Content uploaded by Norman P Li
Author content
All content in this area was uploaded by Norman P Li on Dec 16, 2020
Content may be subject to copyright.
British Journal of Psychology (2016), 107, 675–697
©2016 The British Psychological Society
www.wileyonlinelibrary.com
Country roads, take me home... to my friends:
How intelligence, population density, and
friendship affect modern happiness
Norman P. Li
1
and Satoshi Kanazawa
2
*
1
School of Social Sciences, Singapore Management University, Singapore
2
Managerial Economics and Strategy Group, Department of Management, London
School of Economics and Political Science, UK
We propose the savanna theory of happiness, which suggests that it is not only the
current consequences of a given situation but also its ancestral consequences that
affect individuals’ life satisfaction and explains why such influences of ancestral
consequences might interact with intelligence. We choose two varied factors that
characterize basic differences between ancestral and modern life –population density
and frequency of socialization with friends –as empirical test cases. As predicted by
the theory, population density is negatively, and frequency of socialization with friends
is positively, associated with life satisfaction. More importantly, the main associations
of life satisfaction with population density and socialization with friends significantly
interact with intelligence, and, in the latter case, the main association is
reversed among the extremely intelligent. More intelligent individuals experience
lower life satisfaction with more frequent socialization with friends. This study
highlights the utility of incorporating evolutionary perspectives in the study of
subjective well-being.
Positive psychology and evolutionary psychology are two subfields of psychology that
have made significant advances in the last few decades (Cosmides & Tooby, 2013;
Diener, 2012). While several evolutionary psychologists have written on happiness
(Buss, 2000; Hill & Major, 2013; Lewis, Al-Shawaf, Russell, & Buss, 2015; Nesse, 2004),
with only a couple of exceptions (Diener, Kanazawa, Suh, & Oishi, 2015; Heintzelman &
King, 2014), positive psychologists have not drawn on insights from evolutionary
psychology. At the same time, while positive psychologists have accumulated an
impressive amount of empirical knowledge in the last few decades about who is happier
than whom, when, and how, there are few systematic general theories of happiness –
evolutionary or otherwise –that can explain why some individuals are happier than
others. In this study, we propose an evolutionary psychological theory of subjective
well-being that we call the savanna theory of happiness, and provide empirical support
for the theory.
*Correspondence should be addressed to Satoshi Kanazawa, Managerial Economics and Strategy Group, Department of
Management, London School of Economics and Political Science, Houghton Street, London WC2A 2AE, UK
(email: S.Kanazawa@lse.ac.uk).
DOI:10.1111/bjop.12181
675
The savanna theory of happiness
One of the fundamental observations in evolutionary psychology is that, just like any other
organ of any other species, the human brain is designed for and adapted to the conditions
of the ancestral environment, not necessarily the current environment, and is therefore
predisposed to perceive and respond to the current environment as if it were the ancestral
environment (Tooby & Cosmides, 1990). Known variously as the Savanna Principle
(Kanazawa, 2004b), the evolutionary legacy hypothesis (Burnham & Johnson, 2005), or
the mismatch hypothesis (Hagen & Hammerstein, 2006), this observation suggests that
the human brain may have difficulty comprehending and dealing with entities and
situations that did not exist in the ancestral environment, roughly the African savanna
during the Pleistocene Epoch.
The Savanna Principle can explain why some otherwise elegant scientific theories of
human behaviour, such as game theory, often fail empirically, because they posit entities
and situations that did not exist in the ancestral environment. For example, nearly half the
players of one-shot Prisoner’s Dilemma games make the theoretically irrational choice to
cooperate with their partner (Sally, 1995). The Savanna Principle suggests that this may
possibly be because the human brainhas difficulty comprehending completely anonymous
social exchange and absolutely no possibility of knowing future interactions (which
together make the game truly one-shot and defection the only rational choice; Hagen &
Hammerstein, 2006; Kanazawa, 2001). Neither of these situations existed in the ancestral
environment, where all social exchanges were in person and potentially repeated;
however, they are crucial for the game-theoretical prediction of universal defection.
Further, recent developments in evolutionary psychology indicate that general
intelligence may have evolved to solve evolutionarily novel problems (Kanazawa, 2004a,
2010). Psychological mechanisms evolved to solve adaptive problems that recurrently
presented themselves in different domains of life throughout human evolutionary history,
such as social exchange, infant care, and incest avoidance (Tooby & Cosmides, 1990).
They are domain-specific and operate only within narrow domains of life, taking as input
very specific types of information.
Recent theoretical developments suggest that general intelligence, far from being
domain-general, may also have evolved as such a domain-specific evolved psychological
mechanism. It may have evolved to allow individuals to solve a wide variety of non-
recurrent adaptive challenges that also directly or indirectly affected survival or
reproduction. All such non-recurrent adaptive problems were evolutionarily novel.
General intelligence may thus have evolved to solve evolutionarily novel problems, as a
psychological adaptation for the domain of evolutionary novelty.
This suggests that the evolutionary constraints on the human brain proposed by the
Savanna Principle may be stronger among less intelligent individuals than among more
intelligent individuals. More intelligent individuals, who possess higher levels of general
intelligence and thus greater ability to solve evolutionarily novel problems, may face less
difficulty in comprehending and dealing with evolutionarily novel entities and situations.
In contrast, less intelligent individuals may face greater difficulty in dealing with
evolutionarily novel entities and situations than more intelligent individuals.
Consistent with this reasoning, more intelligent individuals are more likely to make the
theoretically rational choice to defect in one-shot Prisoner’s Dilemma games (Kanazawa &
Fontaine, 2013). This may be because more intelligent individuals are better able to
comprehend the evolutionarily novel situations of complete anonymity and absolutely no
possibility of knowing future interactions and make the rational decision to defect. In
contrast, less intelligent individuals may have greater difficulty comprehending such
676 Norman P. Li and Satoshi Kanazawa
evolutionarily novel situations and, as a result, make the theoretically irrational (albeit
evolutionarily rational) decision to cooperate.
The Savanna Principle in evolutionary psychology, applied to life satisfaction, may
suggest that it may not be only the consequences of a given situation in the current
environment that influence individuals’ life satisfaction but also what its consequences
would have been in the ancestral environment. Having implicit difficulty comprehending
and dealing with evolutionarily novel situations, the human brain may respond to the
ancestral consequences of the current situation and individuals’ life satisfaction may
fluctuate accordingly. The evolutionary constraints on the human brain may incline
individuals to experience a given situation as if it were taking place in the ancestral
environment, not in the current environment, and be subject to its ancestral
consequences for life satisfaction. Further, the effect of such ancestral consequences of
current situations on life satisfaction may be greater among less intelligent individuals, for
whom the evolutionary constraints specified by the Savanna Principle are stronger, than
among more intelligent individuals, for whom they are weaker.
The savanna theory of happiness therefore suggests that, having implicit difficulty
comprehending and dealing with evolutionarily novel situations, the human brain may
respond to the ancestral consequences of the current situation and individuals’ subjective
well-being may fluctuate accordingly (Kanazawa & Li, 2015). Situations and circum-
stances that would have increased our ancestors’ life satisfaction in the ancestral
environment may still increase our life satisfaction today, and those that would have
decreased their life satisfaction then may still decrease our life satisfaction today. The
savanna theory further suggests that such effects of ancestral consequences on current life
satisfaction may be stronger among less intelligent individuals than among more
intelligent individuals.
Positive psychologists have long debated the precise definition of happiness and
related concepts, such as life satisfaction and subjective well-being (Miao, Koo, & Oishi,
2013; Oishi, Graham, Kesebir, & Galinha, 2013; Pavot & Diener, 2013). Even though our
empirical analyses below used a measure of global life satisfaction, the savanna theory of
happiness is not committed to any particular definition and is compatible with any
reasonable conception of happiness, subjective well-being, and life satisfaction (cognitive
vs. affective; hedonic vs. eudaimonic, etc.). The theory does, however, treat happiness as a
state, rather than a trait; it cannot explain the (partly genetically determined) ‘happiness
set point’ (Headey & Wearing, 1989), to which individuals tend to return after momentary
and situational perturbations to their baseline levels of happiness. The theory instead
explains such temporary and situational fluctuations from the happiness baseline as a
function of the potential evolutionary consequences of the current situations and
circumstances.
In this study, we provide empirical tests of the two hypotheses derived from the
savanna theory of happiness. As empirical test cases, we focus on two factors that
characterize basic differences in the social landscape of ancestral versus modern
environments and thus might affect life satisfaction: population density and friendships.
Population density
The beauty of the country besides, the pleasures of a country life, the tranquility of mind
which it promises, and wherever the injustice of human laws does not disturb it, the
independency which it really affords, have charms that more or less attract everybody.
Adam Smith, Wealth of Nations (1776, III.1.3)
The savanna theory of happiness 677
Ruralites in economically developed nations tend to be happier than their urbanite
counterparts (Berry & Okulicz-Kozaryn, 2009; Easterlin, Angelescu, & Zweig, 2011). Even
in the still developing China, rural residents report higher levels of subjective well-being
than urban residents, despite the fact that city dwellers are vastly wealthier (Knight &
Gunatilaka, 2010). In the United States, there is an ‘urban–rural happiness gradient’,
whereby residents of rural areas and small towns are happier than those in suburbs, who
in turn are happier than those in small central cities, who in turn are happier than those in
large central cities (Berry & Okulicz-Kozaryn, 2011). What accounts for the differences in
happiness across these residential settings? Why are ruralites happier than urbanites?
A current leading explanation for the lower level of life satisfaction in cities is that
urban life is accompanied by numerous ‘social ills’, such as anomie, alienation, social
disorganization, and depression (Berry & Okulicz-Kozaryn, 2011; Evans, 2009; Wirth,
1938). An fMRI study shows that the brains of current city dwellers and those who grew
up in cities respond to stress with greater activities than those of current country dwellers
and those who grew up in the country (Lederbogen et al., 2011). These studies, however,
simply raise another question: Why does the human brain perceive urban life, but not rural
life, as stressful, alienating and depressing? Why does urban life, but not rural life, induce
alienation and depression?
The savanna theory of happiness offers one potential answer. There is converging
evidence to suggest that our ancestors may have lived in groups of about 150 individuals.
Comparative data on relative neocortex size in the brain and the group size among 38
genera of primates suggest that the natural size for human groups given its neocortex ratio
is about 150 (Dunbar, 1992). Indeed, the mean band or village size of nine modern hunter–
gatherer societies is 148.4 (Dunbar, 1993). Computer simulations of the evolution of risk
aversion suggest that it can only evolve in small groups of about 150 individuals (Hintze,
Olson, Adami, & Hertwig, 2013). The mean size of personal networks suggested by the
number of annual Christmas cards sent is 153.5 (Hill & Dunbar, 2003). The mean size of
social networks suggested by two ‘small world’ experiments is 134 (Killworth, Bernard, &
McCarty, 1984). The typical size of Neolithic villages in Mesopotamia was 150–200 (Oates,
1977); the mean size of Hutterite farming communities in Canada is 107 (Mange & Mange,
1980); and the mean size of Amish parishes in central Pennsylvania is 112.8 (Hurd, 1985).
The typical size of military unit in the classical Roman army was 120–130, and the mean
company size of armies in World War II was 180 (MacDonald, 1955). Gautney and Holliday
(2015) estimate the population density in Africa and Eurasia during the Pleistocene Epoch
to be between 0.03 and 0.12 individuals/km
2
, about one-tenth of the population density
of the least dense state in the United States (Alaska =0.46 individuals/km
2
) in 2010 but
denser than the least dense counties in the United States (Yukon-Koyukuk Census Area,
Alaska =0.015; and Lake and Peninsula County, Alaska =0.027).
When the numberof individuals in a group exceeds 150–200, the group typicallyfissions
into and forms two separate groups, because in larger groups social organization based on
cooperation and reciprocity becomes exceedingly difficult (Chagnon, 1979). Because the
major constraint on human group size is cognitive (Dunbar, 1992, 1993), it is possible that,
as the population density becomes too high, the human brain feels uneasy and
uncomfortable, and such unease and discomfort may translate into reduced subjective
well-being. For example, job satisfaction is significantly negatively associated with
organizational size (Indik, 1965; Porter & Lawler, 1965). The savanna theory of happiness
may therefore suggest that group sizes and population densities much higher than were
typical in the ancestral environmentmay decrease subjective well-being. It further suggests
678 Norman P. Li and Satoshi Kanazawa
that such a negative effect of population density on happiness may interact with general
intelligence, such that the negative effect is greater among less intelligent individuals than
among more intelligent individuals. We tested these hypotheses in Study 1A.
Friendships
‘Friends show their love in times of trouble, not in happiness’.
Euripides (480BC–406BC)
‘Lots of people want to ride with you in the limo, but what you want is someone who will take
the bus with you when the limo breaks down’. Oprah Winfrey (1954–)
One of the most important determinants of life satisfaction is the quality of social
relationships, in particular friendships (Diener & Seligman, 2004, pp. 18–20; Dolan,
Peasgood, & White, 2008, pp. 106–108). The more friends one has, and the more time
one spends with them, the happier one tends to be on average, although recent
studies suggest that the quality of friendships is more important for happiness than
their quantity (Demir, Orthel, & Andelin, 2013; Demir, Orthel-Clark, €
Ozdemir, &
€
Ozdemir, 2015). The association between satisfaction with friendships and life
satisfaction is particularly stronger in more individualistic cultures (Diener & Diener,
1995; Li & Cheng, 2015). While the strong impact of friendships on subjective well-
being may make intuitive sense, why are friends important for life satisfaction
theoretically?
Perhaps the strong effect of friendships on life satisfaction is too obvious to
explain; to our knowledge, only one scholar has offered a systematic explanation for
why friendships increase happiness. Meliksßah Demir and colleagues (Demir, 2015;
Demir & Davidson, 2013; Demir & €
Ozdemir, 2010; Demir, €
Ozen, & Do
gan, 2012;
Demir, €
Ozen, Do
gan, Bilyk, & Tyrell, 2011) argue that friendships increase happiness
because they satisfy some basic psychological needs, such as relatedness, the
knowledge that one matters to others, and the desire to share and amplify good news
and events (captured in the Swedish proverb ‘Shared joy is a double joy, shared
sorrow is half a sorrow’). Demir’s explanation, however, raises even more
fundamental questions. Why do humans have these basic psychological needs in
the first place? And why can they be satisfied only (or primarily) by friends?
The savanna theory of happiness can provide one potential answer to such
fundamental questions. As noted before, our ancestors lived as hunter–gatherers in small
bands of about 150 individuals (Dunbar, 1992, 1993). In such settings, having frequent
contact with lifelong friends and allies was likely necessary for survival and reproduction
for both sexes, as evidenced by studies of both contemporary hunter–gatherers (Apicella,
Marlowe, Fowler, & Christakis, 2012; Hruschka, 2010; Lewis et al., 2015) and our primate
cousins (Smuts, 1985; de Waal, 1982). For instance, cooperative alliances may have
allowed men to overcome critical challenges posed by hunting and warfare (Bowles,
2009; Geary, Byrd-Craven, Haord, Vigil, & Numtee, 2003), and close relationships among
unrelated women may have facilitated joint childcare and allomothering (Hrdy, 2009).
Likewise, reciprocal food-sharing among group members occurs commonly in modern-
day hunter–gatherers and may have allowed our ancestors to survive despite success or
failure in hunting and gathering on any given day (Hill & Hurtado, 1996).
The savanna theory of happiness 679
The evolutionary significance of friendships and alliances is suggested by
numerous studies indicating that ostracism is invariably painful and distressful across
various contexts and sources (Williams, Forgas, & von Hippel, 2005). In one
experiment, participants earned money to be excluded in a game and lost money to
be included. Despite earning more money than others, those who were ostracized still
experienced pain (van Beest & Williams, 2006). Indeed, fMRI studies show that being
ostracized activates the same region of the brain that lights up when individuals
experience physical pain (Eisenberger, Lieberman, & Williams, 2003). Given the
available evidence, it is reasonable to assume that humans evolved to detect ostracism
(Gruter & Masters, 1986) largely because friendship ties and alliances were very
important for the survival and reproductive success of our ancestors (Lewis et al.,
2015).
In contrast, survival and reproduction today depend increasingly more on one’s
ability to navigate myriad evolutionarily novel entities such as the internet, govern-
ments, banks, corporations, trusts, and the legal system. Instead of relying on reciprocal
cooperation with friends and allies for basic needs, modern-day individuals deal with
strangers or faceless entities and have no way of identifying those involved in the
procurement and processing of necessities such as food (Pollan, 2006). It is entirely
possible for individuals in modern society to survive and reproduce successfully without
having any friends; friendships are not as critically necessary today for day-to-day living
as they were in the ancestral environment. Hruschka (2010, p.2) notes, in a book
entirely devoted to the importance of friendship, that ‘while friends make us happy and
help us in small ways, it is not entirely clear that they are important in the high-stakes
game of survival and reproduction’. In 1998, 9% of respondents in the General Social
Survey in a representative sample of non-institutionalized American adults responded
that they did not have any good friends to whom they felt close (Smith, Marsden, &
Hout, 2015, p. 639).
The savanna theory of happiness therefore suggests that the human brain may have
implicit difficulty comprehending and dealing with life without frequent contact with
close friends and allies, and such difficulty may decrease individuals’ subjective well-
being. Further, such an effect of friendships on life satisfaction may be particularly
stronger among less intelligent individuals, who are likely less able to adapt to
evolutionarily novel circumstances such as a dearth of close friends. Thus, we expect
friendships to have a positive effect on subjective well-being and further (and more
importantly) that such an effect will be stronger among less intelligent individuals. We
tested these hypotheses in Study 1B.
STUDY 1A
Methods
Data and participants
We used Wave III data from the National Longitudinal Study of Adolescent Health (Add
Health), which consisted of personal interviews held in 2001–2002 with 15,197
individuals aged 18–28 (M=21.96, SD =1.77). The participants were part of a large
sample of students originally selected in 1994–1995 (Wave I) from middle and high
schools that were representative of US schools with respect to region of country,
urbanicity, school size, school type, and ethnicity. For further details of the sampling and
study design, see http://www.cpc.unc.edu/projects/addhealth/design.
680 Norman P. Li and Satoshi Kanazawa
Dependent variable: Global life satisfaction
Add Health asked its respondents ‘How satisfied are you with your life as a whole?’:
1=very dissatisfied,2=dissatisfied,3=neither satisfied nor dissatisfied,4=satis-
fied, and 5 =very satisfied (reverse coded). We used this measure of life satisfaction as the
dependent variable in our ordinal regression analysis.
1
Independent variable: Population density
Add Health measured the population density at the level of census block group (a
subdivision of a census tract and the smallest geographic unit for which the Census Bureau
tabulates aggregate data), census tract, county, and state. It was measured as the number
of persons in thousands/km
2
.
The distributions of population density at all levels were extremely positively skewed
(skewness: block group =6.780; census tract =7.449; county =8.702; state =17.460).
We therefore took the natural log of the measures of population density, which nearly
eliminated the skewness (skewness after log normalization: block group =.809; census
tract =.684; county =.023; state =.121). We used the natural logs of measures of
population density in our regression analyses below.
Independent variable: Intelligence
Add Health measured respondents’ intelligence by an abbreviated version of the Peabody
Picture Vocabulary Test. Their raw scores were transformed into the standard IQ metric,
with a mean of 100 and a standard deviation of 15. The Peabody Picture Vocabulary Test is
properly a measure of verbal intelligence. However, verbal intelligence is known to be
highly correlated with and thus heavily load on general intelligence (Huang & Hauser,
1998; Miner, 1957; Wolfle, 1980).
Control variables
In addition, we controlled for the following characteristics of the respondent: sex
(0 =female,1=male); age (in years); ethnicity (with three dummies for African
American, Asian American, and Native American, with White American as the reference
category); education (in years of formal schooling); and current marital status (1 =cur-
rently married,0 =otherwise). All of these variables are known correlates of happiness
(Dolan et al., 2008). Preliminary analysis showed that the respondent’s earnings had no
association with life satisfaction among Add Health respondents (r=.013, p=.116,
n=14,414), perhaps because of their relative youth and little variance in earnings
(M=11,744, median =8,000 SD =17,289, IQR =16,500, n=14,425). This was
consistent with earlier studies, which showed that variance in earnings generally
increased with age (Beach, Finnie, & Gray, 2010; Caswell & Kluge, 2015; Lam & Levison,
1992).
1
More sophisticated statistical procedures like structural equation modelling (SEM) or multilevel modelling (MLM) are not
feasible with our data. SEM is not feasible because we have only one indicator each for all of the variables in our analysis, and MLM
is not feasible because, while we know the population density of the county or the state of residence, for example, we do not know
in which county or state the Add Health respondents reside. (Add Health is extremely concerned about privacy issues and does not
make much individually identifiable information available in the data.) So we cannot perform MLM by nesting individual
respondents in the county or state of their residence.
The savanna theory of happiness 681
Results
The results of the ordinal regression analysis appear in Table 1. Whether measured at the
level of block group, census tract, county, or state, population density was significantly
negatively associated with Add Health respondents’ life satisfaction (Columns 1–4; block
group: b=.058; census tract: b=.055; county: b=.076; state: b=.100;
p<.001 for all). This did not change at all when we controlled for sex, age, ethnicity,
education, and current marital status (Columns 5–8). Consistent with the prediction
derived from the savanna theory of happiness, the higher the population density of the
immediate environment, the less happy Add Health respondents were.
Further analyses (Columns 9–12) showed that, consistent with the prediction, the
negative association between population density and life satisfaction was significantly
stronger among less intelligent individuals than among more intelligent individuals. The
interaction terms between population density and intelligence were statistically
significantly positive for block group (b=.002, p<.001), census tract (b=.002,
p<.001), county (b=.002, p<.001), and state (b=.003, p=.010).
Figure 1 presents the statistical interaction graphically. While county population
density had a significantly negative association with life satisfaction among both less
intelligent (with IQ of 81.39, one standard deviation below the mean) and more
intelligent (with IQ of 115.57, one standard deviation above the mean) individuals,
the negative association was greater among less intelligent individuals (M=4.2617
vs. 4.1090) than among more intelligent individuals (M=4.2161 vs. 4.1495). Put
another way, in a county with low population density (41 persons/km
2
, one
standard deviation below the mean), less intelligent individuals had higher mean life
satisfaction than more intelligent individuals did. In contrast, in a county with high
population density (937 persons/km
2
, one standard deviation above the mean), more
intelligent individuals had higher mean life satisfaction than less intelligent
individuals did.
2
There is currently no accepted method of computing effect sizes or standardized
regression coefficients in ordinal regression and other generalized linear models, partly
because the effects of independent variables on the dependent variable in ordinal
regression are proportional, not constant. However, the mean differences in life
satisfaction in low and high densities in Figure 1 allowed us to compute Cohen’s das
an estimate for the effect of population density on life satisfaction. Given that the
standard deviation of life satisfaction was 0.815, the mean difference in life satisfaction
for low-IQ individuals (0.1527) translated to d=.19, and that for high-IQ
individuals (0.0666) to d=.08. The effect of population density on life satisfaction
was therefore more than twice as large for low-IQ individuals than for high-IQ
individuals.
Given that our data are correlational and population density and life satisfaction were
measured at the same time, we cannot rule out an opposite causal order to what we
hypothesize in the savanna theory of happiness, where people who experience higher life
satisfaction are more likely to move to rural areas. This does not appear to be the case.
While life satisfaction at Wave III was significantly positively associated with the distance
2
In generalized linear models (such as ordinal regression that we employed here), the independent variables have proportional
effects on the dependent variable and constant effects on the logit (natural log of odds of a respondent being in one category of
the dependent variable versus another). As a result, simple slope analysis, of the kind described in Aiken and West (1991) for OLS
regression, cannot be performed on raw dependent variable and must instead be performed on the logit. Because the logit has no
intuitive or readily interpretable meaning, we have chosen not to perform a simple slope analysis.
682 Norman P. Li and Satoshi Kanazawa
Table 1. Population density and life satisfaction
(1)
Block group
(2)
Census tract
(3)
County
(4)
State
Population density .058*** .055*** .076*** .100***
(.008) (.008) (.010) (.019)
Age
Sex
Ethnicity
African American
Asian American
Native American
Education
Currently married
Intelligence
Intelligence*population density
Threshold
Y=15.106 5.095 4.996 4.837
(.106) (.106) (.107) (.118)
Y=23.135 3.124 3.025 2.866
(.041) (.041) (.044) (.066)
Y=31.594 1.584 1.484 1.326
(.022) (.022) (.027) (.056)
Y=4 .570 .581 .682 .836
(.017) (.017) (.024) (.055)
2Log Likelihood 23549.434*** 19396.470*** 5169.710*** 993.527***
Cox & Snell pseudo-R
2
.003 .003 .004 .002
Number of cases 14,877 14,877 14,877 14,877
(5)
Block group
(6)
Census tract
(7)
County
(8)
State
Population density .049*** .047*** .068*** .079***
(.008) (.008) (.010) (.019)
Age .052*** .053*** .051*** .056***
(.009) (.009) (.009) (.009)
Sex .118*** .118*** .118*** .120***
(.032) (.032) (.032) (.032)
Ethnicity
African American .178*** .181*** .173*** .193***
(.038) (.038) (.038) (.038)
Asian American .253*** .258*** .248*** .289***
(.058) (.057) (.058) (.057)
Native American .203** .204** .209** .226***
(.069) (.069) (.068) (.068)
Education .151*** .152*** .152*** .148***
(.008) (.008) (.008) (.008)
Currently married .715*** .715*** .713*** .724***
(.044) (.044) (.044) (.044)
Intelligence
Intelligence*population density
Threshold
Y=14.237 4.233 4.084 4.152
(.236) (.236) (.239) (.243)
Continued
The savanna theory of happiness 683
Table 1. (Continued)
(5)
Block group
(6)
Census tract
(7)
County
(8)
State
Y=22.258 2.254 2.105 2.173
(.215) (.215) (.218) (.222)
Y=3.690 .686 .536 .606
(.212) (.212) (.216) (.220)
Y=4 1.539 1.543 1.694 1.621
(.212) (.212) (.216) (.220)
2Log Likelihood 32503.061*** 31976.678*** 27210.467*** 21051.828***
Cox & Snell pseudo-R
2
.042 .042 .043 .041
Number of cases 14,811 14,811 14,811 14,811
(9)
Block group
(10)
Census tract
(11)
County
(12)
State
Population density .248*** .222*** .284*** .435**
(.054) (.054) (.059) (.141)
Age .050*** .051*** .049*** .054***
(.010) (.010) (.010) (.009)
Sex .118*** .118*** .119*** .120***
(.032) (.032) (.032) (.032)
Ethnicity
African American .192*** .191*** .188*** .197***
(.041) (.041) (.041) (.040)
Asian American .256*** .263*** .255*** .288***
(.059) (.059) (.059) (.059)
Native American .194** .198** .203** .220**
(.069) (.069) (.069) (.069)
Education .153*** .153*** .154*** .149***
(.009) (.009) (.009) (.009)
Currently married .710*** .710*** .708*** .716***
(.045) (.045) (.045) (.045)
Intelligence .001 .000 .003* .010*
(.001) (.001) (.001) (.004)
Intelligence*population density .002*** .002*** .002*** .003*
(.001) (.001) (.001) (.001)
Threshold
Y=14.227 4.167 3.765 3.098
(.253) (.253) (.266) (.469)
Y=22.258 2.197 1.795 1.129
(.233) (.233) (.247) (.458)
Y=3.691 .630 .227 .436
(.230) (.230) (.244) (.457)
Y=4 1.537 1.596 2.001 2.661
(.231) (.231) (.245) (.458)
2Log Likelihood 31731.139*** 31710.645*** 31321.322*** 30674.162***
Cox & Snell pseudo-R
2
.043 .042 .043 .041
Number of cases 14,278 14,278 14,278 14,278
Note. Main entries are unstandardized regression coefficients.
Numbers in parentheses are standard errors.
*p<.05; **p<.01; ***p<.001.
684 Norman P. Li and Satoshi Kanazawa
Add Health respondents moved between Waves I and III (r=.022, p=.008, n=14,801),
the distance moved was more strongly positively associated with Wave III population
density (block group: r=.072; census tract: r=.076; county: r=.060; state: r=.046;
p<.001, n=14,813 for all). In other words, longer-distance movers were more likely to
move to urban areas, not rural areas, and they became more satisfied with their life despite
their long-distance move (to urban areas), not because of it. As a result, controlling for the
distance moved strengthens the negative association between population density and life
satisfaction, not weakens or eliminates it, at all levels except for state, where the
association remains unchanged.
Discussion
Consistent with our prediction derived from the savanna theory of happiness, population
density measured at the block group, census tract, county, and state levels had a
significantly negative association with Add Health respondents’ life satisfaction. The
lower the population density of the immediate environment and the closer it was to what
it was in the ancestral environment, the higher life satisfaction Add Health respondents
experienced.
Further, as predicted, the association between population density and life satisfaction
was significantly stronger among less intelligent individuals than among more intelligent
individuals. Less intelligent individuals might have had greater difficulty comprehending
and dealing with the evolutionary novelty of living in a high population density area and
become less satisfied with life as a result. In contrast, more intelligent individuals might
have had less difficulty with living in a high population density area and their life
satisfaction might not have been affected as much. In low population density, less
intelligent individuals on average had higher life satisfaction than more intelligent
individuals did, but in high population density, more intelligent individuals on average had
higher life satisfaction than less intelligent individuals did.
Figure 1. Interaction effect between county population density and intelligence on life satisfaction.
The savanna theory of happiness 685
Interestingly, Add Health respondents’ intelligence was significantly negatively
associated with the natural log of population density (block group: r=.041, p<.001;
census tract: r=.027, p=.001; county: r=.038, p<.001; state: r=.028,
p<.001; n=14,351 for all). It means that more intelligent individuals did not selectively
migrate to urban areas, and less intelligent individuals did not selectively migrate to rural
areas, in order to take advantage of their respective levels of intelligence to become more
satisfied with life. We believe there are two potential (and non-mutually exclusive)
reasons for this. First, individuals in general may not be (either consciously or
unconsciously) aware of the negative effect of population density on happiness and its
divergent effects by intelligence. Second, individuals may not have complete freedom to
move where they want in order to pursue life satisfaction, especially at such a young age.
They may be constrained by the requirements of their education, employment, and family.
STUDY 1B
Methods
Data and participants, dependent variable (global life satisfaction), one of the indepen-
dent variables (intelligence), and all control variables for Study 1B were identical to those
in Study 1A.
Independent variable: Frequency of socialization with friends
Add Health asked its respondents ‘In the past 7 days, how many times did you just “hang
out” with friends, or talk on the telephone for more than five minutes?’: 0 =not at all to
7=7 or more times. Add Health did not clearly define or explain to its respondents
exactly who counted as friends. This is unlikely to be a problem in the current study,
however, because Demir (2015, p. vii) concludes, based on the review of a large number
of studies in the literature, that ‘friendship is related to happiness regardless of the ways
the constructs were assessed’.
Results
The results of the ordinal regression analyses appear in Table 2. As Column (1) shows,
when entered alone, frequency of socialization with friends had no significant association
with life satisfaction (b=.008, p=.201; r=.010, p=.208). This was because current
marital status masked their association. Currently married Add Health respondents
simultaneously were happier (4.11 vs. 4.35), t(15,155) =13.840, p<.001, and
socialized with their friends less frequently (4.56 vs. 3.24), t(15,117) =26.201, p<.001.
This is consistent with the dyadic withdrawal hypothesis (Johnson & Leslie, 1982).
3
As
Column (2) shows, once current marital status was controlled, frequency of socialization
with friends had a significantly positive association with life satisfaction (b=.031,
p<.001), and, as Column (3) shows, this did not change even when we further controlled
for age, sex, ethnicity, and education.
Column (4) shows that, consistent with the prediction, the positive association
between frequency of socialization with friends and life satisfaction was significantly
3
We thank an anonymous reviewer for alerting us to the dyadic withdrawal hypothesis.
686 Norman P. Li and Satoshi Kanazawa
stronger among less intelligent individuals than among more intelligent individuals. The
interaction term between intelligence and frequency of socialization with friends was
significantly negative (b=.001, p=.014).
Figure 2 presents the statistical interaction graphically. Among less intelligent
individuals (with a mean IQ of 81.39), frequency of socialization with friends had a
significantly positive effect on life satisfaction. Those who socialized with friends more
frequently (6.71, nearly every day) had a significantly higher life satisfaction (M=4.1586)
than those who socialized with friends less frequently (1.95, less than twice a week)
(M=4.1163). In contrast, among more intelligent individuals (with a mean IQ of 115.57),
those who socialized with friends more frequently were actually less satisfied with life
Table 2. Frequency of socialization with friends and life satisfaction
(1) (2) (3) (4)
Frequency of socialization with friends .008 .031*** .017* .103**
(.006) (.007) (.007) (.035)
Currently married .663*** .773*** .765***
(.043) (.045) (.045)
Age .060*** .059***
(.009) (.009)
Sex .126*** .125***
(.031) (.032)
Ethnicity
African American .194*** .203***
(.038) (.040)
Asian American .301*** .301***
(.057) (.058)
Native American .215** .210**
(.068) (.069)
Education .145*** .147***
(.008) (.009)
Intelligence .003
(.002)
Intelligence*frequency of
socialization with friends
.001*
(.000)
Threshold
Y=15.060 4.876 4.389 4.028
(.108) (.109) (.236) (.282)
Y=23.113 2.928 2.434 2.084
(.049) (.051) (.215) (.264)
Y=31.573 1.382 .866 .518
(.035) (.037) (.213) (.262)
Y=4 .590 .806 1.364 1.710
(.033) (.036) (.213) (.262)
2Log Likelihood 255.142 408.003*** 15615.357*** 29472.587***
Cox & Snell pseudo-R
2
.000 .016 .041 .041
Number of cases 15,111 15,111 15,047 14,513
Note. Main entries are unstandardized regression coefficients.
Numbers in parentheses are standard errors.
*p<.05; **p<.01; ***p<.001.
The savanna theory of happiness 687
(M=4.1063) than those who socialized with friends less frequently (M=4.1311). The
statistical interaction was such that more intelligent individuals were actually less satisfied
with life if they socialized with their friends more frequently. Among low-IQ individuals,
the mean difference in life satisfaction between those who socialize with friends more and
less frequently (0.423) translated to d=.05, and that among high-IQ individuals
(0.0248) to d=.03.
Given that our data are correlational and frequency of socialization with friends and life
satisfaction were measured at the same time, we cannot rule out an opposite causal order
to what we hypothesize, where happier people choose to socialize with their friends more
frequently. This may potentially be a problem because our measure of frequency of
socialization with friends referred to recent past (‘In the past 7 days’), while the measure
of life satisfaction was global (‘as a whole’). We are sure there are some mutual influences
between life satisfaction and frequency of socialization with friends, but there are a few
considerations, suggesting that the results largely reflect our hypothesized causality. For
instance, Baker, Cahalin, Gerst, and Burr (2005) showed that the positive effect of seeing
family and friends on subjective well-being remained even after controlling for the earlier
level of life satisfaction in a previous wave of a longitudinal survey. Similarly, in our data,
frequency of socialization with friends was still significantly associated with life
satisfaction even after happiness at Waves I and II (measured by the question ‘How
often was each of the following things true during the past seven days? You were happy’.
0=never or rarely,1 =sometimes,2 =a lot of the time,3 =most of the time or all of
the time), in addition to current marital status, was controlled (b=.018, p=.016).
Discussion
Consistent with our prediction derived from the savanna theory of happiness, and with
past empirical studies, frequency of socialization with friends had a significantly positive
Figure 2. Interaction effect between frequency of socialization with friends and intelligence on life
satisfaction.
688 Norman P. Li and Satoshi Kanazawa
association with life satisfaction among Add Health respondents, once current marital
status, which acted as a suppressor, was controlled. The more frequently individuals
socialized with their friends, the more satisfied they were with their lives.
Further, as predicted by the savanna theory of happiness, the association between
socialization with friends and life satisfaction was significantly stronger among less
intelligent individuals than among more intelligent individuals. Less intelligent individuals
might have had greater difficulty comprehending and dealing with the evolutionary
novelty of not associating with friends and allies regularly and become less satisfied with
life as a result. In contrast, more intelligent individuals might have had less difficulty with
not associating with friends and allies regularly and their life satisfaction might not have
been affected as much. In fact, extremely (+1 SD) intelligent individuals even appeared to
become more satisfied with life when their frequency of socialization with friends was
lower.
Interestingly, Add Health respondents’ intelligence was significantly positively
associated with the frequency of socialization with friends (r=.121, p<.001,
n=14,581); more intelligent individuals socialized with their friends more frequently.
The association between intelligence and frequency of socialization with friends was
stronger among currently unmarried individuals (r=.131, p<.001, n=12,091) than
among currently married individuals (r=.083, p<.001, n=2,490). It means that
more intelligent individuals did not voluntarily decrease their frequency of socialization
with friends, and less intelligent individuals did not voluntarily increase it, in order to
take advantage of their respective levels of intelligence to increase their life satisfaction.
As in Study 1A, we believe there are two potential (and non-mutually exclusive)
reasons for this. First, individuals in general may not be (either consciously or
unconsciously) aware of the divergent effect of socialization with friends on happiness
by intelligence. Second, individuals may not have complete control over how
frequently to socialize with their friends (or how many friends to have). Friendship
is a two-way street, and friends must mutually seek each other to establish friendship
and socialize together –something that may be increasingly difficult to do in transient
modern environments. More intelligent individuals may simply have more friends to
begin with.
Combined models
To investigate the independence of the effects observed above, we examined a model in
which the two major predictors of life satisfaction in our study (population density and
friendship) were simultaneously considered. The two variables very weakly correlated
with each other in the Add Health data (correlation with frequency of socialization with
friends: census block r=.046, p<.001; census tract r=.050, p<.001; county r=.034,
p<.001; state r=.021, p=.012; n=14,839 for all). As a result, the associations of these
predictors with life satisfaction were statistically independent. Net of each other (and
current marital status), population density at all levels, and frequency of socialization with
friends were still statistically significantly associated with life satisfaction, and this did not
change at all when we further controlled for age, sex, ethnicity, and education. When
entered together in the same ordinal regression equations, the predicted interaction terms
between each predictor and intelligence still remained statistically significant. It therefore
appeared that population density and socialization with friends (and their interactions
with intelligence) were statistically independent predictors of life satisfaction. There was
The savanna theory of happiness 689
no significant three-way interaction effect among intelligence, population density
(measured at any level), and friendship.
GENERAL DISCUSSION
While positive psychologists in the last few decades have accumulated an impressive
amount of empirical knowledge about who under what circumstances or in what
conditions are happier than whom, there have been few general theories in positive
psychology that explain why some individuals are happier than others (other than to point
to genetic predisposition and the heritability of happines s). In this paper, we proposed the
savanna theory of happiness, which suggests that it is not only the current consequences
of a given situation but also its ancestral consequences that affect subjective well-being
(Kanazawa & Li, 2015). That is, individuals’ life satisfaction may fluctuate with what the
situation would have meant in the ancestral environment. The savanna theory of
happiness further suggests that such an effect of the ancestral consequences on life
satisfaction may be greater among less intelligent individuals than among more intelligent
individuals.
We chose two varied factors representing basic differences in the social landscape
between modern and ancestral environments –population density and socialization with
friends –as empirical test cases for the theory in this paper. Despite their widely varied
nature, population density and socialization with friends had remarkably similar
associations with life satisfaction among the Add Health respondents. Both factors had
the anticipated main association: Population density, whether measured at the level of
census block group, census tract, county or state, was significantly negatively associated
with life satisfaction, and socialization with friends, once current marital status was
controlled, was significantly positively associated with life satisfaction. More importantly,
we observed the predicted statistical interaction with intelligence for both factors; in fact,
in the case of socialization with friends, we observed the reversal of the main association
among the extremely intelligent. More intelligent individuals actually experienced higher
life satisfaction with lower frequency of contact with friends.
Our studies uniquely highlight three important determinants of individual differences
in life satisfaction: population density, friendship, and intelligence. The importance of
population density can potentially explain the ‘urban–rural happiness gradient’ (Berry &
Okulicz-Kozaryn, 2011) observed in the United States and other nations, where ruralites
are significantly happier than suburbanites, who are in turn significantly happier than
urbanites. In our study, we treated population density as a continuous variable and
showed that it was negatively associated with life satisfaction.
The importance of friendship in our analysis is consistent with a large number of
previous studies on life satisfaction (Diener & Seligman, 2004, pp. 18–20; Dolan et al.,
2008, pp. 106–108). However, to the best of our knowledge, no one else has
demonstrated the statistical interaction between socialization with friends and intelli-
gence. Nor has anyone demonstrated that extremely intelligent individuals may be less
satisfied with life if they socialized with their friends more frequently.
The current research adds to a growing body of literature indicating that the human
brain may have difficulty with conditions that are mismatched to the natural environments
of the ancestral past, when psychological mechanisms are hypothesized to have evolved
(Tooby & Cosmides, 1990). Building on previous work on the evolutionary constraints on
the human brain (Burnham & Johnson, 2005; Hagen & Hammerstein, 2006; Kanazawa,
690 Norman P. Li and Satoshi Kanazawa
2004b), our research extends the consequences of such cognitive limitations and
mismatch from areas as diverse as relationship satisfaction (Russell, McNulty, Baker, &
Meltzer, 2014), eating disorders (Li, Smith, Griskevicius, Cason, & Bryan, 2010), and
leadership (van Vugt & Ronay, 2014) to affective states like happiness. Our results also
show important interaction between such evolutionary limitations and general intelli-
gence and suggest that more intelligent individuals might suffer from affective
consequences of evolutionary limitations on the brain to a significantly lesser degree
than less intelligent individuals might.
Further, our study is consistent with a large body of literature on life history theory
(Charnov, 1993; Ellis, Figueredo, Brumbach, & Schlomer, 2009; Figueredo et al.,
2006), a framework rooted in the biological sciences that deals with how all
organisms, including humans, assess the environment in order to allocate their time,
energy, and resources optimally on survival and reproductive activities across the
entire lifespan. In an ancestral environment, higher population densities and less
interaction with friends may be indicative of high competition for scarce resources
(Ellis et al., 2009) and a dearth of coalitional support, respectively. As such, overall
mood may be calibrated downward and life history strategy adjustments may be made
accordingly. If individuals who are less generally intelligent are less capable of
weathering disadvantageous conditions, they may experience greater shifts in life
history strategy in conjunction with decrements in life satisfaction. Further research
may benefit from an investigation of the role that life satisfaction plays in relation to
life history strategies.
Limitations, implications, and future directions
Despite using a large, nationally representative sample with over 15,000 participants, a
limitation of the current studies is that the data are correlational. Although we have
considered and potentially ruled out some alternative explanations, we cannot be sure of
any causal relationships until experiments are conducted. Accordingly, future research
should attempt to manipulate the key factors and observe and measure the appropriate
outcomes. Of course, it may not be feasible to manipulate where people live or the
number of friends with whom they regularly socialize. Nevertheless, studies using modern
media such as printed photographs, television, and computers briefly to present social
and environmental stimuli have effectively induced people into believing that they are
exposed to real-life individuals or certain environments (Gutierres, Kenrick, & Partch,
1999; Li et al., 2010). Such methods may be adopted in order to manipulate exposure to
rural versus urban environments (van der Wal, Schade, Krabbendam, & van Vugt, 2013),
and larger versus smaller number of friends.
Aside from the alternative explanations that we tested within the studies, there may be
other explanations for our findings. For instance, general intelligence might equip
individuals to better handle all situations including high population densities and a lack of
interactions with friends. Although plausible, other studies seem to suggest that this is not
likely the case. General intelligence appears not to matter for solving evolutionarily
familiar problems in domains spanning mating, parenting, interpersonal relationships,
and wayfinding (Kanazawa, 2004a, 2010). Moreover, such an explanation does not
simultaneously address why higher population density poses a problem requiring greater
intelligence to rectify and why associating with friends is assumed to be desirable. Indeed,
we know of no alternative explanation that can parsimoniously explain all our findings
and address the ultimate causes of each issue. Nevertheless, the literature will benefit from
The savanna theory of happiness 691
a greater consideration and integration of proximate and ultimate explanations. We
encourage researchers from different perspectives to construct and test alternative
models.
Another potential limitation concerns the very small mean differences that are
reported here. While the mean differences we report are small by the standards of
experimental psychology, which relies on direct manipulations of independent variables
in controlled experiments, their magnitudes are reasonable by the standards of survey
research (see De Neve, Christakis, Fowler, and Frey (2012) for an example of a non-
experimental study which uses the same dependent variable from the same survey data
that we use here). Nevertheless, further studies are necessary to corroborate the pattern of
findings reported here.
Far from being conclusive, our findings raise numerous questions for future research.
For example, we do not know exactly what it is about denser populations that reduces
happiness. Is it the density itself, or other evolutionarily novel factors closely associated
with it, such as less access to greenery and nature (Kim et al., 2010), greater interactions
with strangers versus kin, friends, and acquaintances? Nor do we know exactly how more
intelligent individuals are better able to handle the evolutionary novelty of urban life.
Future research can explore specific factors to gain a better understanding of the
mechanisms underlying the relationships between population density, intelligence, and
happiness.
Likewise, (how) does modern technology change the nature and magnitude of the
effect of friendships and socialization with friends? The only way our ancestors could
interact and socialize with their friends and allies was face-to-face. Add Health only asks
about ‘hanging out’ with friends in person or talking to them on the phone. Is talking to
them via Skype or FaceTime the same as or better than talking to them on the phone? Do
Facebook friends (and online interactions with them) count as socializing with them?
Consistent with the Savanna Principle, there is some suggestion that humans have implicit
difficulty distinguishing between real friends and characters they repeatedly see on TV
because realistic electronic images of other humans did not exist in the ancestral
environment (Derrick, Gabriel, & Hugenberg, 2009; Gardner & Knowles, 2008;
Kanazawa, 2002). This seems to suggest that staying in touch on Skype and FaceTime is
just as good as doing so in person and better than talking on the phone, but socializing on
Facebook or with text messages has little effect. Indeed, Facebook use has been linked to
depression (for a review and evolutionary explanations, see Blease, 2015; Pinker, 2014),
which suggests that it is not an effective substitute for live encounters with friends.
Investigating such distinctions in light of the present findings may have implications for
public policy and policy-related research by suggesting how interventions can be
designed to increase people’s actual or perceived level of interaction with friends and thus
life satisfaction.
Similarly, our research here can be tied together with work on environmental
psychology to shed light on how spaces can be designed to reduce real or perceived
population density. For instance, urban centres can be designed to diffuse large
populations across wider areas. At the same time, given that many people spend a
significant amount of time commuting to and from work on crowded trains, busses, or
highways, it is possible that their perceptions of population density may be skewed
upwards –and their life satisfaction downwards –by the high population density
encountered during the commute. As such, increasing the number of trains and busses as
well as shifting work schedules by an hour or two may be cost-effective methods of
increasing life satisfaction.
692 Norman P. Li and Satoshi Kanazawa
Another potentially fruitful avenue to explore involves investigating people who are
satisfied with life despite living in crowded cities or associating with friends
infrequently. Aside from intelligence, there may be other traits, conditions, or
behaviours that are effectively boosting their life satisfaction. One such candidate is
the strength of one’s family ties, which tends to be positively associated with life
satisfaction (Schilling & Wahl, 2002) and negatively associated with depression
(Hammen & Brennan, 2002). It may be the case that strong relations with family
members can insulate individuals from the potential harms of evolutionary novelty
regardless of intelligence.
Finally, ancestral environments differ from modern ones in many ways. For example,
our ancestors tended to interact largely with kin and familiar others. In modern
environments, people vary more widely on such interac tions, with some individuals rarely
interacting with kin and dealing largely with strangers. Investigating this and other key
differences may yield interesting discoveries on what can impact life satisfaction. In
addition to helping inform the larger body of research on subjective well-being, such
work, and the avenues described above, may have implications for interventions aimed at
improving well-being in the modern society.
Conclusion
The empirical evidence from the two studies presented above provide tentative support
for the savanna theory of happiness, which explains why rural Americans tend to be
happier than their urban counterparts, and why Americans who socialize with friends
more frequently are happier. More importantly, the studies illustrate the value of
incorporating evolutionary perspectives to the study of subjective well-being. The current
paper adds to the growing body of knowledge on evolutionary mismatch theory
indicating that many of the ills of modern society might owe themselves to the disparity
between modern environments and the ancestral environments in which our brain
evolved and to which it is adapted (Buss, 2000; Hill & Major, 2013). Such work cuts across
all areas of psychology, including mating and relationships (Russell et al., 2014),
cooperation (Hagen & Hammerstein, 2006), clinical health (Li et al., 2010), behavioural
economics (van der Wal et al., 2013), and industrial–organizational psychology (van Vugt
& Ronay, 2014). In this rapidly growing area with far-reaching implications, the savanna
theory of happiness provides a novel answer to the question of what makes individuals
happier and why.
Acknowledgements
We thank Shimon Edelman for his comments on an earlier draft and Oi-Man Kwok, Felix
Thoemmes, Ming-Hong Tsai, and Stephen G. West for statistical help. See Add Health
acknowledgments at http://www.cpc.unc.edu/projects/addhealth/faqs/addhealth/index.
html#what-acknowledgment-should-be.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions.
Newbury Park, CA: Sage.
Apicella, C. L., Marlowe, F. W., Fowler, J. H., & Christakis, N. A. (2012). Social networks and
cooperation in hunter–gatherers. Nature,481, 497–501.
The savanna theory of happiness 693
Baker, L. A., Cahalin, L. P., Gerst, K., & Burr, J. A. (2005). Productive activities and subjective well-
being among older adults: The influence of number of activities and time commitment. Social
Indicators Research,73, 431–458.
Beach, C. M., Finnie, R., & Gray, D. (2010). Long-run inequality and short-run instability of men’s and
women’s earnings in Canada. Review of Income and Wealth,56, 572–596.
Berry, B. J. L., & Okulicz-Kozaryn, A. (2009). Dissatisfaction with city life: A new look at some old
questions. Cities,26, 117–124.
Berry, B. J. L., & Okulicz-Kozaryn, A. (2011). An urban–rural happiness gradient. Urban Geography,
32, 871–883.
Blease, C. (2015). Too few ‘likes’, too many ‘friends’? What Evolutionary Psychology tells us about
Facebook Depression. Review of General Psychology,19,1–13.
Bowles, S. (2009). Did warfare amongst ancestral hunter–gatherers affect the evolution of human
social behaviors? Science,324, 1293–1298.
Burnham, T. C., & Johnson, D. D. P. (2005). The biological and evolutionary logic of human
cooperation. Analyse & Kritik,27, 113–135.
Buss, D. M. (2000). The evolution of happiness. American Psychologist,55,15–23.
Caswell, H., & Kluge, F. A. (2015). Demography and the statistics of lifetime economic transfers
under individual stochasticity. Demographic Research,32, 536–588.
Chagnon, N. A. (1979). Mate competition, favoring close kin, and village fissioning among the
Yanomam€
o Indians. In N. A. Chagnon & W. Irons (Eds.), Evolutionary biology and human
social behavior: An anthropological perspective (pp. 86–131). North Scituate, MA: Duxbury.
Charnov, E. L. (1993). Life history invariants: Some explorations of symmetry in evolutionary
ecology. Oxford, UK: Oxford University Press.
Cosmides, L., & Tooby, J. (2013). Evolutionary psychology: New perspectives on cognition and
motivation. Annual Review of Psychology,64, 201–229.
De Neve, J.-E., Christakis, N. A., Fowler, J. H., & Frey, B. S. (2012). Genes, economics, and happiness.
Journal of Neuroscience, Psychology, and Economics,5, 193–211.
de Waal, F. B. M. (1982). Chimpanzee politics: Power and sex among apes. Baltimore, MD: Johns
Hopkins University Press.
Demir, M. (Ed.) (2015). Friendship and happiness: Across life-span and cultures. Dordrecht, the
Netherlands: Springer.
Demir, M., & Davidson, I. (2013). Toward a better understanding of the relationship between
friendship and happiness: Perceived responses to capitalization attempts, feelings of mattering,
and satisfaction of basic psychological needs in same-sex best friendships as predictors of
happiness. Journal of Happiness Studies,14, 525–550.
Demir, M., Orthel, H., & Andelin, A. K. (2013). Friendship and happiness. In I. Boniwell, S. David & A.
C. Ayers (Eds.), The Oxford handbook of happiness (pp. 860–870). Oxford, UK: Oxford
University Press.
Demir, M., Orthel-Clark, H., €
Ozdemir, M., & €
Ozdemir, S. B. (2015). Friendship and happiness among
young adults. In M. Demir (Ed.), Friendship and happiness: Across the life-span and cultures
(pp. 117–135). Dordrecht, the Netherlands: Springer.
Demir, M., & €
Ozdemir, M. (2010). Friendship, need satisfaction and happiness. Journal of
Happiness Studies,11, 243–259.
Demir, M., €
Ozen, A., & Do
gan, A. (2012). Friendship, perceived mattering and happiness: A study of
American and Turkish college students. Journal of Social Psychology,152, 659–664.
Demir, M., €
Ozen, A., Do
gan, A., Bilyk, N. A., & Tyrell, F. A. (2011). I matter to my friend,
therefore I am happy: Friendship, mattering, and happiness. Journal of Happiness Studies,
12, 983–1005.
Derrick, J. L., Gabriel, S., & Hugenberg, K. (2009). Social surrogacy: How favored television
programs provide the experience of belonging. Journal of Experimental Social Psychology,45,
352–362.
Diener, E. (2012). New findings and future directions for subjective well-being research. American
Psychologist,67, 590–597.
694 Norman P. Li and Satoshi Kanazawa
Diener, E., & Diener, M. (1995). Cross-cultural correlates of life satisfaction and self-esteem. Journal
of Personality and Social Psychology,68, 653–663.
Diener, E., Kanazawa, S., Suh, E., & Oishi, S. (2015). Why people are in a generally good mood.
Personality and Social Psychology Review,19, 235–256.
Diener, E., & Seligman, M. E. P. (2004). Beyond money: Toward an economy of well-being.
Psychological Science in the Public Interest,5,1–31.
Dolan, P., Peasgood, T., & White, M. (2008). Do we really know what makes us happy? A review of
the economic literature on the factors associated with subjective well-being. Journal of
Economic Psychology,29,94–122.
Dunbar, R. I. M. (1992). Neocortex size as a constraint on group size in primates. Journal of Human
Evolution,20, 469–493.
Dunbar, R. I. M. (1993). Coevolution of neocortical size, group size and language in humans.
Behavioral and Brain Sciences,16, 681–735.
Easterlin, R. A., Angelescu, L., & Zweig, J. S. (2011). The impact of modern economic growth on
urban–rural differences in subjective well-being. World Development,39, 2187–2198.
Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. (2003). Does rejection hurt? An fMRI study of
social exclusion. Science,302, 290–292.
Ellis, B. J., Figueredo, A. J., Brumbach, B. H., & Schlomer, G. L. (2009). Fundamental dimensions of
environmental risk: The impact of harsh versus unpredictable environments on the evolution
and development of life history strategies. Human Nature,20, 204–268.
Evans, R. J. (2009). A comparison of rural and urban older adults in Iowa on specific markers of
successful aging. Journal of Gerontological Social Work,52, 423–438.
Figueredo, A. J., V
asquez, G., Brumbach, B. H., Schneider, S. M. R., Sefcek, J. A., Tal, I. R., & Jacobs, W.
J. (2006). Consilience and life history theory: From genes to brain to reproductive strategy.
Developmental Review,26, 243–275.
Gardner, W. L., & Knowles, M. L. (2008). Love makes you real: Favorite television characters are
perceived as “real” in a social facilitation paradigm. Social Cognition,26, 156–168.
Gautney, J. R., & Holliday, T. W. (2015). New estimations of habitable land area and human
population size at the Last Glacial Maximum. Journal of Archeological Science,58, 103–112.
Geary, D. C., Byrd-Craven, J., Haord, M. K., Vigil, J., & Numtee, C. (2003). Evolution and development
of boys’ social behavior. Developmental Review,23, 444–470.
Gruter, M., & Masters, R. D. (1986). Ostracism as a social and biological phenomenon: An
introduction. Ethology and Sociobiology,7, 149–158.
Gutierres, S. E., Kenrick, D. T., & Partch, J. J. (1999). Beauty, dominance, and the mating game:
Contrast effects in self-assessment reflect gender differences in mate selection. Personality and
Social Psychology Bulletin,25, 1126–1134.
Hagen, E. H., & Hammerstein, P. (2006). Game theory and human evolution: A critique of some
recent interpretations of experimental games. Theoretical Population Biology,69, 339–348.
Hammen, C., & Brennan, P. A. (2002). Interpersonal dysfunction in depressed women: Impairments
independent of depressive symptoms. Journal of Affective Disorders,72, 145–156.
Headey, B., & Wearing, A. (1989). Personality, life events, and subjective well-being: Toward a
dynamic equilibrium model. Journal of Personality and Social Psychology,57, 731–739.
Heintzelman, S. J., & King, L. A. (2014). Life is pretty meaningful. American Psychologist,69,
561–574.
Hill, R. A., & Dunbar, R. I. M. (2003). Social network size in humans. Human Nature,14,53–72.
Hill, K., & Hurtado, A. M. (1996). Ache life history. Hawthorne, NY: Aldine.
Hill, S. E., & Major, B. (2013). An evolutionary psychological perspective on happiness. In I.
Boniwell, S. A. David & A. Ayers (Eds.), Oxford handbook of happiness (pp. 875–886). Oxford,
UK: Oxford University Press.
Hintze, A., Olson, R. S., Adami, C., & Hertwig, R. (2013). Risk aversion as an evolutionary adaptation.
arXiv:1310.6338v1 [q-bio.PE].
Hrdy, S. B. (2009). Mothers and others: The evolutionary origins of mutual understanding.
Cambridge, UK: Harvard University Press.
The savanna theory of happiness 695
Hruschka, D. J. (2010). Friendship: Development, ecology, and evolution of a relationship.
Berkeley, CA: University of California Press.
Huang, M.-H., & Hauser, R. M. (1998). Trends in black-white test-score differentials: II. The
WORDSUM vocabulary test. In U. Neisser (ed.), The rising curve: Long-term gains in IQ and
related measures (pp. 303–332). Washington, DC: American Psychological Association.
Hurd, J. P. (1985). Sex differences in mate choice among the ‘Nebraska Amish of central
Pennsylvania’. Ethology and Sociobiology,6,49–57.
Indik, B. P. (1965). Organization size and member participation: Some empirical tests of alternative
explanations. Human Relations,18, 339–350.
Johnson, M. P., & Leslie, L. (1982). Couple involvement and network structure: A test of the dyadic
withdrawal hypothesis. Social Psychology Quarterly,45,34–43.
Kanazawa, S. (2001). De gustibus est disputandum. Social Forces,79, 1131–1163.
Kanazawa, S. (2002). Bowling with our imaginary friends. Evolution and Human Behavior,23,
167–171.
Kanazawa, S. (2004a). General intelligence as a domain-specific adaptation. Psychological Review,
111, 512–523.
Kanazawa, S. (2004b). The Savanna Principle. Managerial and Decision Economics,25,41–54.
Kanazawa, S. (2010). Evolutionary psychology and intelligence research. American Psychologist,
65, 279–289.
Kanazawa, S., & Fontaine, L. (2013). Intelligent people defect more in a one-shot prisoner’s dilemma
game. Journal of Neuroscience, Psychology, and Economics,6, 201–213.
Kanazawa, S., & Li, N. P. (2015). Happiness in modern society: Why intelligence and ethnic
composition matter. Journal of Research in Personality,59, 111–120.
Killworth, P. D., Bernard, H. R., & McCarty, C. (1984). Measuring patterns of acquaintanceship.
Current Anthropology,25, 381–397.
Kim, T.-H., Jeong, G.-W., Baek, H.-S., Kim, G.-W., Sundaram, T., Kang, H.-K., ... Song, J.-K. (2010).
Human brain activation in response to visual stimulation with rural and urban scenery pictures: A
functional magnetic resonance imaging study. Science of the Total Environment,408, 2600–
2607.
Knight, J., & Gunatilaka, R. (2010). The rural–urban divide in China: Income but not happiness?
Journal of Development Studies,46, 506–534.
Lam, D., & Levison, D. (1992). Age, experience, and schooling: Decomposing earnings inequality in
the United States and Brazil. Sociological Inquiry,62, 220–245.
Lederbogen, F., Kirsch, P., Haddad, L., Streit, F., Tost, H., Schuch, P., ... Meyer-Lindenberg, A.
(2011). City living and urban upbringing affect neural social stress processing in humans.
Nature,474, 498–501.
Lewis, D. M. G., Al-Shawaf, L., Russell, E. M., & Buss, D. M. (2015). Friends and happiness: An
evolutionary perspective on friendship. In M. Demir (Ed.), Friendship and happiness: Across
the life-span and cultures (pp. 37–57). Dordrecht, the Netherlands: Springer.
Li, T., & Cheng, S.-T. (2015). Family, friends, and subjective well-being: A comparison between the
west and Asia. In M. Demir (Ed.), Friendship and happiness: Across the life-span and cultures
(pp. 235–251). Dordrecht, Netherlands: Springer.
Li, N. P., Smith, A. R., Griskevicius, V., Cason, M. J., & Bryan, A. (2010). Intrasexual competition and
eating restriction in heterosexual and homosexual individuals. Evolution and Human
Behavior,31, 365–372.
MacDonald, C. B. (1955). Company. Encyclopedia Britannica (14th ed.). London, UK:
Encyclopedia Britannica.
Mange, A. P., & Mange, E. J. (1980). Genetics: Human aspects. New York, NY: Saunders.
Miao, F. F., Koo, M., & Oishi, S. (2013). Subjective well-being. In I. Boniwell, S. A. David & A. C. Ayers
(Eds.), The Oxford handbook of happiness (pp. 174–184). Oxford, UK: Oxford University Press.
Miner, J. B. (1957). Intelligence in the United States: A survey –with conclusions for manpower
utilization in education and employment. New York, NY: Springer.
696 Norman P. Li and Satoshi Kanazawa
Nesse, R. M. (2004). Natural selection and the elusiveness of happiness. Philosophical Transactions
of the Royal Society of London, Series B,359, 1333–1347.
Oates, J. (1977). Mesopotamian social organisation: Archaeological and philological evidence. In J.
Friedman & M. J. Rowlands (Eds.), The evolution of social systems (pp. 457–485). London, UK:
Duckworth.
Oishi, S., Graham, J., Kesebir, S., & Galinha, I. C. (2013). Concepts of happiness across time and
cultures. Personality and Social Psychology Bulletin,39, 559–577.
Pavot, W., & Diener, E. (2013). Happiness experienced: The science of subjective well-being. In I.
Boniwell, S. A. David & A. C. Ayers (Eds.), The Oxford handbook of happiness (pp. 134–151).
Oxford, UK: Oxford University Press.
Pinker, S. (2014). The village effect: Why face-to-face contact matters. New York, NY: Random
House.
Pollan, M. (2006). The omnivore’s dilemma: A natural history of four meals. London, UK: Penguin.
Porter, L. W., & Lawler, E. E. III (1965). Properties of organization structure in relation to job
attitudes and job behavior. Psychological Bulletin,64,23–51.
Russell, V. M., McNulty, J. K., Baker, L. R., & Meltzer, A. L. (2014). The association between
discontinuing hormonal contraceptives and wives’ marital satisfaction depends on husbands’
facial attractiveness. Proceedings of the National Academy of Sciences,111, 17081–17086.
Sally, D. (1995). Conversation and cooperation in social dilemmas: A meta-analysis of experiments
from 1958 to 1992. Rationality and Society,7,58–92.
Schilling, O., & Wahl, H. W. (2002). Family networks and life-satisfaction of older adults in rural and
urban regions. Kolner Zeitschrift f €
ur Soziologie und Sozialpsychologie,54, 304–317.
Smith, A. (1981). An inquiry into the nature and causes of the wealth of nations. R. H. Campbell &
A. S. Skinner (Eds.), Indianapolis, IN: Liberty Fund. (Original work published 1776.)
Smith, T. W., Marsden, P. V., & Hout, M. (2015). General social surveys, 1972–2014: Cumulative
codebook. Chicago, IL: National Opinion Research Center.
Smuts, B. B. (1985). Sex and friendship in baboons. New York, NY: Aldine.
Tooby, J., & Cosmides, L. (1990). The past explains the present: Emotional adaptations and the
structure of ancestral environments. Ethology and Sociobiology,11, 375–424.
van Beest, I., & Williams, K. D. (2006). When inclusion costs and ostracism pays, ostracism still hurts.
Journal of Personality and Social Psychology,91, 918–928.
van der Wal, A. J., Schade, H. M., Krabbendam, L., & van Vugt, M. (2013). Do natural landscapes
reduce future discounting in humans? Proceedings of the Royal Society of London, Series B,
280, 20132295.
van Vugt, M., & Ronay, R. D. (2014). The evolutionary psychology of leadership: Theory, review, and
roadmap. Organizational Psychology Review,4,74–95.
Williams, K. D., Forgas, J. P., & von Hippel, W. (Eds.) (2005). The social outcast: Ostracism, social
exclusion, rejection, and bullying. New York, NY: Psychology Press.
Wirth, L. (1938). Urbanism as a way of life. American Journal of Sociology,44,1–24.
Wolfle, L. M. (1980). The enduring effects of education on verbal skills. Sociology of Education,53,
104–114.
Received 25 May 2015; revised version received 13 December 2015
The savanna theory of happiness 697