical factor structure emerged from
studies of personality traits in Chinese populations reported by Cheung
et al. (2001). In those studies, impulsivity and fantasy loaded negatively,
whereas order and deliberation loaded positively, on one and the same fac-
tor. Therefore, assertions that the DNC factor discussed in the present study
is a random hit that is not susceptible to replication would be unfounded.
The single DNC factor captures a societal perception of a personality
characterized by a strong, impulsive, and fantasy-prone emotionality at the
expense of order and deliberation. It correlates with RTRC at .62 (p
< .000,
n
= 28) and casts additional light on its nature: Impulsivity and a lack of
deliberation are easily associated with risk taking.
Members of societies that score high on RTRC tend to perceive their
average fellow citizens as more emotional, impulsive, and disorderly than
do members of societies where RTRC is lower. But are these perceptions
real? As McCrae and colleagues (in press) correctly point out, they may be
nothing but stereotypes, especially if they refer to the average citizen. They
could be derived from observations of a certain minority of individuals who
are more likely to get involved in fierce competition and accept risks that
are unacceptable to the majority of the population. Yet, these minorities
may be larger in some societies than in others and the effects of their
actions are more strongly felt.
Of course, it is interesting to know how RTRC relates not only to DNC
but also to self-reports and peer-reports of Big Five traits that have been
aggregated at the national level. So far, three studies have been published
with sufficiently large numbers of countries that report such national scores
(McCrae, 2002; McCrae & Terracciano, 2005, and Schmitt, Allik, McCrae,
& Benet-Martinez, 2007). Unfortunately, these three studies have produced
puzzling results and it is not at all clear what exactly they have measured.
5
Despite this, I ran a correlational analysis and found only two significant
correlations between RTRC and any of the Big Five from the three studies:
−.43 (p = .039, n = 23) with agreeableness in McCrae (2002) and .48
(p
= .019, n = .23) with conscientiousness in the same study. The murder
index does not yield significant correlations with any dimension in any of
the three studies.
Minkov / Risk-Taking Reproductive Competition
19
Although it is conceptually acceptable that cultures that allow fierce
competition might have less agreeable people, it is not at all clear what
stronger competition should have to do with higher conscientiousness. The
national Big Five scores from self-reports and peer-reports cannot be used
to elucidate the nature of RTRC. Contrary to the assertion made by McCrae
and colleagues (in press), it has not been demonstrated that self-reports and
peer-reports yield a more accurate image of national personality traits than
descriptions of average citizens, despite the indisputable potential for
stereotyping and crude overgeneralization when the latter approach is used.
As the nature of the RTRC factor has been elucidated in this way, that
dimension can be compared with the murder index. The correlation
between the two variables is .70 (p
< .000, n = 59). This is a strikingly high
correlation across a fairly large number of countries. It is higher than any
correlation between murder rates and any other item in a long list of soci-
etal and psychological variables reported in Lim et al. (2005)—probably
the most extensive recent comparison of national murder rates with other
variables.
The murder index correlates significantly also with GDP per person
(r
= −.42 Spearman, p < .000, n = 110). However, its correlation with the
UNDP’s (2006) education index is
−.01 (n = 115). All other societal vari-
ables that might hypothetically predict murder rates, listed in Lim et al.
(2005), yield weak zero-order correlations with that variable and are not
good candidates for a regression analysis. This applies also to youth rates
(the available proxy variable is the percentage of the population aged from
0 to 14; UN Statistics Division, 2007). Thus, contrary to Deane’s (1987)
recommendation, this variable does not need to be considered as a potential
predictor of murder rates. In fact, including a large number of independent
variables in a regression model, despite their weak zero-order correlations
with the dependent variable, might raise the R
2
value, and thus improve the
model. However, as Lim and colleagues (2005) correctly point out, a larger
number of variables usually means a smaller sample of observations
(because of missing values) and that renders the findings less reliable.
This leaves only one variable—GDP per person—as a potential reliable
predictor of the murder rates in addition to RTRC.
Across 59 countries, RTRC and GDP explain .49% of the variance in
murder rates (R
2
= .486, F change = 26.451, df = 2,56, p = .000). GDP has
a beta of
−.060 (t = −.535, p < .595), which rules it out as a predictor of
murder rates. RTRC has a beta of .664 (t
= 5,953, p < .000). It is the only
reliable predictor.
20
Cross-Cultural Research
Admittedly, this parsimonious regression model is not the only possible
one. Other models, with more variables (but probably fewer countries), can
hypothetically be built, in which GDP per person could reach statistical sig-
nificance as a predictor of murder rates in addition to RTRC. It is also pos-
sible that some other measurement of educational achievement can be found
that yields a higher correlation with murder rates than the UNDP (2006)
index does. But RTRC is unlikely to be ruled out as a good predictor.
This validates Hypothesis 3. There remains one more question: Does the
composite RTRC measurement predict murder rates better than socioeco-
nomic inequality alone? The two variables cannot be compared in a regres-
sion model because the result would be high collinearity: Gini is highly
correlated with RTRC. However, the three facets of RTRC—road deaths,
adolescent fertility, and Gini—can be compared in multiple regression to
determine their relative contribution as predictors of the murder index. This
regression model has an R
2
of .502 across 59 countries (F change
= 18.474,
df
= 3,55, p = .000). Road deaths are the best predictor (beta = .395, t = 3.562,
p
< .001), followed by adolescent fertility rates (beta = .256, t = 1.799,
although p
< .077). Gini is the weakest predictor, far from statistical signifi-
cance (beta
= .195, t = 1.366, p < .177). Collinearity is not a significant prob-
lem in this model. The variance inflation factor (VIF) value for road deaths is
1.35, whereas the other two variables have VIF values of 2.24 and 2.25.
Perhaps even more convincingly, a regression model can be built in
which Gini and GDP per person predict the murder index. The R
2
value in
this model is only .21. This is far lower than the .49 of the model in which
the predictors are RTRC and GDP per person.
Gini coefficients make a very weak contribution as a predictor of the
murder index. Road deaths and adolescent fertility—or more precisely the
factor that they define without Gini—can be a sufficient predictor, without
a contribution from Gini. That factor explains about half of the variance in
murder rates. This validates Hypothesis 4.
Discussion
This study presented statistical evidence that reproductive competition
and the risk taking that it involves represent a single measurable phenome-
non at the national level. The measurement is statistically strong and theo-
retically sound. The resulting robust dimension—RTRC—explains up to
one half of the variance in national murder rates, depending on the nature
and size of the country sample. It appears that its predictive power is higher
(closer to 50%) for larger and geographically well-balanced samples.
Minkov / Risk-Taking Reproductive Competition
21
RTRC clearly reflects national differences in endorsement of interper-
sonal competition because it is highly correlated with explicitly stated com-
petition-related values and life perceptions in the nationally representative
WVS (2006). RTRC also predicts a higher work goal orientation, which
further confirms its competition-related nature.
It is crucially important to note that RTRC and its correlates probably do
not capture trends that are typical or average in any culture. In an absolute
sense, the percentages of people who die in road accidents and the percent-
age of girls who have children at a very young age are low in all countries,
as is the percentage of the population who have ever committed a murder
or killed somebody in a car crash.
Similarly, Caldwell (2002) and the African social scientists who worked
with him found that the average number of sexual partners that their West
African respondents had had in their lives was comparable to that in
American, French, and British populations. However, the researchers found
some unusual extremes: About half of the men and one quarter of the
women in the Nigerian state of Ekiti identified lifetime sexual relationships
with ten or more persons.
It must also be noted that RTRC is highly correlated with the national
percentages of people who perceive full personal life control and fully
approve of competition and large income differences, but the same index
yields low and insignificant correlations with average national perceptions
of personal life control or average national approval of competition and
large income differences. Strong competition is driven by strongly compet-
itive individuals who are not necessarily typical representatives of their
respective nations.
However, despite the relative marginality of strong risk-taking competi-
tion in terms of the percentages of people who are involved in it or approve
of it, its effect can be felt by large segments of a country’s population. The
Pew Research Center’s (2007) most recent cross-national study of 47 nations
on all continents, using mostly nationally representative samples, provides a
good illustration. The respondents were asked whether there is an area
within a kilometer of their homes where they are afraid to walk after dark.
The percentages of respondents who chose the affirmative answer option
range from 8 in Jordan to 84 in Venezuela. These percentages correlate with
the murder index at .56 (p
< .000, n = 39) and with RTRC at .77 (p < .000,
n
= 25). RTRC may be a marginal phenomenon but its effect is not.
It was demonstrated that members of nations that score high on RTRC
tend to describe their fellow citizens as impulsive, emotional, disorderly,
and lacking deliberation. Again, these descriptions should not be taken at
22
Cross-Cultural Research
face value in the sense that they are unlikely to be correct portrayals of the
average member of any nation, despite the fact that the survey respondents
were instructed to depict precisely this average profile. Ordinary respon-
dents are not personality psychologists and their descriptions of national
character are likely to be stereotypical. However, stereotypes are usually
overblown generalizations of some reality. In this case, the kernel of truth
in these descriptions of national character is to be sought in the most
fiercely competitive individuals who accept unusual risks by committing
murder, driving very aggressively, getting involved in unprotected sex, rap-
ing women, and so forth. Fierce competition and risk taking can result in
recklessness on the part of the fiercest competitors. Although they are a
minority in any population, their recklessness creates such a strong impres-
sion on their fellow citizens that it takes center stage in their memories and
causes stereotypical overgeneralizations.
Although the findings of this study seriously compromise the theory that
explains national murder rates as a direct result of socioeconomic inequal-
ity, it does not completely invalidate the strain theory; it is its strain-from-
socioeconomic-inequality version that needs a revision. Instead, it is more
plausible to speak of strain from competition. This type of strain may be
responsible for the reckless risk taking that accompanies the fiercest types
of fitness contests.
As in the case of other cultural dimensions, the historical origins of the
differences in RTRC are shrouded in uncertainty, if not mystery, and any
explanations are bound to be speculative. The first evident observation is the
correlation between RTRC and GDP per person at PPP,
−.50 (p < .000,
n
= 64); as well as between RTRC and the UNDP (2006) education index,
−.64 (p = .000, n = 63). But, in a regression model with 42 common cases,
these two variables explain only 41% of the variance in RTRC. Actually, the
only predictor is the education index, whereas GDP per person has a beta of
−.01 and a p value of .487. It must also be noted that although differences in
GDP per person and education explain about 60% of the variance in adoles-
cent fertility, they explain only 18% of the variance in Gini and 16% of the
variance in road deaths. Thus, only one of the facets of RTRC is satisfacto-
rily explained in terms of wealth and education differences.
Nevertheless, the rich countries are concentrated in the lower part of the
RTRC ranking. This suggests that, after all, an abundance of material
resources could depress RTRC, although it may not create full convergence
among the rich countries. But there is another even more interesting ques-
tion: What accounts for the dispersal of the poor countries throughout the
whole RTRC ranking?
Minkov / Risk-Taking Reproductive Competition
23
Keeping in mind the fact that although sub-Saharan Africa is not repre-
sented in Table 2, it is certainly characterized by very high RTRC, there
emerges a somewhat clear geographic pattern: Latin America and sub-
Saharan Africa versus the rest of the poor world. The main difference
between these two parts of the globe is that the former does not have a long
history of intensive agriculture. The predominant traditional economy there
used to be horticulture and hunting-gathering. There is a well-known link
between horticulture and competition for women, which results in polyg-
yny and aggression (Haviland, 1990; Marlowe, 2003, Oswalt, 1986).
Intensive agriculture appears to have suppressed these phenomena. This
can be explained in several ways. Ember and Ember (1992) indicate that a
number of studies show that intensive agriculturalists typically work longer
hours than horticulturalists. When men work longer, they have less time to
compete for women. On the other hand, in horticulturist societies it is eas-
ier for women to raise children without a very significant contribution from
men (Fisher, 1992), therefore such societies are characterized by lower
paternal provisioning and lower male contribution to subsistence (Marlowe,
2000, 2003). Intensive agriculture involves activities such as plowing and
digging of irrigation systems, which are harder for women than hoeing. As
a result, intensive agriculturalists have to concentrate on their work in the
fields and cannot afford to devote significant efforts to competition for
women. Without intensive paternal provisioning their offspring would die.
There is also another explanation. In many cases, intensive agriculture is
a collective activity. Men who cultivate crops together should learn (and do
learn) how to cooperate rather than compete.
Because adolescent fertility rates are highly correlated with differences
in education, this facet of reproductive competition could hypothetically be
manipulated to some extent by national governments (but see Minkov,
2008, for evidence that educational achievement is highly correlated with
cultural values and is not at all easily manipulated). The other facets may
be more stable and resistant to pressure for change. As a result, although a
global increase of wealth may depress RTRC and homicide rates across
many countries, that does not mean that national differences on that dimen-
sion will disappear altogether or change dramatically in the near future.
A reviewer of this article called its findings “socially dangerous” and
perceived an element of indictment in it. However, various evolutionists
(Duntley & Buss, 2004; Wilson et al., 2002) consider reproductive compe-
tition and the violence that it involves a normal evolutionary adaptation.
There is nothing abnormal in the fact that some environments are more con-
ducive to this adaptation. There is, for instance, evidence that one of the
24
Cross-Cultural Research
possible outcomes of mating competition, polygyny, may act as pathogen
defense in both animals (Read, 1991) and humans (Barber, 2008). From
this perspective, mating competition is a meaningful adaptation in environ-
ments with a lot of infectious diseases.
In fact, disapproval of risk-taking reproductive competition strategies, or
of studies that reveal the existence of cross-cultural differences in this
respect, reveals ethnocentricity. I fully endorse Wilson and colleagues’
(2002) view in this respect. They indicate that violence has been interpreted
in terms of short time-horizons, myopia, and low intelligence. But these
judgmental views disregard important facts: Short time-horizons involving
risk-taking competition are logical adaptations to particular environments
where “the expected fitness returns of present striving are positively accel-
erated rather than exhibiting diminishing marginal returns” (p. 392).
Notes
1. An example of an exotic theory is provided by Mawson and Jacobs (1978). According
to those authors, the high homicide rates in Latin America may be because of an excessive
consumption of corn. The explanation is that corn reduces brain tryptophan and/or serotonin
and that can cause mood disturbances. At the time of their study, the authors could not have
known that Latin Americans would consistently rank at the top of all large-scale cross-cultural
studies of subjective well-being and have very low suicide rates, which is hard to square with
the hypothesis that they may suffer from a serotonin deficiency. An example of an unconvinc-
ing and highly controversial theory is Rushton and Whitney’s (2002) attempt to explain
national homicide rates in terms of racial differences. Rushton and Whitney consistently claim
that Blacks have the highest homicide rates across the world but are silent about the very high
rates in northern Latin America.
2. Because the empirical approach is an exploration that does not necessarily start from an
existing theory, it is sometimes called dustbowl empiricism or a fishing trip. This reveals a sus-
picion that the findings may be accidental: If the net had been cast elsewhere, some other dis-
covery would have been made. However, this suspicion is unfounded when the research is not
random, but exhaustive; in other words, it involves scouring all variables from all public data-
bases that include anything that can conceptually be linked to nation-level cultural differences.
3. An exception was made for a few variables, such as youth rates, that have been sug-
gested by some researchers as potential predictors of murder rates.
4. Life control is not the same as success control. The latter was measured twice by the
Pew Research Center, most recently in 2007 (Pew Research Center, 2007). The participants
were asked how much they agreed with the statement that success depends mostly on external
forces. Mean levels of life control and mean agreement with external success control are neg-
atively correlated but the correlation is modest. Also, the percentages of people who perceive
full life control are insignificantly correlated with the percentages of people who fully agree
or disagree with external success control. Obviously, the life control item in the WVS (2006)
refers to a perceived personal freedom to act in accordance with one’s desires even if that does
not necessarily result in any achievement.
Minkov / Risk-Taking Reproductive Competition
25
26
Cross-Cultural Research
5. Here are some correlations (*p
< .05; **p < .01) between the nation-level Big Five
scores from the three studies, numbered in their chronological order: Agreeableness (1)
×
Agreeableness (2)
= .12 (n = 27); Extraversion (1) × Extraversion (3) = .42* (n = 28); but
Extraversion (1)
× Openness (3) = .72* (n = 28); Conscientiousness (1) × Conscientiousness
(2)
= .34 (n = 27); but Conscientiousness (1) × Neuroticism (3) = −.66** (n = 28).
References
Agnew, R. (1992). Foundation for a general strain theory. Criminology, 30(1), 47-87.
Agnew, R. (1999). A general strain theory of community differences in crime rates. Journal of
Research in Crime and Delinquency, 36, 123-155.
Agnew, R., & White, H. (1992). An empirical test of general strain theory. Criminology 30(4),
475-499.
Amuyunzu-Nyamongo, M., Tendo-Wambua, L., Babishangire, B., Nyagero, J., Yitbarak, N.,
Matasha, M., et al. (1999). Barriers to behavior change as a response to STD including
HIV/AIDS: The East African experience. In J. C. Caldwell, P. Caldwell, J. Anarfi, K.
Awusabo-Asare, J. Ntozi, I. O. Orubuloye, et al. (Eds.), Resistances to behavioral change
to reduce HIV/AIDS infection predominantly heterosexual epidemics in Third World coun-
tries (pp. 1-11). Canberra: Australian National University.
Avison, W. R., & Loring, P. L. (1986). Population diversity and cross-national homicide: The
effects of inequality and heterogeneity. Criminology, 24(4), 733-749.
Awusabo-Asare, K., Abane, A. M., Badasu, D. M., & Anarfi, J. (1999). “All die be die”:
Obstacles to change in the face of HIV infection in Ghana. In J. C. Caldwell, P. Caldwell,
J. Anarfi, K. Awusabo-Asare, J. Ntozi, I. O. Orubuloye, et al. (Eds.), Resistances to behav-
ioral change to reduce HIV/AIDS infection predominantly heterosexual epidemics in Third
World countries (pp. 125-132). Canberra: Australian National University.
Barber, N. (2006). Why is violent crime so common in the Americas? Aggressive Behavior,
32(5), 442-450.
Barber, N. (2007). Evolutionary explanations for societal differences and historical change in
violent crime and single parenthood. Cross-Cultural Research, 41, 123-148.
Barber, N. (2008). Explaining cross-national differences in polygyny intensity: Resource-
defense, sex ratio, and infectious diseases. Cross-Cultural Research, 42, 103-117.
Braithwaite, J., & Braithwaite, V. (1980). The effect of income inequality and social democ-
racy on homicide: A cross-national comparison. British Journal of Criminology, 20, 45-53.
Buss, D. M., & Duntley, J. D. (2004). Homicide: An evolutionary psychological perspective
and implications for public policy. In R. W. Bloom & N. Dess (Eds.), Evolutionary psy-
chology and violence (pp. 115-128). Westport, CT, and London: Praeger.
Butchart, A., & Engstrom, K. (2002). Sex- and age-specific relations between economic devel-
opment, economic inequality, and homicide rates in people aged 0-24 years: A cross-sec-
tional analysis. Bulletin of the World Health Organization, 80(10), 797-805.
Caldwell, J. C. (2000). Rethinking the African AIDS epidemic. Population and Development
Review, 26, 117-135.
Caldwell, J. C. (2002). The African AIDS epidemic: Reflections on a research program—
Perspective. Journal of Population Research, 19, 173-190.
Cheung, F. M., Leung, K., Zhang J. X., Sun, H. F., Gan, Y. Q., Song, W, Z., et al. (2001).
Indigenous Chinese personality constructs: Is the five-factor model complete? Journal of
Cross-Cultural Psychology, 32, 407-433.
Minkov / Risk-Taking Reproductive Competition
27
Deane, G. D. (1987). Cross-national comparison of homicide: Age/sex-adjusted rates using the
1980 U.S. homicide experience as a standard. Journal of Quantitative Criminology, 3, 215-227.
Duntley, J. D., & Buss, D. M. (2004). The evolution of evil. In A. G. Miller, (Ed.), The social
psychology of good and evil (pp. 102-123). New York: Guilford.
Ember, C., & Ember, M. (1992). Anthropology. Englewood Cliffs, NJ: Prentice Hall.
Fajnzylber, P., Lederman, D., & Loayza, N. (1998). Determinants of crime rates in Latin
America and the world: An empirical assessment. Washington, DC: The International
Bank for Reconstruction and Development / The World Bank.
Fajnzylber, P., Lederman, D., & Loayza, N. (2002). Inequality and violent crime. Journal of
Law and Economics, 45, 1-40.
Faurie, C., & Raymond, M. (2004). Handedness, homicide and negative frequency-dependent
selection. Proceedings of the Royal Society B: Biological Sciences, 272 (1558), 25-28.
Fisher, H. (1992). Anatomy of love. London: Simon & Schuster.
Hanges, P. J., & Dickson, M. (2004). The development and validation of the GLOBE culture
and leadership scales. In R. J. House, P. J. Hanges, J. Mansour, P. W. Dorfman, & V. Gupta.
(Eds.), Culture, leadership, and organizations: The GLOBE study of 62 societies.
Thousand Oaks, CA: Sage.
Haviland, W. A. (1990). Cultural anthropology. Orlando, FL: Holt, Rinehart & Winston.
Heine, S., Buchtel, E., & Norenzayan, A. (2008). What do cross-national comparisons of per-
sonality traits tell us? The case of conscientiousness. Psychological Science, 19, 309-313.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, and
organizations across nations (2nd ed.). Thousand Oaks, CA: Sage.
Hui, C. H., & Triandis, H. C. (1989). Effects of culture and response format on extreme
response style. Journal of Cross-Cultural Psychology, 20(3), 296-309.
International Bank for Reconstruction and Development / The World Bank. (2007). Global
monitoring report 2007. Retrieved January 25, 2008, from www.worldbank.org/gmr2007
Krohn, M. D. (1976). Inequality, unemployment and crime: A cross-national analysis.
Sociological Quarterly, 17(3), 303-313.
Lee, R. B. (1979). The !Kung San. Cambridge, UK: Cambridge University Press.
Lim, F., Bond, M. H., & Bond, M., K. (2005). Linking societal and psychological factors to
homicide rates across nations. Journal of Cross-Cultural Psychology, 36, 515-536.
Marlowe, F. W. (2000). Paternal investment and the human mating system. Behavioral
Processes, 51, 45-61.
Marlowe, F. W. (2003). The mating system of foragers in the standard cross-cultural sample.
Cross-Cultural Research, 37, 282-306.
Marlowe, F. W. (2004). Mate preferences among Hadza hunter-gatherers. Human Nature,
15(4), 365-376.
Mawson, A. R., & Jacobs, K. W. (1978). Corn consumption, tryptophan and cross-national
homicide rates. Orthomolecular Psychiatry, 7(4), 227-230.
McAlister, A. F. (2006). Acceptance of killing and homicide rates in nineteen nations.
European Journal of Public Health, 16(3), 259-265.
McCrae, R. R. (2002). NEO-PI-R data from 36 cultures: Further intercultural comparisons. In
R. R. McCrae & J. Allik. (Eds.), The five-factor model of personality across cultures
(pp. 105-126). New York: Kluwer Academic/Plenum.
McCrae, R. R., & Terracciano, A. (2005). Personality profiles of cultures: Aggregate person-
ality traits. Journal of Personality and Social Psychology, 89(3), 407-425.
McCrae, R. R., Terraciano, A., Realo, A., & Allik, A. (in press). Climatic warmth and national
wealth: Some culture-level determinants of national character stereotypes. European
Journal of Personality.
28
Cross-Cultural Research
Merton, R. K. (1938). Social structure and anomie. American Sociological Review, 3, 672-682.
Messner, S. F., Raffalovich, L. E., & Shrok, P. (2002). Reassessing the cross-national rela-
tionship between income inequality and homicide rates: Implications of data quality con-
trol in the measurement of income distribution. Journal of Quantitative Criminology,
18(4), 377-395.
Minkov, M. (2007). What makes us different and similar: A new interpretation of the World
Values Survey and other cross-cultural data. Sofia, Bulgaria: Klasika i Stil.
Minkov, M. (2008). Self-enhancement and self-stability predict school achievement at the
national level. Cross-Cultural Research, 42(2), 172-196.
Moore, A. R., & Oppong, J. (2006). Sexual risk behavior among people living with HIV/AIDS
in Togo. Social Science and Medicine, 64(5), 1057-1066.
Neumayer, E. (2003). Good policy can lower violent crime: Evidence from a cross-national
panel of homicide rates, 1980-97. Journal of Peace Research, 40(6), 619-640. Retrieved
March 27, 2008, from http://eprints.lse.ac.uk/618/1/JPR_(Homicide).pdf
Orubuloye, I. O., Caldwell, J. C., & Caldwell, P. (1992). Diffusion and focus in sexual
networking: Identifying partners and partners’ partners. Studies in Family Planning, 23(6
Pt 1), 343-351.
Orubuloye, I. O., Caldwell J. C., & Caldwell, P. (1997). Perceived male sexual needs and male
sexual behavior in southwest Nigeria. Social Science and Medicine, 44(8), 1195-1207.
Oswalt, W. H. (1986). Life cycles and lifeways. Palo Alto, CA: Mayfield.
Peden, M. et al. (Eds.) (2004). World report on road traffic injury prevention. Geneva: World Health
Organization. (Corrigenda retrieved October 11, 2008, from http://www.who.int/violence_
injury_prevention/publications/road_traffic/world_report/wrrtip_corrigenda_en.pdf
Pew Research Center. (2007). Global opinion trends 2002-2007. Retrieved February 20, 2008,
from http://pewglobal.org/reports/pdf/257.pdf and http://pewglobal.org/reports/pdf/257
topline-trend.pdf
Read, A. F. (1991). Passerine polygyny: A role for parasites? American Naturalist, 138(2),
434-459.
Rushton, J. P., & Whitney, G. (2002). Cross-national variation in violent crime rates: Race,
r-K theory, and income. Population and Environment, 23(6), 501-511.
Schmitt, D. P.; International Sexuality Description Project (2004). Patterns and universals of
mate poaching across 53 nations: The effect of sex, culture, and personality on romanti-
cally attracting another person’s partner. Journal of Personality and Social Psychology,
86(4), 560-584.
Schmitt, D. P., Allik, J., McCrae, R. R., & Benet-Martinez, V. (2007). The geographic distrib-
ution of Big Five personality traits: Patterns and profiles of human self-description across
56 nations. Journal of Cross-Cultural Psychology, 38, 173-212.
Terracciano, A., Abdel-Khalek, A. M., Adam, N., Adamovova, L., Ahn, C.-k., Ahn, H.-n.,
et al. (2005). National character does not reflect mean personality trait levels in 49 cul-
tures. Science, 310(5745), 96-100.
UN Development Program. (2006). Human development report 2006. Retrieved January 20,
2007, from http://hdr.undp.org/hdr2006/statistics/indicators/
UN Development Program. (2007/2008). Human development reports. Retrieved March, 14,
2008, from http://hdrstats.undp.org/indicators/147.html
UN Office on Drugs and Crime. (2004). Seventh United Nations survey of crime trends and
operations of criminal justice systems, covering the period 1998-2000. Retrieved January
20, 2006, from http://www.unodc.org/pdf/crime/seventh_survey/7sv.pdf
Minkov / Risk-Taking Reproductive Competition
29
UN Office on Drugs and Crime. (2008). Compiling and comparing international crime:
Statistics. Retrieved August 22, 2008, from http://www.unodc.org/unodc/en/data-and-
analysis/Compiling-and-comparing-International-Crime-Statistics.html
UN Statistics Division. (2007). Statistical databases; Common database. Internet publication.
Retrieved Oct. 28, 2007, from http://unstats.un.org/unsd/databases.htm
UN Statistics Division. (2008). UN Statistical databases; Social indicators. Indicators on
child-bearing. Adolescent fertility rate. Internet publication. Retrieved Oct. 12, 2008, from
http://unstats.un.org/unsd/demographic/products/socind/childbearing.htm
White, D. R. (1988). Rethinking polygyny: Co-wives, codes and cultural systems. Current
Anthropology, 29(4), 529-572.
Wilkinson, R. G., Kawachi, I., & Kennedy, B. P. (1998). Mortality, the social environment,
crime and violence. Sociology of Health and Illness, 20(5), 578-597.
Wilson, M., Daly, M., & Pound, N. (2002). An evolutionary psychological perspective on the
modulation of competitive confrontation and risk taking. In D. W. Pfaff, A. P. Arnold,
A. M. Etgen, S. E. Fahrbach, & R. T. Rubin (Eds.), Hormones, brain and behavior (Vol.
5, pp. 381-408). San Diego: Academic Press.
World Values Survey. (2006). On-line data analysis. Retrieved in June through November
2006 from www.worldvaluessurvey.com
Michael Minkov is an Associate Professor of Cross-Cultural Awareness at International
University College, Sofia, Bulgaria, and Sofia University Saint Kliment Ohridski, where he is
also doing a PhD in the field of Scandinavian culture. His book "What Makes Us Different; A
New Interpretation of the World Values Survey and Other Cross-Cultural Data" discusses
some new cultural dimensions that Geert Hofstede has accepted as a potential expansion of his
well-known five-dimensional model.
Wyszukiwarka
Podobne podstrony:
Risk Taking
RISK TAKING AND REASONS FOR LIVING IN
Risk taking, global diversification, and growth
Distributed Denial of Service Attacks Explain nation, classification and suggested Solutions
Carolyn Levine Topol A Risk Worth Taking
Reprodukcja ludności Polska wyklad 6 cz 1
REPRODUKCJA POZYCJI SPOŁECZNEJ 7
Reprodukcja ludno ci Polska wyklad 6 cz[1][1] 2
reprod w3 2008
Nationalism and Nihonjinron, Harumi Befu20090112 00001
competence vs performance
Current Clinical Strategies, Psychiatry History Taking (2004) BM OCR 7 0 2 5
Dealing with competency?sed questions
UNCERTAINTY RISK STABILITY?nking 04
Risk score
Bourdieu i Passeron opracowania reprodukcja
więcej podobnych podstron