On the evolution of personality –
Insights from contemporary Western humans
Venla Berg
Institute of Behavioural Sciences
University of Helsinki, Finland
Academic dissertation to be publically discussed,
by due permission of the Faculty of Behavioural Sciences,
at the University of Helsinki in Auditorium 5, Fabianinkatu 33,
on the 14th of October 2016, at 12 o’clock
University of Helsinki
Institute of Behavioural Sciences
Studies in Psychology 121: 2016
Supervisors
Associate Professor Markus Jokela
Institute of Behavioural Sciences, University of Helsinki, Helsinki, Finland
Research Professor Anna Rotkirch
Population Research Institute, Väestöliitto, Helsinki, Finland
Reviewers
Professor Lars Penke
Biological Personality Psychology, Georg August University Göttingen, Göttingen,
Germany
Assistant Professor Marco Del Giudice
Department of Psychology, University of New Mexico, Albuquerque, New Mexico, USA
Opponent
Professor Daniel Nettle
Centre for Behaviour & Evolution, Newcastle University, Newcastle, UK
ISSN-L 1798-842X
ISSN 1798-842X
ISBN 978-951-51-2489-0 (pbk.)
ISBN 978-951-51-2490-6 (PDF)
http://www.ethesis.helsinki.fi
Helsinki University Print
Helsinki 2016
Contents
Abstract………………………………………………………………….………………………………………………5
Tiivistelmä……………………………………………………………………………………………………………..6
Acknowledgements……………………………………………………………………………….……………...7
Abbreviations…………………………………………………………………………………………..……………9
List of original publications.............................................................................................9
1 Introduction…………………………………………………………………….…………….………………….10
1.1 Personality and fitness……………………………………………….………………………..……..…..14
1.2 Genetics of personality…………………………………………………………………………....……..18
1.3 Theories on the evolution of personality………………………………………..……………..….23
1.3.1 Selective neutrality……………………………………………………………………...…………..24
1.3.2 Mutation–selection balance……………………………………………………………………..25
1.3.3 Balancing selection………………………………………………………………………………….27
2 Research aims and scope…………………………………………………………………………………30
2.1 Is personality phenotypically associated with lifetime reproductive
success?.............................................................................................................................31
2.2 Are the phenotypic and genetic associations between personality and
lifetime reproductive success different?..........................................................................32
2.3 Is personality involved in a reproductive quality–quantity trade-off?.....................34
2.4 Are there novel selection pressures on personality in contemporary
humans?...........................................................................................................................36
3 Methods……………………………………………………………………………………..………………..…..38
3.1 Participants………………………………………………………………………………………………..…38
3.2 Measures……………………………………………………………………………………………………...40
3.3 Statistical analyses………………………………………………………………………………………...43
3.3.1 Statistical methods for sub-study I…………………………………………………………...43
3.3.2 Statistical methods for sub-study II………………………………………………………….45
3.3.3 Statistical methods for sub-study III………………………………………………………...45
4 Results………………………………………………………………………………………….……………..……46
4.1 Personality and lifetime reproductive success in contemporary
humans…………………………………………………………………………..…………………….……………46
4.2 Phenotypic and genetic assessment of response to
selection (sub-study I)…………………………………………………………………………..…………….48
3
4.3 Reproductive quality–quantity trade-offs and
personality (sub-study II)…………………………………………………………………………………….50
4.4 Personality and planned and non-planned
pregnancies (sub-study III)……………………………………………………….…………………………52
5 Discussion…………………………………………………………………………………………………………56
5.1 Personality, reproductive success, and fluctuating environments………………………..56
5.2 Personality and the phenotypic gambit…………………………………………………………….59
5.3 Personality and reproductive life-history trade-offs……………………………...…………..62
5.4 Temporal changes in the selection pressures on personality……………………………...63
5.5 Methodological considerations…………………………………………………………………….….66
5.6 Conclusions and theoretical implications………………………………………………………...69
6 References…………………………………………………………………………………………………..……75
4
Abstract
Recently, personality variation has been found to be associated with number of
offspring, pace of reproduction, and other components of fitness in humans and
other animals. The link between personality and fitness has inspired a surge of
theoretical accounts on the evolution of personality. However, the associations
between personality and fitness are yet to be sufficiently empirically elaborated for
the theoretical advances to be solidly based. This thesis explores the pathways
between personality and fertility in three contemporary Western human
populations.
Specifically, this thesis i) provides further evidence on the phenotypic associations
between personality and lifetime reproductive success; ii) examines the differences
between a phenotypic and genetic approach to natural selection on personality; iii)
investigates whether personality is involved in a trade-off between the number and
“quality” of offspring; and iv) explores possible novel selection pressures on
personality. The results are discussed in relation to the theoretical accounts on the
evolution of personality.
The results suggest that the phenotypic associations between personality and
lifetime reproductive success vary by populations. Further, it seems that the
phenotypic and genetic approaches to natural selection on personality differ and that
modern environments can impose novel selection pressures on personality. Lastly,
it seems that personality is similarly associated with number of children and
grandchildren, implying no quality–quantity trade-offs in modern environments.
Crucially, the results show that personality is not associated with fitness in any
one way, or because of any one evolutionary mechanism. Further, the results suggest
that the behavioural personality differences might not be evolutionarily relevant, but
that personality correlates genetically with something else that affects fitness. Thus,
future studies should concentrate on the individual traits and their genetic
correlates, as trying to find a unified evolutionary answer for “personality” may prove
an elusive task.
5
Tiivistelmä
Persoonallisuuden vaihtelun on viimeaikoina havaittu olevan yhteydessä jälkeläisten lukumäärään ja muihin kelpoisuuden tekijöihin niin ihmisillä kuin muillakin
eläimillä. Persoonallisuuden ja kelpoisuuden välinen yhteys on synnyttänyt paljon
teoreettista tutkimusta, jossa pohditaan persoonallisuuden evoluutiota. Tämä
teoreettinen pohdinta tarvitsee tuekseen kuitenkin yksityiskohtaisempaa empiiristä
tietoa, jota on tällä hetkellä niukasti saatavilla. Tämä väitöskirjatutkimus tarkastelee
ja tarkentaa sitä, miksi ja millaisilla tavoilla persoonallisuus ja lasten lukumäärä ovat
yhteydessä. Osatutkimukset on toteutettu kolmella länsimaisella aikalaisaineistolla
Suomesta, Iso-Britanniasta ja Yhdysvalloista.
Tarkemmin keskitytään i) vertailemaan persoonallisuuden ja lasten lukumäärän
välisiä geneettisiä ja ilmiasun tason yhteyksiä; ii) tutkimaan liittyykö persoonallisuuteen vaihtokauppaa jälkeläisten määrän ja ”laadun” välillä; ja iii) tarkastelemaan,
luoko moderni länsimainen ympäristö uusia valintapaineita persoonallisuudelle.
Näiden tutkimuskysymysten tuloksia arvioidaan suhteessa persoonallisuuden
evoluutiota käsittelevään teoreettiseen kirjallisuuteen.
Persoonallisuuden ja syntyvyyden väliset ilmiasun tason yhteydet vaihtelivat osatutkimuksesta toiseen ja geneettisen tason yhteydet eivät olleet samanlaisia kuin
ilmiasun tason yhteydet. Näyttäisi myös siltä että nyky-ympäristöt luovat uudenlaisia valintapaineita persoonallisuudelle. Lisäksi persoonallisuus oli samalla tavalla
yhteydessä niin lasten kuin lastenlastenkin lukumäärään. Tämä viittaa siihen, että
persoonallisuuteen liittyen vaihtokauppaa lasten lukumäärän ja ”laadun” välillä ei
ole.
Tämän väitöskirjatutkimuksen perusteella vaikuttaa siltä, että eri persoonallisuuspiirteet ovat yhteydessä kelpoisuuteen hyvin eri tavoin. On lisäksi mahdollista,
että persoonallisuuserot eivät käyttäytymisen tasolla ole evolutiivisesti merkittäviä,
vaan persoonallisuus korreloi geneettisesti joidenkin muiden kelpoisuuteen vaikuttavien tekijöiden kanssa. Tutkimuksen tulisikin tulevaisuudessa keskittyä yksittäisiin persoonallisuuspiirteisiin ja niiden geneettiseen taustaan, sillä yhtä evolutiivista
selitystä persoonallisuudelle tuskin tullaan löytämään.
6
Acknowledgements
I am immensely grateful for the time, effort, thought, and material assets that several
individuals and institutions have allocated on this project of mine. First and
foremost, I would like to thank my supervisors Associate Professor Markus Jokela
and Research Professor Anna Rotkirch. I thank Markus for trusting in me, first by
inviting me to work on my Master’s thesis in his project, and then by continuing our
collaboration with my doctoral studies. With his deep understanding of statistical
methods, the academia, and the quirks of writing and publishing articles, he is an
inspiration. I wish to thank Anna for taking me under her wings at Population
Research Institute. Alongside with her brilliant scientific mind, I am so grateful for
her introducing me to her large academic network, and showing that research can be
a lot of fun, including mushroom picking and rowing boats in spectacularly beautiful
countryside.
I am grateful for the research facilities provided by the Institute of Behavioural
Sciences at the University of Helsinki, where I started my doctoral studies, and
Väestöliitto, Population Research Institute, where I have had the pleasure to work at
for the past couple of years. I would like especially to thank Väestöliitto, and its CEO
Eija Koivuranta, for providing me with such a stimulating environment to do
research. The financial help from Research Grants of the University of Helsinki and
the Academy of Finland is also gratefully acknowledged. This research would not
have been possible without the prior work of numerous anonymous academics who
have compiled the datasets used in this thesis. I am grateful to Jaakko Kaprio for
giving me the permission to work on the Finnish Twin Cohort, and the research
organisations responsible for collecting and keeping up the HRS and NCDS datasets.
I would like to express my deepest gratitude to Professor Lars Penke and Associate
Professor Marco Del Giudice for reviewing this thesis, for giving me valuable
feedback, and also for their encouraging words. I also wish to warmly thank
Professor Daniel Nettle for agreeing to act as the opponent. For the scientific articles
that are part of the present thesis, I own great debt of gratitude to all my co-authors.
I wish to thank Markus Jokela, Anna Rotkirch, Virpi Lummaa, Mirkka Lahdenperä,
7
Heini Väisänen, Ian J. Rickard, Karri Silventoinen, and Jaakko Kaprio for their
insightful and invaluable input. I also greatly appreciate them having borne with me
through the almost unbearably strenuous journeys to get our multidisciplinary
papers published. Especially, I would like to thank Virpi for her straightforward,
honest, yet kind feedback and support she has given me. I truly believe I am now a
better researcher because of what she has taught me. And, I wish to thank her and
Anna for showing me that female scientists really rock.
I am grateful to my current and former colleagues at Population Research
Institute in Väestöliitto, who – after some rather lonely years at the university – gave
me an academic home. Anneli Miettinen, Lassi Lainiala, Tiina Helamaa, Osmo
Kontula, Minna Säävälä, Eevi Lappalainen, and Michael Laakasuo, among others, I
would like to thank you for making getting up in the morning and going to work every
day such a delightful task. In addition, this sometimes-bumpy ride to become a PhD
has greatly been smoothed by my dearest fellow-doctoral-student friends, Heta
Moustgaard, Kanerva Lahdensuo, and Kristiina Tammisalo, and my equally dear
non-academic friends Jenni Kokander and Anna Meling. I am also very thankful for
the peer support provided by Paula Virtala and Christian Hakulinen during our
longish loosely scientific luncheons throughout these years.
This thesis would not have been finished without the support from my childhood
and current families. Mom, dad, thank you for patiently watching me, without a hint
of frustration, go through multiple faculties and several majors before finding solid
ground on psychology. My odyssey was what it took me to be the happy scientist I
am today. Finally, I wish to thank the two single most wonderful creatures in the
world, my partner Ari Helo and our son Helmeri, for being there and carrying me
through this process. I am sorry I gave you such a hard time. You are my everything.
Helsinki, September 2016
Venla Berg
8
Abbreviations
FFM
Five-Factor Model of personality
SES
Socio-economic status
GWAS
Genome wide association study
GCTA
Genome-wide complex trait analysis
List of original publications
I
Berg, V., Lummaa, V., Rickard, I.J., Silventoinen, K., Kaprio, J., and Jokela,
M. (2016). Genetic associations between personality traits and lifetime
reproductive success in humans. Behavior Genetics. Published online before
print. DOI:10.1007/s10519-016-9803-5
II
Berg, V., Lummaa, V., Lahdenperä, M., Rotkirch, A., & Jokela, M. (2014).
Personality and long-term reproductive success measured by the number of
grandchildren. Evolution and Human Behavior, 35, 533-539.
III
Berg, V., Rotkirch, A., Väisänen, H., & Jokela, M. (2013). Personality is
differentially associated with planned and non-planned pregnancies.
Journal of Research in Personality, 47, 296-305.
The articles are reprinted with the kind permission of the copyright holders, and
they are referred to in the text by their roman numerals.
9
1 Introduction
Personality is a central concept in human psychology. Human beings seem to be
innately different in the ways we are predisposed to act, think and feel across
different situations and time, and these behavioural tendencies are called
personality. Personality is known to be associated with many aspects of human life,
including mental and physical well-being (e.g., Kendler, Gatz, Gardner, & Pedersen,
2006), or educational aspirations (e.g., Poropat, 2009). Recently, personality
variation has also been found to be associated with reproductive behaviour, such as
number of offspring and pace of childbearing. The link between personality variation
and reproduction is likely to shape human evolution, as differential amounts of genes
related to personality variation are passed on to the next generations. Similar
findings on associations between personality variation and fitness in other, nonhuman animals, have inspired a surge of theoretical accounts explaining and
contemplating the evolution of personality. However, the empirical evidence on the
associations between personality and reproductive behaviour are yet to be
sufficiently elaborated for the theoretical advances to be solidly based. The present
thesis is the first study to explore the pathways of how personality is related to
fertility in contemporary populations in developed Western societies. Identifying
pathways between personality and reproduction provides information vital to
theoretical accounts on the evolution of personality.
During the 1990s, after decades of research, personality psychology converged
into a loose consensus of two to five higher order factors, the exact amount and
content of which vary somewhat between different theories (e.g., Bouchard &
Loehlin, 2001; Digman, 1990; Stelmack, 1991). The so called Five Factor Model
(FFM) of Costa & McCrae (Costa & McCrae, 1992) is perhaps the most used
personality model. It comprises the personality traits extraversion, neuroticism
agreeableness, conscientiousness, and openness to experience (sometimes referred
to as intellect). The five factor structure of human personality seems fairly stable
across different populations and cultures (McCrae, 2002; Rolland, 2002; Yamagata
10
et al., 2006), although not universally so (e.g., Gurven, von Rueden, Massenkoff,
Kaplan, & Lero Vie, 2013). Personality is heritable, meaning that the variation in
personality across individuals is partly related to the genetic variation across
individuals (Vukasovic & Bratko, 2015). In addition, personality traits in humans are
meaningfully associated with various outcomes. For example, neuroticism is
associated with an increased risk of major depression (Kendler et al., 2006; Malouff,
Thorsteinsson, & Schutte, 2005), and conscientiousness and openness to experience
with academic achievement in both children and young adults (Poropat, 2009;
Poropat, 2014; Richardson, Abraham, & Bond, 2012; Spengler, Luedtke, Martin, &
Brunner, 2013). In other words, personality traits are not mere statistical or
theoretical concepts, but have predictive validity for behaviour and development.
Consistent individual variation in behavioural tendencies in other, non-human
animals has been a topic in biological behavioural sciences for a long time (reviewed
in Fuller, 1960). In behavioural ecology, however, individual variation in behaviour
was for a long time assumed to be spurious noise around the species-typical adaptive
patterns (see e.g., Nettle & Penke, 2010). Recent years have, nevertheless, seen a
rapid growth of interest in this consistent individual behavioural variation in the
biological sciences (van Oers & Sinn, 2011). By now, it seems clear that it is a
ubiquitous phenomenon, present in a wide variety of taxa, not just in humans or
other primates (e.g., Massen, Antonides, Arnold, Bionda, & Koski, 2013), not just in
mammals (e.g., Dingemanse & Réale, 2005), but in birds (Dingemanse, Both, Drent,
& Tinbergen, 2004) and fish (Quinn, Patrick, Bouwhuis, Wilkin, & Sheldon, 2009),
and possibly not even just in vertebrates (for personality in field crickets, see
DiRienzo, Niemelä, Skog, Vainikka, & Kortet, 2015). Compared to human
psychological personality research, animal personality research is taking its first
steps in how to reliably measure, classify, and validate personality (Carter, Feeney,
Marshall, Cowlishaw, & Heinsohn, 2013; Koski, 2011). In non-humans animals,
consensus is yet to be reached even on what to call this phenomenon, and various
terms such as coping styles, temperament (Réale, Reader, Sol, McDougall, &
Dingemanse, 2007), and behavioural syndromes (Sih, Bell, Johnson, & Ziemba,
11
2004) have been proposed. In this thesis I will use the term “personality” for both
humans and other animals. Importantly, the biological take on personality has
introduced new insights into studying it. The focus has turned into the evolution of
personality – the key question being, why does heritable personality variation exist?
A pivotal concept of evolutionary theory is fitness, i.e., the relative proportion of
genes one individual passes on to the next generation (e.g., Stearns, 1986). All else
being equal, natural selection should deplete a trait that is detrimental to fitness
from any genetic variation (Falconer & Mackay, 1996; Merilä & Sheldon, 1999). This
means that genes that affect fitness negatively are passed on in fewer and fewer
amounts, and eventually disappear from the genetic pool entirely. The main concern
in the origins and maintenance of genetic variation in personality is therefore finding
out whether personality has fitness consequences. Empirical studies on non-human
personality have focused on just that. The associations between personality and
different components of fitness, such as survival, reproductive success, or parental
care, have recently been examined in many different species (e.g., Dingemanse et al.,
2004; Hollander, Van Overveld, Tokka, & Matthysen, 2008; Sinn, Apiolaza, &
Moltschaniwskyj, 2006). Lately, the fitness correlates of human personality have
also been gaining attention. There is a body of research showing associations
between personality and mortality (e.g., Jokela, Pulkki-Råback, Elovainio, &
Kivimäki, 2014; Lee, Wadsworth, & Hotopf, 2006; Shipley, Weiss, Der, Taylor, &
Deary, 2007; Turiano, Chapman, Gruenewald, & Mroczek, 2015). The present thesis
focuses on another component of fitness, lifetime reproductive success, i.e., the total
number of offspring born to an individual during reproductive lifespan. Lifetime
reproductive success also shows associations with personality traits in humans (see
section 1.1.), and is the most important factor defining fitness in contemporary
industrialised societies with low mortality (Falconer & Mackay, 1996; but see Jones
& Bird, 2014).
The purpose of the present thesis is twofold. Firstly, as previous research has
established associations between personality and reproductive success, I attempt to
shed light on the mechanisms underlying these associations. The primary research
12
question can be formulated as why and how personality and reproductive success
are related. Although it is beyond the scope of this thesis to answer this question
comprehensively, it has inspired the specific research questions at hand:
i)
Are personality traits and lifetime reproductive success phenotypically
associated? This thesis provides further evidence on this subject matter
with large, representative datasets from contemporary Western
populations.
ii)
Are personality traits and lifetime reproductive success genetically
associated? Recent advances in the field of quantitative population
genetics have shown that concentrating on the phenotypic level is not
sufficient for making inferences on evolutionary changes. Sub-study I
provides the first results on the associations between personality and
fitness on a genetic level.
iii)
Is personality involved in a reproductive quality–quantity trade-off
between the number and “quality” of offspring? Sub-study II examines
this question by looking at the associations between personality and
reproduction over two generations.
iv)
Do the associations between personality and reproduction have features
that are unique to contemporary Western societies? Reproductive
behaviour may be very differently associated with personality in
societies that have undergone the demographic transition (Lee, 2003)
than in societies with high fertility and high mortality. Sub-study III
tackles this question by examining the associations between personality
traits and planned and non-planned pregnancies in the contemporary
UK.
The second purpose of this thesis is related to the fact that the researchers
studying human personality and the researchers studying personality in non-human
animals do not seem to communicate very effectively (Gosling, 2001; Koski, 2011;
Nettle & Penke, 2010). Psychological personality research tends to concentrate on
the structure and correlates of personality traits. In contrast, the focus in biological
13
sciences is on the evolution of this individual behavioural variation. In consequence,
both fields miss out on the potential that a different point of view could provide. In
order to advance co-operation between these two research fields, I incorporate
research from both traditions in the present thesis.
The next sections of this Introduction outline the empirical and theoretical
background for the thesis. Firstly, I review existing research concerning the
associations between personality and different fitness components, focusing on
fertility (section 1.1.). I then briefly summarise the research on the genetics of
personality – that is, the proximate mechanisms causing an individual to have
certain behavioural tendencies (section 1.2.). In section 1.3, I will introduce some
theories on the ultimate mechanisms of personality, i.e., theoretical suggestions
about the origins and evolution of personality. Chapter 2 turns in more detail to the
aims and research questions of the present thesis.
1.1 Personality and fitness
Personality seems to be associated with many different components of fitness (e.g.,
Jokela et al., 2014). The present thesis focuses on the associations between
personality and fertility, and especially, lifetime reproductive success. Recently, the
associations between personality variation and fertility in humans have been
examined in a dozen of studies. Most of the studies have been conducted on
contemporary, Western, industrialised societies, but some results concern highmortality and high-fertility populations as well. Overall, the associations between
personality and reproductive outcomes are rather weak, and typically, personality
variation explains only a few percent of the variation in number of children. In
addition, the findings are inconsistent through the studies. However, even the
slightest fitness effects are important and can have large consequences in the
evolutionary timescale (Penke, Denissen, & Miller, 2007). Further, the inconsistency
of the associations between personality and reproductive success may actually
provide essential information on the evolution of personality (see section 1.3.). It is
of importance to note, however, that empirical studies on the associations between
14
personality and fitness in humans seldom make explicit inferences about the
evolutionary forces working on personality. With few exceptions (see, e.g., Del
Giudice, 2012; Penke et al., 2007), the theoretical advances considering the
evolution of personality largely come from outside the psychology departments.
Regarding personality variation and fertility, the most consistent finding seems to
be that higher extraversion, and related traits such as leadership and sociability, are
associated with higher probability of parenthood and having subsequent children in
both sexes – but even more consistently in men. This has been observed in
contemporary Western samples (Dijkstra & Barelds, 2009; Jokela, Alvergne, Pollet,
& Lummaa, 2011; Jokela & Keltikangas-Järvinen, 2009; Jokela, Kivimäki, Elovainio,
& Keltikangas-Järvinen, 2009; Skirbekk & Blekesaune, 2014), as well as in highfertility, high-mortality populations in developing countries (Alvergne, Jokela, &
Lummaa, 2010; Bailey et al., 2013; Gurven, von Rueden, Stieglitz, Kaplan, & Eid
Rodriguez, 2014). Some studies, however, have not found such an association,
especially in women (Alvergne et al., 2010; Bailey et al., 2013; Eaves, Martin, Heath,
Hewitt, & Neale, 1990; Gurven et al., 2014; Mealey & Segal, 1993; Nettle, 2005;
Perkins et al., 2013; Skirbekk & Blekesaune, 2014). These discrepancies may be due
to lack of power in smaller samples, different personality measures, or some real
differing effects of extraversion in different environments.
Higher neuroticism and related traits, such as harm avoidance and negative
emotionality, have also quite consistently been associated with reproductive success,
namely, with lower probability of parenthood and fewer number of children in
contemporary Western samples (Jokela et al., 2011; Jokela et al., 2009; Jokela,
Hintsa, Hintsanen, & Keltikangas-Järvinen, 2010; Reis, Doernte, & von der Lippe,
2011; Skirbekk & Blekesaune, 2014) and in one sample of Bolivian Tsimane men
(Gurven et al., 2014). This association has not, however, been observed in all studies
(Bailey et al., 2013; Dijkstra & Barelds, 2009; Perkins et al., 2013), and one study in
rural Senegal found neuroticism to increase rather than decrease the number of
children in women (Alvergne et al., 2010).
15
The other three traits of the Five Factor Model of personality have been studied
less, and the results have been more inconsistent. Higher agreeableness (and a
comparable trait, reward dependence) has been shown to correlate with higher
number of children, especially in women (Dijkstra & Barelds, 2009; Jokela et al.,
2011; Jokela et al., 2010), but many studies do not report associations between
agreeableness and reproductive success (Bailey et al., 2013; Gurven et al., 2014;
Perkins et al., 2013; Skirbekk & Blekesaune, 2014). Higher conscientiousness, and
its temperamental counterpart persistence, have been found to be associated with
both higher (Dijkstra & Barelds, 2009; Gurven et al., 2014) and lower number of
children, especially in women (Jokela et al., 2011; Jokela et al., 2010; Perkins et al.,
2013; Skirbekk & Blekesaune, 2014), suggesting that this personality dimension may
involve both positive and negative associations with fertility depending on the
population. Finally, higher openness to experience was not associated with
reproductive success in Senegalese men (Alvergne et al., 2010) or in Dutch and
Norwegian women (Dijkstra & Barelds, 2009; Skirbekk & Blekesaune, 2014). It was,
however, associated with lower number of children in contemporary Norwegian
men, Australian women, and in both sexes in two contemporary American samples
(Jokela et al., 2011; Perkins et al., 2013; Skirbekk & Blekesaune, 2014), and
conversely, with higher number of children in Tsimane men (Gurven et al., 2014). In
addition, one study on an American sample found that the associations between
openness to experience and conscientiousness and number of children only became
evident in cohorts born in the latter part of the 20th century (Jokela, 2012). It is
possible that some aspects of the contemporary Western environments, such as
women’s increasing labour participation, moderate the fitness consequences of at
least some personality traits.
Apart from one early finding (Eaves et al., 1990), all studies report linear
associations between personality and reproductive success in humans. Linear
associations between a trait and fitness imply directional natural selection, whereas
curvilinear associations could be signs of stabilising or disruptive selection.
Stabilising selection is implied if intermediate personality levels are associated with
16
greatest fitness, and disruptive selection if both ends of a personality scale are
associated with greatest fitness (see section 1.3.). However, in a study conducted in
rural Senegal (a high-fertility, high-mortality population), high neuroticism in
women in low socio-economic position increased the number of children but
decreased their physical well-being. This led to the prediction that highest fitness
would be associated with intermediate levels of neuroticism (Alvergne et al., 2010).
This same pattern was not evident among women in a higher socio-economic
position. In sum, although several findings on personality and reproductive success
have been documented in humans, the specific fitness consequences of different
personality traits are yet unclear. The differences seem to be at least partly associated
with the instruments used to measure personality traits, as well as the characteristics
of the study population. In addition, the associations between personality and
fertility are not strong, and seem to require a large sample in order to be detected.
In biological sciences, there is an innate tendency to view animal behaviour as an
outcome of evolution. As a consequence, personality in non-human animals has been
studied in relation to its fitness associations throughout the surge of interest in
animal personality. Based on their results, studies on animal personality often make
explicit inferences on the possible mechanisms maintaining genetic variance in
personality. A meta-analysis comprising studies on mammals, birds, fish, and some
insects, concluded that “boldness” increases reproductive success but shortens lifespan, that “exploration” has no effect on reproductive success but a small positive
effect on survival, and that “aggression” has a small positive effect on reproductive
success and a positive but non-significant trend on survival especially in females
(Smith & Blumstein, 2008). The results varied somewhat by sex and by whether the
studied species was captive or domesticated, or wild.
The results of the meta-analysis (Smith & Blumstein, 2008) suggest that
personality might be affecting different components of fitness antagonistically. For
example, boldness seems to increase reproductive success while it is detrimental to
survival. In consequence, the net effect of boldness on fitness could be neutral, and
personality variation in boldness could be maintained through such a trade-off. The
17
maintenance of genetic variation in traits that seem to be under directional selection
remains a puzzle (see section 1.3. on evolutionary mechanisms that could maintain
genetic variation). However, many studies have also found varying associations
between personality traits and fitness with changing environments, for example,
variation in predation risks (Réale & Festa-Bianchet, 2003), food availability
(Dingemanse et al., 2004), or in habitat complexity (Höjesjö, Johnsson, & Bohlin,
2004). Such studies indicate that heritable variation in personality may be
maintained through environmental heterogeneity, with fitness effects varying
through time or environments (Penke et al., 2007; see also section 1.3.). Finally, as
would be expected, not all studies on animal personality have found associations
with any components of fitness (e.g., Brent et al., 2014).
1.2 Genetics of personality
On average, two randomly selected humans share approximately 99.9 percent of
their DNA, leaving only 0.1 per cent varying between individuals (NHGRI, 2005).
The genes that do show genetic variability are called segregating genes, and the loci
that show variation, polymorphic. The genetic sciences are involved in explaining
how the genetic polymorphism is associated with phenotypic individual variation.
Many clinically relevant traits, such as Huntington’s disease, have been associated
with a single gene, and with a certain polymorphism (e.g., MacDonald et al., 1993).
Such traits are often binary – they are either present or not. However, many other
physical or psychological traits, such as height or personality, are quantitative: they
are approximately normally distributed in populations and therefore assumed to be
affected by polymorphic variation in several genes. The genetics of quantitative traits
are studied, in general, from two distinct perspectives. Quantitative genetics use
statistical techniques that employ information on the relatedness between
individuals in a population in order to examine how much of the phenotypic
variation in a trait can be accounted for by the variation in the segregating genes (i.e.,
heritability). By contrast, molecular genetics investigate phenomena on the level of
the genes. Candidate gene linkage studies look for associations between a certain
18
trait and previously identified genes that could potentially be linked to that trait.
Genome wide association studies (GWAS), in turn, scan through the whole genome
to find any genes, without prior assumptions, that are associated with a certain trait.
Most recently, genome-wide complex trait analyses (GCTA) examine how much of
the observed phenotypic variation in a trait can be accounted for by observed
common genetic polymorphism in general (so called SNP-based heritability).
Over the course of the past five decades or so, numerous quantitative genetic
studies have studied the heritability of personality in humans. A recent metaanalysis concluded that human personality traits have an average heritability of .39
(Vukasovic & Bratko, 2015). This means that approximately 40 per cent of the
phenotypic personality variation between individuals is accounted for by genetic
variation between individuals. The meta-analysis showed that twin studies on
average yield significantly higher estimates of heritability (.47) compared to family
or adoption studies (.22) (Vukasovic & Bratko, 2015). The methodology of classical
twin studies results in so called broad-sense heritability estimates, which means that
it includes both additive and non-additive genetic effects. Additive genetic effects
refer to polymorphisms whose effect on the phenotype only depends on that
polymorphism. Non-additive genetic variation refers to situations where
polymorphisms in one locus (dominance) or several loci (epistasis) interact to
produce phenotypes. Family or adoption studies, on the contrary, yield narrow-sense
heritability estimates including additive genetic variation only. The authors
hypothesise that the difference between the estimates from family or adoption
studies and twin studies is indicative of personality having a lot of non-additive
genetic variation.
Another meta-analysis on the heritability of extraversion and neuroticism with
twin data similarly concluded the broad sense heritabilities to be .48 and .49,
respectively (van den Berg et al., 2014). This study was conducted on classical twin
designs, the power of which to detect non-additive effects is generally low (Posthuma
& Boomsma, 2000). However, the large meta-analytic sample size allowed for
testing non-additive genetic effects as well. In accordance with the meta-analysis by
19
Vukasovic and Bratko (2015), it was concluded that approximately half of the genetic
variation in extraversion and neuroticism is non-additive in nature (van den Berg et
al., 2014).
In conclusion, on the basis of quantitative genetic studies it can be stated that
personality variation in humans is moderately driven by genetic variation and that
at least part of this genetic variation is non-additive in nature. Natural selection is
more effective in depleting additive genetic variation than non-additive genetic
variation (Merilä & Sheldon, 1999). Thus, the evolutionary changes in personality
are probably considerably slower than what could have been expected on the basis
of classical twin studies.
Quantitative genetic studies on non-human animals reporting the heritability of
personality are more difficult to summarise because of the variety in ways of
conceptualising and measuring personality traits. Heritability estimates of almost
everything from 0.00 to 1.00 have been reported (Adams, 2011; van Oers & Sinn,
2011), with a trend towards higher heritabilities in more benign environments
(Charmantier & Garant, 2005). The low heritabilities in some studies might be a
result of measuring behaviour in one context only, which may not capture
personality as a genetically driven tendency to behave consistently in a certain way
(Dochtermann, Schwab, & Sih, 2015). When trying to account for the intraindividual variation across situations and tease apart the personality from that
variation, Dochtermann’s et al. (2015) meta-analysis concluded that the heritability
of animal personality is on average .52. This is remarkably similar to the broad sense
heritability estimates from human personality meta-analyses mentioned above (van
den Berg et al., 2014; Vukasovic & Bratko, 2015). Whether the genetic variation in
non-human animal personality include non-additive genetic variation is yet to be
established. However, some evidence of non-additive genetic variance in animal
personality has been accumulating. One study found significant non-additive genetic
variance in orangutans (Adams, King, & Weiss, 2012) and another in great tits (van
Oers, Drent, de Jong, & van Noordwijk, 2004). The fact that non-additive genetic
variation in personality is reported in such phylogenetically distant species as
20
primates (humans included) and birds, might be a sign of non-additive genetic
variance being an evolutionarily common feature in personality.
With the overwhelming evidence from quantitative genetics that phenotypic
personality variation is associated with genetic variation, molecular genetic studies
have tried to identify specific genes related to personality variation. Early candidate
gene studies reported promising findings linking, for example, a serotonin
transporter gene (SLC6A4) with neuroticism and a dopamine receptor gene (DRD4)
with extraversion (e.g., Benjamin, Liz, Pattersonz, & Hamer, 1996; Lesch et al.,
1996). However, later research showed conflicting results on the association between
these genes and personality traits, and reviews and meta-analytic studies have
concluded that if they are involved in personality variation at all, their effects are tiny
(Balestri, Calati, Serretti, & De Ronchi, 2014; Munafò & Flint, 2011; Munafò et al.,
2009; Munafò, Yalcin, Willis-Owen, & Flint, 2008). Some studies in non-human
animals have also tried to find genes for personality. In accordance with studies on
humans, polymorphism in the DRD4-gene has been associated with exploratory
behaviour in great tits (Fidler et al., 2007). Such studies, however, are few in number
and most likely to suffer a similar lack of repeatability as are studies on humans (see,
e.g., Korsten et al., 2010).
With the lack of success to link certain candidate genes with personality variation,
genes for personality have been searched for in genome-wide association studies
(GWAS). Although some new genes for some personality traits have been identified
in GWAS, findings seem to be inconsistent and hard to replicate (de Moor et al.,
2012; de Moor et al., 2015; Service et al., 2012). In fact, one large meta-analytic
GWAS found no significant associations between Cloninger’s temperament scales
(Cloninger,
Svrakic,
&
Przybeck,
1993)
and
common
single
nucleotide
polymorphisms at all (Service et al., 2012). Another large meta-analytic GWAS
examining extraversion also failed to find any polymorphisms significantly related
to its variation (van den Berg et al., 2016). Thus, no genes with replicated substantial
effects on personality are currently identified.
21
In addition to the failure to find specific genes underlying personality variation,
personality traits show a uniquely large quantity of “missing heritability” (see e.g.,
Penke & Jokela, 2016). That is, while the results from twin studies suggest that
around 40 percent of the variation in personality traits is genetic in nature, true
genetic variance detected in genome-wide complex trait analyses (GCTA) seem to
account for only a little of that variance (van den Berg et al., 2016; de Moor et al.,
2015; Power & Pluess, 2015; Verweij et al., 2012; Vinkhuyzen et al., 2012). Such
studies have, in general, found that from 0 per cent to around 20 per cent of the
variance in personality is associated with common single nucleotide polymorphisms.
Furthermore, results from these GCTA studies vary, despite their large sample sizes
and meta-analytical approach. For example, the estimates for extraversion range
from zero percent (van den Berg et al., 2016; Power & Pluess, 2015) to twelve percent
(Vinkhuyzen et al., 2012). The conclusion so far seems to be that at best circa 30 per
cent of the heritability estimated on the basis quantitative genetic studies is
accounted for by actual genetic variation.
Possible reasons for such an apparent paradox are manifold (see van den Berg et
al., 2016; de Moor et al., 2015; Power & Pluess, 2015; Verweij et al., 2012;
Vinkhuyzen et al., 2012). Firstly, it may be that personality variation is mostly driven
by uncommon genetic polymorphisms (i.e., recent mutations). GWAS and GCT
analyses generally detect polymorphisms that are present in at least 0.5 per cent of
the sample. Polymorphisms that are rarer go undetected. Secondly, GWAS and
GCTA look for single nucleotide polymorphisms while it is possible that a substantial
fraction of genetic variation in personality is explained by, for example, copy number
variants. Thirdly, personality may be affected by unusually high amounts of nonadditive genetic variation like epistasis or dominance. And fourthly, it is likely that
personality is affected by substantial interplay between genes and environments:
gene–environment
interactions
or
gene–environment
correlations.
Gene–
environment interaction refers to situations where the resulting phenotype of a
genetic effect is dependent on the environment. Gene–environment correlation, in
turn, refers to situations where individuals are non-randomly selected into specific
22
environmental circumstances driven by their genes. Indeed, some studies have
reported gene–environment interactions regarding personality (Badcock et al., 2011;
Krueger, South, Johnson, & Iacono, 2008; Reiner & Spangler, 2011). Also, it has been
suggested that the human capability of niche-picking or niche construction, i.e.
choosing and building environments suitable for themselves, render geneenvironment correlations in personality very likely (Montiglio, Ferrari, & Réale,
2013; Penke & Jokela, 2016). In certain cases, the presence of gene–environment
interactions or correlations can inflate the estimates of heritability in twin designs
(Purcell, 2002). Heritability in personality might therefore be smaller than the .40
estimated in the recent meta-analyses (van den Berg et al., 2014; Vukasovic &
Bratko, 2015). Further, if a presence of a certain genetic polymorphism leads to
different manifest personalities depending on environmental conditions, this will
hamper the search for genes for personality.
In conclusion, the current research evidence considering the genetics of human
personality (and it is reasonable to assume that this applies to non-human animal
personality as well) suggests that personality is most likely affected by a very large
quantity – thousands even – of genes with infinitesimal individual effects.
Additionally, personality is potentially affected by dominance and epistasis to a large
extent, and shows interplay between genes and environments. Further, it is likely
that the “genes for personality” are “genes for a good many other things” as well. This
is because the genetic pattern for numerous human behavioural traits largely
appears to be the pattern of very many genes with tiny individual effects (see Chabris,
Lee, Cesarini, Benjamin, & Laibson, 2015). With only a little over 20 000 genes in
the human genome, genes would quickly run out without the existence of pleiotropic
effects, or one gene affecting many phenotypic phenomena. To conclude, not much
is currently known about the genetic basis of personality.
1.3 Theories on the evolution of personality
When heavily simplified, evolutionary theory predicts “survival of the fittest”, i.e.,
that natural selection will deplete traits of variation, leaving only the form associated
23
with the greatest fitness (Merilä & Sheldon, 1999). In genetic terms this implies that
traits that have been under natural selection in the evolutionary past, should show
no genetic variation. Recent decades, however, have seen an abundance of evidence
of genetic variation in many traits in both humans and non-human animals (see e.g.,
Polderman et al., 2015). By now, it has become apparent that virtually every
psychological or behavioural trait in humans contains substantial amounts heritable
genetic variance – to the extent that the heritability of behavioural traits is proposed
to be “the first law of behavioural genetics” (Bouchard & McGue, 2003; Turkheimer,
Pettersson, & Horn, 2014). Personality is one of the traits that has recently been
found to show heritable variation across a wide range of taxa (see section 1.2.). Thus,
the existence of genetic variance in personality presents a puzzle. With the surge of
interest in personality in biological sciences, several theories on the mechanisms of
maintenance of this variation have been suggested, some of which are outlined
below.
1.3.1 Selective neutrality
Genetic variation is constantly being introduced into the genetic pool of a population
by mutation (Falconer & Mackay, 1996). Most mutations are deleterious, because
they randomly interfere with the evolved genetic information (Eyre-Walker &
Keightley, 2007). Deleterious mutations are, in their most severe form, lethal, and
may also cause other negative fitness effects such as decreased reproductive success.
Therefore, such mutations are removed from the genetic pool – the rate of them
disappearing depends on how deleterious they are (Falconer & Mackay, 1996). In
rare situations, a mutation improves fitness. A mutation that improves fitness should
eventually spread through the population and become fixated, i.e., become the new
“species-typical” (Merilä & Sheldon, 1999). In addition, some mutations are
selectively neutral: they are neither deleterious nor beneficial to fitness. Because
natural selection does not clean such mutations out nor fixates them, they induce
genetic variation into the genetic pool of a population.
24
Selective neutrality is proposed to be one of the mechanisms maintaining genetic
variation in personality (see e.g., Penke et al., 2007). In the 1990s, Tooby and
Cosmides, early advocates of evolutionary psychology, even stated that selective
neutrality would be the most important reason for “species-typical, adaptively
designed developmental and psychological mechanisms” to show genetic variation
(Tooby & Cosmides, 1990, p. 17). In their view, evolved psychological adaptations
can have genetic variation, as long as “heritability localizes in nonessential parts of
the design” (Tooby & Cosmides, 1990, p. 39), i.e., as long as the heritable variation
does not involve fitness-related consequences. They continue that traits that are not
under intense natural selection tolerate more heritable variation. In other words, the
more heritable a trait is, the less likely it is to be an evolutionarily relevant trait.
The concept of selective neutrality thus predicts that personality variation is not
associated with fitness-related outcomes. This perspective, along with the notions of
Tooby and Cosmides, are strongly challenged by empirical findings. A vast majority
of human traits, be it psychological or physiological, show considerable genetic
variation (Polderman et al., 2015). In personality, genetic variation is usually
estimated to account for about half of the phenotypic variation (see section 1.2.). It
has also become increasingly clear that personality variation is associated with
fitness (see section 1.1.), making it highly unlikely that selective neutrality can
account for the persistent genetic variance in personality (e.g., Bouchard & Loehlin,
2001; Penke et al., 2007). However, evidence concerning the associations between
personality and fitness is so far only based on the phenotypic level, from which it is
not straightforward to draw evolutionary conclusions. In section 2.2. and sub-study
I, I explore this point in more detail.
1.3.2 Mutation–selection balance
The mutation–selection balance hypothesis for the maintenance of genetic variation
involves mutations introducing deleterious genetic variation into the genetic pool of
a population, and unidirectional natural selection clearing it out (Falconer &
Mackay, 1996; Merilä & Sheldon, 1999). The genetic variation in a trait at any
25
moment would therefore be the result of mutations that have yet to be cleared out.
Recent evidence suggests that personality is most likely affected by large numbers of
genes (see section 1.2.). Thus, personality traits have a large “mutational target size”,
i.e., many genes to potentially be affected by mutation (Penke et al., 2007). If so,
personality traits are likely to suffer from relatively large amounts of deleterious
genetic variance caused by recently occurred mutations at any moment (see e.g.,
Verweij et al., 2012). While natural selection works to decrease the deleterious
genetic variation, new mutations occur constantly, and thus the genetic variance is
maintained in a balance between new mutations and selection against them. It is
important to note that the mutation–selection balance hypothesis implies that
personality traits have optimal levels, i.e., a species-typical adaptation which is being
corrupted by mutations. It predicts consistent directional selection favouring one
end (the adaptive one) of the variation in personality traits (Penke & Jokela, 2016).
The mutation–selection balance hypothesis also predicts inbreeding depression: if
close relatives procreate, the offspring will have lower fitness because of increased
concentration of deleterious mutations (Penke et al., 2007).
Some theoretical accounts contemplating the role of mutation–selection balance
conclude that it is an unlikely source of genetic variance in personality (Penke &
Jokela, 2016; Penke et al., 2007). Most importantly, personality does not seem to be
under consistent unidirectional selection (see section 1.1.). However, a recent large
scale study with genome wide single nucleotide polymorphism data concluded that
for the personality traits of the Cloninger’s temperament model (Cloninger et al.,
1993), mutation–selection balance was the most plausible mechanism to have
maintained their variance (Verweij et al., 2012). They detected both a genetic
architecture suggesting mutation–selection balance, and potential inbreeding
depression. It remains to be seen whether the same holds for the much better
validated personality traits of the Five Factor Model. In sum, the genetic architecture
of personality might show signs that are consistent with mutation–selection balance,
while research on the associations between personality traits and fitness yields
conflicting results.
26
1.3.3 Balancing selection
Balancing selection refers to situations where natural selection actively works to
maintain genetic variance in a population, rather than deplete it. Some of the
potential mechanisms for balancing selection will not be discussed here (e.g.,
heterozygote advantage), but I will concentrate on mechanisms possibly involved in
the maintenance of heritability in personality.
One form of balancing selection is called antagonistic pleiotropy. This means that
natural selection can be maintaining genetic variation if a trait affects different
components of fitness antagonistically, for example, by increasing reproductive
success but decreasing survival. Specifically, personality traits could be involved in
life history trade-offs (Nettle, 2005; Nettle, 2006; Wolf, van Doorn, Leimar, &
Weissing, 2007). Life history theory starts from the premise that individuals aim
(unconsciously, at most cases) at maximising their fitness, but have limited
resources to do so (Hill & Kaplan, 1999). Therefore, individuals are faced with
decisions regarding how to allocate their limited recourses, and these decisions
involve trade-offs. The defining feature of a trade-off is that investing resources in
one asset necessarily involves costs to some other. The specific costs and benefits
associated with allocation decisions depend on the environmental conditions as well
as on the characteristics of the individual making the decisions. Personality
variation, therefore, could result from the differentially beneficial allocation
decisions that individuals face.
Nettle (2006) has suggested particular trade-off patterns for each of the
personality traits of the Five Factor Model. For example, neuroticism might be
involved in avoiding risk prone situations and escaping predators more successfully
(see also Lee et al., 2006) and thereby increasing survival. Simultaneously, it might
impair physical health through a highly activated HPA-axis and thereby decrease
survival. Wolf et al. (2007), in turn, created a model resulting in a personality trait
with correlating facets such as tendency to explore and aggressiveness, in response
to the future reproductive value of the individual. They summarise their findings as
“the more an individual has to lose, the more risk-averse it should be” (Wolf et al.,
27
2007, p. 583). The trade-off hypothesis of personality suggests that if measuring only
one fitness component, evolutionary inferences may be misled. The genetic variance
in personality is maintained because the net effect of personality on fitness is neutral
across personality variation, even if associations with individual fitness components
may be directional. The possible role of one specific trade-off, between the quality
and quantity of offspring, in maintaining heritable personality variation in humans
is addressed in section 2.3. and sub-study II.
Another form of balancing selection is fluctuating selection due to environmental
heterogeneity. This means that different ends of a personality trait continuum are
beneficial in different environmental circumstances. For example, a heightened
sensitivity to risks in individuals with high neuroticism might be beneficial in
circumstances of high predation risk, while its detrimental effects on physical health
would overcome the benefits where predation risk is low (Nettle, 2006). In their
studies on the great tit, Dingemanse et al. (2004) showed that the fitness effects of a
personality trait they called “exploratory behaviour” varied between sexes and
environmental conditions (food availability). When food was scarce, females who
explored more had a higher survival rate. In contrast, when food was abundant
females who explored more had a lower survival rate, presumably because overt
aggression lead to more conflicts and thus excess mortality. For males, the pattern
was reversed, and the authors hypothesised this to be due to territorial competition
being more important in males than in females. With abundant food, more
individuals survive through the winter, territorial competition increases, and males
with higher exploration tendencies fare better (Dingemanse et al., 2004). It is
plausible that environmental heterogeneity could cause differential benefits and
costs of other personality traits in other animals as well. The question of
environmental heterogeneity is inherent in the present thesis, as all three substudies are performed in different human populations.
Also other mechanisms maintaining genetic variation in personality related to
environmental heterogeneity have been proposed. For instance, negative frequencydependent selection implies that a trait is only beneficial to an individual’s fitness if
28
it is rare in the social environment. Such a mechanism could maintain variation in
personality through costs and benefits encountered in interactions with other
individuals in the population (see e.g., Wolf & McNamara, 2012). In principle,
aggressive individuals outperform non-aggressive ones only if they are in the
minority – if they are the majority, the costs related to aggressive encounters
overcome the benefits. Also, the fitness consequences of personality traits could vary
in response to the sex-ratio, i.e. the proportion of men in the population (Del
Giudice, 2012). For example, male-biased sex ratios increase male–male
competition and promote relationship stability and paternal investment in offspring,
which could provide fitness benefits to personalities able to cope with such demands.
It is, however, beyond the scope of this thesis to examine such hypotheses
empirically.
Finally, genetic variation in personality, or any other trait for that matter, could
be maintained by stabilising or disruptive selection (Falconer & Mackay, 1996).
Stabilising selection refers to situations where intermediate characteristics are
favoured by natural selection – for instance, if average levels of extraversion would
be associated with the greatest fitness, as opposed to either extreme extraversion or
extreme introversion. Disruptive selection, in turn, is implied when both extremes
show highest fitness and intermediate levels the lowest. With stabilising or
disruptive selection, non-linear associations between the trait and fitness are
expected. Each of the three sub-studies of this thesis tested for the presence of nonlinear associations between personality and reproductive success.
29
2 Research aims and scope
The research field concerning the evolution of personality is relatively young, and
further scattered by being studied from two rather different perspectives with little
communication in between (Koski, 2011; Nettle & Penke, 2010). Psychologists,
possessing intricate knowledge on how to conceptualise and measure personality,
are insufficiently informed and interested in examining the “what is it there for and
how”. And vice versa, behavioural ecologists and other scholars from the biological
sciences have partly jumped over the “what is it”, and rushed on to study the “why is
it”. As a result, the field is filled with gaping holes, some of which the present thesis
attempts to bridge.
All three sub-studies of the present thesis provide further empirical evidence on
the phenotypic associations between personality traits and reproductive success in
contemporary humans. In addition, each of the sub-studies bring about new
perspectives on the matter. Firstly, to date, studies on the evolution of personality
are based on the so called “phenotypic gambit”: the notion that observed phenotypic
fitness associations correspond to underlying genetic patterns in a way that warrants
evolutionary conclusions (van Oers & Sinn, 2011). An increasing body of research in
the field of quantitative and population genetics suggests that this is not nearly
always the case, especially when studying wild populations (Morrissey, Kruuk, &
Wilson, 2010). Crucially, no studies to date have looked at the justification of the
phenotypic gambit in studies of personality evolution, a neglect that sub-study I tries
to correct. Secondly, empirical studies on the associations between human
personality and fitness have largely overlooked the possible evolutionary
mechanisms maintaining heritable variation in personality, despite these having
been hypothesised in more theory-driven publications (e.g., Penke et al., 2007). In
sub-study II, the question of directional selection on personality is examined by
testing whether personality is associated with a trade-off between quality and
quantity of offspring. Thirdly, studying contemporary humans can seem far-fetched
from an evolutionary point of view. In fact, it is sometimes argued that with modern
contraception, humans have ceased to evolve (see e.g., Stearns, Byars, Govindaraju,
30
& Ewbank, 2010), or that “counting babies” is not informative of evolutionarily
relevant behaviour in contemporary humans (see e.g., Crawford, 2000). Sub-study
III examines the associations between personality and reproductive behaviour in the
contemporary UK, and draws inferences on the possible current vs. past selective
pressures. The specific research questions of the present thesis and backgrounds for
the three sub-studies are detailed below.
2.1 Is personality phenotypically associated with lifetime
reproductive success?
We know that personality is associated with reproductive success in humans and in
non-human animals. All three sub-studies of the present thesis provide further
evidence on the question of phenotypic associations between personality traits and
fertility. More specifically, they provide information on the associations between
personality traits and lifetime reproductive success, as studies with sufficiently old
human participants are not abundant. Lifetime reproductive success is often
considered to be the most reliable fitness measure: it provides one single measure of
the proportion of genes an individual passes on to the next generation (although it
has been argued that the pace of reproduction also plays an important role in nonstationary, growing populations, see e.g., Jones & Bird, 2014). Furthermore, even
though associations between fitness and personality are by now relatively solidly
established, the strength and even direction of those associations vary from study to
study, leaving room for further research. For instance, consistent variation in the
fitness consequences of personality between different populations is indicative of
balancing selection through environmental heterogeneity, whereas consistent
directional selection could point to mutation–selection balance (see sections 1.3.2.
and 1.3.3.). The three sub-studies of this thesis are performed on three different
samples of contemporary Western humans. These societies are quite similar in terms
of longevity, rates of mortality and fertility, but differ in, for example, level of welfare
benefits and social inequality. Such subtler differences provide grounds for
31
examining the consistency or inconsistency in the associations between personality
and fitness.
2.2 Are the phenotypic and genetic associations between
personality and lifetime reproductive success different?
Theoretical accounts on the evolution of personality usually start from the premises
that personality is under natural selection and that the phenotypically observed
natural selection induces evolutionary responses in personality (e.g., Dingemanse &
Réale, 2005; Dingemanse & Wolf, 2010).
Sub-study I examines whether the
phenotypic approach is justified when considering the evolution of personality.
Natural selection observed at the phenotypic level has no evolutionary consequences
unless the trait correlates genetically with fitness (van Tienderen & de Jong, 1994).
Especially in wild populations, where environmental variation affects phenotypes,
the underlying genetic associations between a trait and fitness may not correspond
to the phenotypic associations (Morrissey et al., 2010). This poses a problem for
evolutionary theory and predictions.
The traditional way to predict selection response, that is, the change in population
mean across two generations, is based on the breeder’s equation (Morrissey et al.,
2010). According to the breeder’s equation, the selection response equals the
product of the trait’s heritability and phenotypic selection differential. Selection
differential is the phenotypic covariance between the trait and relative fitness
(Falconer & Mackay, 1996). In controlled conditions with little environmental
variation, genetic differences are likely to be manifested on the phenotypic level, so
that selection for phenotypes correlates with genotypes as predicted by the breeder’s
equation (Hill, 2014). In wild populations, this rarely seems to be the case, and there
are numerous examples with no selection response despite an apparent directional
selection for a heritable trait (Merilä, Sheldon, & Kruuk, 2001).
There are several possible reasons for the breeder’s equation to fail to predict
selection responses correctly. Firstly, if two traits are genetically correlated, selection
pressures on one trait may induce evolutionary changes in the other (Dochtermann
32
& Roff, 2010). A multivariate form of the breeder’s equation can incorporate more
complex information on the genetic correlations between multiple traits, which
improves the prediction of selection responses (Lande & Arnold, 1983). Secondly,
and even more importantly, the breeder’s equation yields unbiased estimates only if
the phenotypic selection differential reflects a causal association between the trait
and fitness (Morrissey et al., 2010). In other words, the estimates of selection
responses may be biased by, for instance, confounding environmental factors.
For example, a study on wild red deer (Kruuk et al., 2002) found antler size to be
both heritable and phenotypically associated with fitness, yet there was no evolution
on antler size during the 30-year study period. Nutritional status and other
environmental factors may have influenced both antler size and fitness, thus creating
a spurious selection differential for antler size. Similarly, in passerine birds the
condition of fledglings was both heritable and positively associated with fitness, but
the average phenotypic condition still decreased rather than increased during the
20-year study period (Merilä et al., 2001). The fledglings’ condition was selected for
at the genetic level, and average genetic condition did indeed increase over time.
However, this positive genetic change was probably concealed by simultaneously
deteriorating environmental conditions, i.e., decreasing food supply (Merilä et al.,
2001). As these two examples clearly illustrate, it may not be justified to make
inferences about (micro)evolution on the basis of phenotypic selection differentials.
The fact that all previous studies examining the fitness consequences and evolution
of personality have only concentrated on the phenotypic approach, represents a
serious omission. We actually do not know whether natural selection works on
personality, or whether the associations between personality and fitness are
indicative of, for example, confounding environmental effects.
An alternative way of predicting microevolution is the Robertson–Price identity,
or the secondary theorem of natural selection, according to which the selection
response equals the additive genetic covariance between the trait of interest and
relative fitness (Falconer & Mackay, 1996; Morrissey et al., 2010). This equation is
less sensitive to environmental confounding factors that may bias the breeder’s
33
equation (Morrissey et al., 2010; Rausher, 1992). The Robertson–Price identity is
biased only if genetic confounders are omitted from the model (Rausher, 1992).
In sub-study I, I and my co-authors investigated whether the expected
microevolution of personality in response to natural selection observed on a
phenotypic level is similar to the expected microevolution based on genetic
information. Using a large twin sample of contemporary Western humans, we
examined the phenotypic and genetic associations between fitness and two
personality traits—extraversion and neuroticism. We conducted analyses based on
the viewpoint of the breeder’s equation and based on the viewpoint of the
Robertson–Price identity to see if they give qualitatively similar results.
2.3 Is personality involved in a reproductive quality–quantity
trade-off?
Many theorists have suggested that balancing selection in the form of life history
trade-offs could be involved in maintaining genetic variation in personality (see
section 1.3.). The trade-offs may be evident within an individual’s life-span – for
example, extraversion could be increasing an individual’s mating success but
decreasing life-expectancy (Nettle, 2005). Therefore, studies involving only one
fitness component (e.g., mating success) could conclude that extraversion is under
positive directional selection, while the net effect of different selective forces would
result in equal fitness across the extraversion continuum. Sub-study II, however,
examines a trade-off whose net effect on fitness only becomes evident when
measured over more than one generation, namely the trade-off between the quality
and quantity of offspring. This trade-off arises from the fact that the more offspring
a parent has, the less resources and time he or she can provide to each individual
offspring (Hill & Kaplan, 1999; Trivers, 1972). Accordingly, individuals may decrease
the number of progeny in order to improve their quality (Hill & Kaplan, 1999). They
might, for example, invest resources to increase the survival of the offspring or,
especially in the case of contemporary Western humans, increase the reproductive
success of the offspring. Therefore, a positive association between a trait and number
34
of children could be inverted in the next generation by poorer fitness of those
children. The hypothesis is that if personality traits work to steer an individual’s
reproductive strategy toward focusing either on quantity or quality of offspring, the
associations between personality and number of children are expected to be nullified
or reversed in the generation of the grandchildren.
A few studies in humans and non-human animals have tackled the question of
personality and quality–quantity trade-offs. For example, in a study on captive
minks, less active females had smaller litters but faster growing kits, suggesting
parental investment in the quality rather than quantity of offspring (Meagher,
Bechard, & Mason, 2012). In a high-fertility and high-mortality human population,
female neuroticism was associated with more offspring but also with lower body
condition of the offspring (Alvergne et al., 2010). However, this trade-off was only
evident in women with low access to resources. Among Tsimane foragerhorticulturalists,
higher
extraversion,
conscientiousness
and
openness
to
experience, and lower neuroticism in men were associated with a higher number of
children (Gurven et al., 2014). But, there were no associations between personality
traits and child mortality before age 15 (Gurven et al., 2014), implying no quality–
quantity trade-offs. In contrast, a study on contemporary British adults found
extraversion to be positively associated both with number of lifetime sexual partners
and likelihood of being hospitalised due to an illness or accident (Nettle, 2005). This
suggests that extraversion could be involved in a trade-off between mating success
and mortality due to risk-taking propensity. Lastly, a study on contemporary
Americans found interaction effects between offspring number and parental
personality on child education: parental neuroticism was detrimental to child
education only in large families and parental openness to experience only to later
born children (Jokela, Alvergne, Rotkirch, Rickard, & Lummaa, 2014). These
studies, however, did not incorporate data on the children’s reproductive success,
and thus do not provide information on the long-term fitness effects of the alleged
quality–quantity trade-off. Sub-study II compares the associations between
personality traits and number of both children and grandchildren to explore if there
35
are indications of a reproductive quality–quantity trade-off. Number of
grandchildren, which combines both offspring number in the first generation
(quantity) and their survival and reproductive output (quality) into one measure, is
an ideal measure to examine such questions.
2.4 Are there novel selection pressures on personality in
contemporary humans?
Sub-study III focuses on some of the specific questions that are involved in studying
the evolution of personality with modern humans. Personality traits are associated
with many aspects of human lives, such as academic achievement (Poropat, 2009),
the propensity to behave promiscuously (Banai & Pavela, 2015), or the likelihood to
use contraception effectively (Niemelä et al., 1981). Contemporary Western
environments are therefore likely to have introduced new selective pressures on
personality while rendering some previous ones unimportant. However, some of the
pathways linking personality and reproductive success in modern humans may well
be similar to those in our evolutionary past (and possibly to non-human animals,
too).
In sub-study III, the associations between personality variation and fitness are
studied from the point of view of reproductive behaviour. This is done by examining
the associations between personality traits and planned pregnancies on one hand,
and between personality and non-planned pregnancies on the other. The hypothesis
is that these associations are indicative of novel vs. older selective pressures on
personality. With regards to non-planned pregnancies, it is hypothesised that a
successful avoidance may be a sign of new selective pressures, such as personality
affecting the ability to use modern contraceptive methods effectively. Higher
incidence of non-planned pregnancies could, in turn, reflect behaviours that would
have increased number of offspring in the past as well, such as higher socio-sexuality.
With planned pregnancies, a higher probability could indicate older selective
pressures, such as an increased desire to have children (Rotkirch, 2007), or being a
more attractive (e.g., trustworthy) partner and potential co-parent (Stone,
36
Shackelford, & Buss, 2012; van der Linden, Figueredo, de Leeuw, Scholte, & Engels,
2012). Conversely, a lower probability of planned pregnancies could be a sign of
more recent selective pressures, such as being more interested in pursuing other life
goals than forming a family.
37
3 Methods
3.1 Participants
The specific research questions in this thesis were examined by using three distinct
pre-existing datasets from three different Western, contemporary populations. All
datasets were from survey-based, longitudinal, nationally representative research
projects, and they can be accessed by interested researchers freely or on request.
Data from three different populations allows the examination of fluctuating selection
pressures in different environments. Further, using data on contemporary Western
humans presents an opportunity to contemplate evolutionary mechanisms that are
still working in spite of low fertility, low mortality, and effective contraception. The
three datasets, in addition, provide reliable personality measures and detailed
fertility history, reaching in most cases the very end of reproductive lifespan. As
lifetime reproductive success seems to be the most important fitness indicator in
post-industrial human populations (see e.g., Goodman & Koupil, 2009), these
datasets provide optimal grounds for looking at personality evolution in such
circumstances.
While the original samples of the datasets used in this thesis were nationally
representative, for the purposes of the analyses, some participants (with missing
data on personality, for example) had to be dropped. However, this did not have large
effects on the representativeness of the samples. In terms of distribution of socioeconomic status, ethnicity, or number of children, the study samples in general
resembled national averages rather closely (see the Methods sections of the original
publications for details).
Sub-study I. In examining the potential differences between the phenotypic and
genetic view on selection responses, the Older Twin Cohort of the Finnish Twin
Cohort Study was used. In 1974, all Finnish twin pairs of the same sex born before
1958 with both co-twins alive (N=13,888) were identified from the Population
Register Centre of Finland (Kaprio & Koskenvuo, 2002). In 1975, a questionnaire
concentrating on genetic and environmental origins of complex diseases was mailed
38
to these twins (response rate 89 %). Extraversion and neuroticism were also assessed
in this questionnaire (Rose, Kaprio, & Koskenvuo, 1988). The sample for sub-study
I consisted of individuals born in 1950–1957, for whom data on live births from the
Population Register of Finland were available.
Sub-study II. The potential reproductive quality–quantity trade-off related to
personality was assessed with data derived from the Health and Retirement Study
(HRS). The HRS is a study of more than 30,000 individuals representing the U.S.
population older than 50 years, started in 1992 (Juster & Suzman, 1995). Telephone
or in-person interviews are conducted every second year, and administered under
the National Institute of Aging and the University of Michigan's Institute for Social
Research. As of 1992, the HRS consists of many different sources of data collection,
and several new subsamples have been included in the study as the original cohorts
have aged. In the case of married couples, both spouses (including spouses who
would otherwise not be age-eligible for the study) have been interviewed. The sample
for sub-study II consisted of individuals with no missing information on number of
grandchildren or personality, and who were at least 55 years old at the time data on
the number of their grandchildren was collected. The data thus include US men and
women born between 1900 and 1947.
Sub-study III. The associations between personality traits and planned and
non-planned pregnancies were investigated with participants from the 1958 British
birth cohort study (also known as the National Child Development Study, NCDS)
(Atherton, Fuller, Shepherd, Strachan, & Power, 2008; Power & Elliott, 2006). The
original participants were 17,634 individuals born in England, Wales, and Scotland
during one week in March 1958. Data have been collected in several follow-up
phases. The analytic sample of sub-study III consisted of men and women who had
participated in at least one of the adulthood rounds in 1991 (at 33 years), 2000 (at
42) or 2004 (at 46) and had information on personality (measured in 2008 at 50
years) and adulthood social class. The details of the analytic samples of the three
sub-studies, along with the variables used in the analyses, can be found in Table 1.
39
Table 1. Details of the samples and variables used in sub-studies.
N (%
Sub-study Sample
Cohorts
Outcome
males)
Register
based
number of
Finnish Twin
I
7669 (48)
1950–57
children until
Cohort Study
the end of
2009
II
Health and
Retirement
Study,
the US
10679 (42)
1900–47
Self-reported
number of
children &
grandchildren at
mean age
67.7
Self-reported
number of
children by
National
age 42–46,
Child
and number
III
1958
Development 8336 (48)
of planned
Study,
and nonthe UK
planned
pregnancies
by age 33
Note. FFM = Five Factor Model; MIDUS = Midlife in the United States.
Personality
measure
Covariates
Eysenck
extraversion
&
neuroticism;
Short form
Age, sex,
children born
prior to
personality
measurement
FFM; adapted
from MIDUS
Age, sex,
race/ethnicity,
educational
level
FFM;
International
Personality
Item Pool
Sample
attrition
indicator¸
socio-economic
status
3.2 Measures
Personality. In all sub-studies, personality was measured with a self-reported
questionnaire. In sub-studies II and III, the FFM-type of personality traits –
extraversion, neuroticism, agreeableness, conscientiousness, and openness to
experience – were assessed. In sub-study I, in turn, only extraversion and
neuroticism as defined by Eysenck’s theory of personality were measured (Eysenck,
1967). In sub-study I, neuroticism (10 items) and extraversion (9 items) were
measured with a short form of the Eysenck Personality Inventory developed by
Floderus (Floderus, 1974). The items were answered on a dichotomous yes/no scale.
Mean scores for the scales were calculated if no more than two items in the scale
were missing. In sub-study II, personality was measured using an instrument
adapted from the MIDI personality scales developed for the MIDUS study (Lachman
40
& Weaver, 1997). The measure has 4–7 items for each of the five higher order traits,
rated on a 4-point Likert scale. Mean scores for personality scales were calculated
for individuals with no more than one missing item in the scale. In sub-study III,
personality was measured with the 50-item Big Five model of the International
Personality Item Pool (Goldberg, 1999). Each personality scale had ten items, rated
on a 5-point Likert scale. A mean score for each personality trait was calculated if no
more than one item was missing in a scale. The Eysenkcian extraversion and
neuroticism closely correspond the FFM ones (Avia et al., 1995; Draycott & Kline,
1995), and in this thesis they are treated as manifestations of the same personality
traits. In order to make comparisons between the sub-studies easier, the emotional
stability scale in sub-study III is hereafter reversed into a neuroticism scale.
Number of offspring. In sub-studies II and III, number of offspring was based
on self-report, whereas in sub-study I, register data on live births was linked to study
participants using a unique personal identification number assigned to all Finnish
citizens at birth. The information concerning offspring in all three sub-studies was
attained towards the end reproductive lifespan. Since all the participants in all substudies were at least strongly approaching, if not yet completely reached, the end of
their reproductive lifespan, number of children in this thesis is considered to be
lifetime reproductive success. In order to prevent outliers to influence the analyses
too strongly, number of offspring was top-coded in all three sub-studies. For the
purposes of sub-study I, lifetime reproductive success was converted into relative
fitness, denoted by w, which is the individual’s number of children relative to the
mean number of children in the population (Falconer & Mackay, 1996).
Socio-economic status. In sub-studies II and III, a subset of analyses was run
controlling for socio-economic status (SES) of the participants. This was done
because previous research has shown that education and social class correlate with
both personality and reproduction (Poropat, 2009; Poropat, 2014; Skirbekk, 2008).
In sub-study II, SES was approximated by using educational level (years of
education; range 0–17). In sub-study III, SES was defined based on the British
Registrar General’s occupational class categorisation (I = professional, II =
41
managerial, IIIa = skilled non-manual, IIIb = skilled manual, IV = semi-skilled, V =
unskilled).
Additional measures. In addition to the measures used in all sub-studies, some
additional variables specific to each sub-study were used. In sub-study I, zygosity
(i.e., whether a twin is identical or fraternal) was assessed in the 1975 questionnaire
with questions about the similarity of appearance of a twin pair at an early school
age. This is a standard procedure used to determine zygosity in twin studies, and it
has been shown to have high validity against genetic markers in the present sample
(Sarna, Kaprio, Sistonen, & Koskenvuo, 1978). Also, since personality was measured
in early adulthood in sub-study I, it was possible to control for a potential problem
of reverse causality (number of children affecting personality, and not vice versa;
Jokela et al., 2009). This was done by adjusting all models for number of children
born prior to the personality assessment in 1975.
In sub-study II, using data from the US, information about the participants’
race/ethnicity (based on self-reports; white/Caucasian, black, Hispanic, or
other/unknown) was included as a covaritate in all regression models. Ethnic
background is strongly related to reproductive output in the US (Sandefur, Martin,
Eggerling-Boeck, Mannon, & Meier, 2001). Adjusting for ethnicity as a potential
confounder thus yields more reliable estimates on the associations between
personality traits and reproductive success.
In sub-study III, the planning status of pregnancies was an additional outcome.
It was assessed with a question “Were you planning to have a baby around this
time?’’ (0 = No/Not sure, 1 = Yes). This question was only asked in one of the
adulthood rounds of the NCDS, when the participants were 33 years old. The
analyses concerning the planning status therefore do not represent completed
reproduction. Also, to account for sample attrition, a variable indicating the last data
point available for the participant was used as a covariate in all models.
42
3.3 Statistical analyses
The associations between personality and number of children were assessed with
linear regression in all three sub-studies. In these, as well as all the other analyses,
the two or five personality traits were always mutually adjusted (i.e., included in the
same model at the same time) to assess their independent effects. The correlated
nature of the data (both two twins from the same family contributing to the results)
in sub-study I was taken into account by using robust variance estimation for clustercorrelated data (Williams, 2000) in the linear regression models.
In all three sub-studies, the possibility of non-linear associations between
personality and lifetime reproductive success was tested by running models with
categorised and quadratic personality traits as independent variables. No signs of
non-linearity were detected in any of the sub-studies. In sub-studies I and II, sex
differences in the associations between personality and number of children were
tested by including sex–personality interaction terms in the linear regression
models. The results are reported separately by sex, where significant sex differences
were detected. In sub-study III, the sexes were modelled separately, but the
difference between the sexes was not statistically tested in the original publication.
Here, however, the statistical significance of sex differences in the associations
between personality traits and number of children is examined. The results reported
in Chapter 4 are from multivariate regression models adjusted for the covariates
listed in Table 1, excluding SES. In sub-studies II and III, the effect of SES in the
associations between personality and number of children was examined by running
models additionally adjusting for SES. When adjusting for SES had a significant
impact on the results, these are reported.
3.3.1 Statistical methods for sub-study I
The similarities and dissimilarities between the phenotypic and genotypic approach
to predicting microevolutionary change in personality were examined with the help
of twin modelling. First, univariate standard biometrical twin modelling (Neale &
Maes, 2003) was used to determine which of the three components, genetic (A),
43
shared environmental (C), and unique environmental (E), contribute to the observed
phenotypic variance in neuroticism, extraversion, and w. Sex differences were tested
by allowing the parameter estimates of the models to differ in men and women.
Based on the univariate analyses, shared environmental components (C) were
dropped from subsequent multivariate models for all three variables. Then, the
standard biometrical twin model was extended into a multivariate Cholesky
decomposition (Neale & Maes, 2003), to attain the genetic (G) and environmental
(E) covariance matrices for neuroticism, extraversion, and w. Again, sex differences
were tested by allowing the parameter estimates of the models to differ in men and
women.
Then, the parameter estimates attained from the phenotypic linear regression
modelling and twin modelling were used to predict the selection responses for
neuroticism and extraversion. The predicted selection responses were calculated by
using the univariate and multivariate forms of the breeder’s equation (Falconer &
Mackay, 1996; Lande & Arnold, 1983) and the Robertson–Price identity (Falconer &
Mackay, 1996; Morrissey et al., 2010). According to the univariate breeder’s
equation, the selection response equals the selection differential multiplied by
heritability. The multivariate breeder’s equation takes into account genetic
correlations between traits. If neuroticism and extraversion are genetically
correlated, selection for extraversion will affect the selection response of neuroticism
as well. According to the Robertson–Price identity, in turn, the selection response
equals the additive genetic covariance between a trait and fitness. In calculating
selection responses according to Robertson–Price identity, the genetic correlation
between extraversion and neuroticism was also taken into account. The detailed
description of the equations used to calculate the selection responses can be found
in the original publication. Finally, the attained estimates of selection responses were
compared against each other to see if the phenotypic and genetic approaches to
selection on personality yield qualitatively similar results.
44
3.3.2 Statistical methods for sub-study II
The extent to which personality is involved in quality–quantity trade-offs was
examined by linear regression of number of children and number of grandchildren
on personality. The attained standardised regression coefficients were compared to
see if they give similar results. The regressions were adjusted for the covariates listed
in Table 1. In addition, a model was run where the regression of number of grandchildren on personality was further adjusted by number of children. This was done
to assess if personality has an independent effect on number of grandchildren above
and beyond its effects on the size of the first generation.
3.3.3 Statistical methods for sub-study III
The associations between personality and planned or non-planned pregnancies
(carried or conceived, modelled separately for men and women) were examined with
multiple-event Cox regression according to the Andersen–Gill model (Andersen &
Gill, 1982). All models were adjusted for the covariates listed in Table 1. The Efron
method was used to handle ties (i.e., simultaneous events in the dataset; Cleves,
Gould, & Gutierrez, 2002). The timing of pregnancies was recorded in months,
starting from the participant’s 12th year birthday. Two time-varying covariates of
fertility history (the number of live children top-coded at five, and the number
pregnancies ending in miscarriage / stillbirth / abortion top-coded at four) were
included in the models to take into account the order of births.
To further examine the associations between personality traits and reproduction,
the timing of live births was investigated. In these analyses, separate linear
regression models at each age were fitted, predicting the number of children the
participant had had up to that age (that is, 22 regression models between ages 12 and
33 for births from planned and non-planned pregnancies). To illustrate the changing
patterns, the regression coefficients obtained from these analyses were plotted
against age.
45
4 Results
4.1 Personality and lifetime reproductive success in
contemporary humans
The associations between personality traits and lifetime reproductive success from
the three sub-studies are summarised in Figure 1. The associations were modelled in
the whole sample when no significant sex differences were found, and separately for
men and women if a statistically significant sex difference was detected. For the
Finnish Twin Cohort of sub-study I, differences between the sexes were significant
for extraversion but not for neuroticism. For the sample from Health and Retirement
Study of sub-study II, no significant sex differences were detected. For the sample
from National Child Development Study of sub-study III, sex differences were
significant for neuroticism, agreeableness and conscientiousness, but not for
extraversion or openness to experience.
In all three sub-studies, extraversion was positively associated with lifetime
reproductive success (in the Finnish twin data of sub-study I, only in men). Openness
to experience and conscientiousness were negatively associated with lifetime
reproductive success in the two sub-studies (II and III) where they were measured.
However, in the British sample of sub-study III, conscientiousness was negatively
associated with number of children only in women. Otherwise, the associations
between personality traits and lifetime reproductive success varied by sample. In
sub-studies II and III, the effect of SES on the associations between personality and
lifetime reproductive success was also examined. In the American data in sub-study
II, controlling for socio-economic status attenuated the association between
openness to experience and lifetime reproductive success, rendering it nonsignificant (b = -0.03, p = .144). In contrast, this adjustment strengthened the
association between neuroticism and lifetime reproductive success to a statistically
significant level (b = -0.04, p = .026). In sub-study III, controlling for socioeconomic status had no statistically significant effects.
46
Figure 1. Associations between personality traits and lifetime reproductive success. Regression
coefficients (for standardised personality traits; mean = 0, sd = 1) and 95 % confidence intervals from
the three sub-studies, Finnish Twin Cohort (FTC; sub-study I), Health and Retirement Study (HRS;
sub-study II), and NCDS (National Child Development Study; sub-study III). E = extraversion, N =
neuroticism, A = agreeableness, C = conscientiousness, O = openness to experience.
47
4.2 Phenotypic and genetic assessment of response to selection
(sub-study I)
The differences between the phenotypic and genetic viewpoints on evolutionary
responses to natural selection were investigated with a Finnish twin dataset in substudy I. When examined phenotypically, neuroticism was not associated with
lifetime reproductive success in women or in men. Extraversion was positively
associated with lifetime reproductive success, but only in men (Figure 1). When
examined with a genetically informative twin model, extraversion and neuroticism
covaried both genetically and environmentally in both men and women, and
extraversion covaried genetically, but not environmentally, with lifetime
reproductive success in men (Table 2). Neuroticism and lifetime reproductive
success did not covary either genetically nor environmentally in either sex in the twin
model (Table 2). The differences in the estimates between men and women were
statistically significant in the multivariate genetic model.
The predicted selection responses from the three different equations (univariate
and multivariate breeder’s equation and the Robertson–Price identity), using
estimates attained from the phenotypic and genetic analyses, are depicted in Figure
2. For neuroticism, the univariate version of the breeder’s equation did not predict a
selection response. Due to indirect selection through the genetically correlated trait
extraversion, the multivariate breeder’s equation and the Robertson–Price identity
yielded negative point estimates of selection response in neuroticism. However, the
point estimate from the Robertson–Price identity was almost three times larger than
the one based on the multivariate breeder’s equation. For extraversion, the point
estimates form the univariate and multivariate breeder’s equations were almost
identical, while the expected selection response based on the Robertson–Price
identity was almost three times larger. In terms of standard deviations, based on the
Robertson–Price identity, the next generation is expected to be .05 standard
deviations less neurotic, and .11 standard deviations more extraverted than the
studied generation.
48
Table 2. Genetic and environmental variances of and covariances between extraversion (E), neuroticism
(N), and relative fitness (w) attained from the multivariate twin model.
(Co)variance
95 % confindence interval
p
N
2.99
2.63, 3.35
E
4.04
3.57, 4.51
w
3.51
2.92, 4.10
N–E
-0.95
-1.25, -0.65
.000
N–w
-0.23
-0.53, 0.07
.135
E–w
-0.03
-0.37, 0.31
.859
N
2.63
2.36, 2.91
E
3.52
3.11, 3.93
w
5.50
4.94, 6.05
N–E
-0.71
-0.95, -0.47
.000
N–w
-0.04
-0.29, 0.21
.752
E–w
0.11
-0.15, 0.37
.399
N
3.31
2.90, 3.73
E
3.12
2.60, 3.63
w
3.91
3.13, 4.69
N–E
-1.20
-1.56, -0.84
.000
N–w
0.12
-0.24, 0.48
.507
E–w
0.59
0.18, 1.00
.005
N
3.03
2.67, 3.39
E
4.10
3.65, 4.56
w
6.08
5.36, 6.79
N–E
-0.74
-1.04, -0.44
.000
N–w
-0.11
-0.40, 0.18
.446
E–w
-0.10
-0.46, 0.26
.573
Women
Genetic
Environmental
Men
Genetic
Environmental
49
Figure 2. Predicted selection responses of neuroticism and extraversion in units of standard
deviation, as estimated on the basis of the three different equations.
4.3 Reproductive quality–quantity trade-offs and personality (substudy II)
In sub-study II, whether or not personality is associated with a quality–quantity
trade-off in respect to number of offspring was examined with the HRS data from
the US. Personality associations with number of children and number grandchildren were compared. The results concerning number of children are summarised
in Figure 1, and section 4.1. Higher extraversion and higher agreeableness, and lower
conscientiousness and lower openness to experience were associated with a higher
number of grandchildren (Table 3, Model 1). When adjusting for number of children
to determine the independent effects of personality on number of grand-offspring,
the association with number of grandchildren was attenuated by 60 per cent for
extraversion, by 25 per cent for agreeableness, by 41 per cent for conscientiousness,
and by 62 per cent for openness to experience (Table 3, Model 2). However, the
association between extraversion and number of grandchildren was the only one to
lose statistical significance in this adjustment. The associations between personality
traits and number of grandchildren were further attenuated when controlling for
50
education (Table 3, Model 3) and openness to experience and agreeableness lost
statistical significance in this adjustment.
Table 3. Associations between personality traits and number of grandchildren.
Model 1
Model 2
b
95 % CI
p
b
95 % CI
p
Model 3
b
95 % CI
p
Extraversion
.24
.11, .37
.000
.10
-.00, .19
.057
.05
-.05, 15
.302
Neuroticism
-.07
-.17, .04
.220
-.02
-.10, .06
.647
-.06
-.14, .02
.116
Agreeableness
.17
.05, .30
.008
.13
.04, .22
.007
.08
-.01, .17
.090
Conscientiousness
-.26
-.37, -.14
.000
-.15
-.24, -.07
.001
-.13
-.21, -.04
.003
Openness
-.37
-.49, -.24
.000
-.14
-.23, -.04
.005
.06
-.04, 16
.218
Note. Linear regression models of number of grandchildren on personality. All models adjusted for age and
race, and the model for the whole sample additionally adjusted for sex. Personality traits are standardised and
entered simultaneously.
Model 1 = Adjusted for age and race/ethnicity.
Model 2 = Additionally adjusted for number of children.
Model 3 = Additionally adjusted for number of children and SES (education).
To compare the associations between children and grandchildren, the results
from the regressions were plotted in one figure. Standardised coefficients and their
95% confidence intervals from three linear regressions of number of children and
grandchildren on personality are depicted in Figure 3. The 95% confidence intervals
of the standardised betas from the first grandchild model overlap the estimates from
the child model (and vice versa), which means that all personality traits were
similarly associated with number of children and number of grandchildren. Thus,
the standardised effect sizes between personality and reproductive success did not
attenuate over two generations. When adjusting for number of children, extraversion
and openness to experience were less strongly associated with number of
grandchildren than with number of children, whereas the other traits were not
differently associated with number of children and number of grandchildren.
51
Figure 3. Associations between personality, number of children and number of grand-children from
three linear regression models. Standardised betas and 95% confidence intervals. All models are
adjusted for race/ethnicity, age, and sex, and in the second grandchild-model, additionally for
number of children. E = extraversion; N = neuroticism; A = agreeableness; C = conscientiousness; O
= openness to experience.
4.4 Personality and planned and non-planned pregnancies (substudy III)
Associations between personality and planned vs. non-planned pregnancies were
analysed in sub-study III, using the NCDS data from the UK. The probability of nonplanned pregnancies was higher among women with higher extraversion and
neuroticism, and lower conscientiousness (Figure 4). The probability of planned
pregnancies was higher among women with higher agreeableness and lower
openness to experience. The probability of conceiving non-planned pregnancies was
higher among men with higher extraversion and lower agreeableness. The
probability of conceiving planned pregnancies was higher among men with higher
extraversion and conscientiousness and lower neuroticism and openness to
experience. These associations were very little affected by adding socio-economic
status to the models. When adjusting for socio-economic status, the effects of
agreeableness on non-planned pregnancies and conscientiousness on planned
52
pregnancies in men were no longer statistically significant (HR = 0.94, p = .052; and
HR = 1.03, p = .068, respectively), whereas the adjustment rendered the effect of
neuroticism on planned pregnancies in women significant (HR = 0.97, p = .031).
In addition, the timing effects between personality and planned and non-planned
childbearing were investigated. Figure 5 illustrates how the strength of associations
between adulthood personality traits and the number of children born from planned
and non-planned pregnancies had changed between ages 15 and 33 (the regression
coefficients before age 15 were all zero and are therefore omitted from the figure).
The associations that were present tended to increase in strength with time. In
women, the associations between personality traits and planned and non-planned
pregnancies were evident from early adolescence while in men, these associations
started to reach significant levels only from around age 20 onwards.
53
Figure 4. Associations between personality and planned and non-planned pregnancies. Hazard
ratios (HR) and 95 % confidence intervals obtained from multiple-event Cox regressions predicting
the occurrence of planned and non-planned pregnancies. Personality traits are standardised and
entered simultaneously. E = extraversion; N = neuroticism; A = agreeableness; C = conscientiousness;
O = openness to experience.
54
Figure 5. Difference in the number of planned (solid lines) and non-planned (dashed lines) children
associated with one standard deviation increment in personality trait scores at different ages. The
lines illustrate regression coefficients (betas) for personality trait scores in separate linear regressions
fitted at each age (separately for men and women) predicting the number of live children from
planned and non-planned pregnancies the participants had had up to that age, with all personality
traits entered simultaneously in the models. Error bars are 95% confidence intervals.
55
5 Discussion
5.1 Personality, reproductive success, and fluctuating
environments
With three distinct samples from contemporary Western populations, this thesis
provides further evidence on the associations between personality traits and fitness.
First, personality traits were consistently associated with reproductive success, but
in accordance with previous studies on the subject, the specific associations varied
between samples (see Figure 1 and section 1.1.). Notably, extraversion increased
number of children in all three sub-studies (although only in men in the Finnish twin
sample). The positive association between extraversion and number of offspring is
the most consistent finding in studies examining such associations in humans
(Penke & Jokela, 2016). While some studies have not found an association between
extraversion and number of children, especially in women (e.g., Alvergne et al.,
2010; Nettle, 2005), only one study has found a negative association between an
extraversion-related personality trait and number of children in humans (Jokela et
al., 2010). Jokela’s and colleagues’ (2010) study employed novelty seeking, a
temperament trait from Cloninger’s model (Cloninger et al., 1993) which is similar
to extraversion. The study was conducted on a Finnish sample from circa the same
period as the Finnish twin sample used in sub-study I of the present thesis, where
extraversion increased the number of children. The contradicting results in these
studies might indicate that novelty seeking captures different aspects of heritable
personality variation than extraversion. In addition, despite the overall negative
association between novelty seeking and number of children in Jokela’s and
colleagues’ study (2010), novelty seeking increased the number of children in men
who were not in a relationship.
Taken together, it seems that more extraverted people, especially men,
consistently have more children in all kinds of environments, be it modernised
Western populations or pre-industrial societies with high mortality and high fertility.
It should be noted, however, that studies in collectivistic Asian cultures are lacking
56
and would provide important insights into the fitness consequences of extraversion.
This points to extraversion being under directional selection, so that higher
extraversion is favoured (see also Penke & Jokela, 2016). The consistent pattern of
natural selection favouring higher extraversion implies that there would be an
adaptive trait, “being extraverted”, that is constantly being corrupted by new
mutations. The more mutational load an individual possesses, the less extraverted
they are, and the lower is their fitness. In other words, genetic variation in
extraversion could be maintained by mutation–selection balance. In addition,
extraversion shows signs of a genetic architecture that is in line with being under
mutation–selection balance. Firstly, as personality traits are most likely affected by
very large quantities of genes (see section 1.2.), they have a large mutational target
size. Secondly, a large proportion of the genetic variance in extraversion seems to be
driven by uncommon polymorphism (i.e., recently occurred mutations; Power &
Pluess, 2015; Verweij et al., 2012). In their recent analyses of the genetic architecture
of Cloninger’s personality traits, Verweij et al. (2012) came to the same conclusion
that with a high proportion of rare polymorphisms, novelty seeking is most likely to
be under mutation–selection balance. Further support for the mutation–selection
balance hypothesis was gained by analyses that showed signs of inbreeding
depression (Verweij et al., 2012). That is, offspring of more closely related
individuals were lower on novelty seeking – possibly because of higher accumulation
of deleterious mutations (Verweij et al., 2012). This finding is contradicted, however,
by at least three studies on non-human animals showing signs of assortative mating
on personality improving fitness (Both, Dingemanse, Drent, & Tinbergen, 2005;
Dingemanse et al., 2004; Sinn et al., 2006). Pairs of birds (great tits) with similar
levels of “exploration” had more offspring in both ends of the exploration continuum
(Both et al., 2005; Dingemanse et al., 2004). And, bold female squids had highest
probability of producing fertilised eggs after copulating with bold male squids as
opposed to shy male squids (Sinn et al., 2006).
Regarding the other four central personality traits, the results of the present thesis
confirm previous findings: the associations between neuroticism, agreeableness,
57
conscientiousness, and openness to experience and reproductive success can be
anything from negative, to no associations, to positive associations. The
inconsistency in the associations between these personality traits and reproduction
could reflect fluctuating selection through environmental heterogeneity. For
example, high neuroticism may be beneficial in an environment with a higher rate of
parasites or infectious diseases, as neuroticism may help being vigilant and avoiding
transmissions (see e.g., Alvergne et al., 2010; Barber & Dingemanse, 2010; Kortet,
Hedrick, & Vainikka, 2010). In environments that are less burdened with such risks
it may show no advantages, or even be detrimental to fitness (see also Nettle, 2006).
Furthermore, the fitness consequences of personality traits in humans seem to
differ not only between modern and developing societies, but also within Western
populations (see e.g., Penke & Jokela, 2016, and the three sub-studies of the present
thesis). Thus, at least in humans, the fitness consequences of personality may
depend not only on the large-scale environmental differences such as the presence
of parasites, but on more intricate differences between populations. For instance,
according to the frequency-dependent selection hypothesis, fitness consequences
could be related to the proportions of the population showing high amounts of some
personality trait (Wolf & McNamara, 2012). Furthermore, it has been proposed that
the sex-ratio, i.e., the proportion of men in a population, could provide grounds for
fluctuating selection pressures on personality (Del Giudice, 2012).
The evidence concerning environmental heterogeneity and fluctuating selection
on humans is, however, circumstantial. In non-humans animals, empirical
longitudinal studies have found fluctuating selection pressures on personality due to
changes in environment. Such effects have been found, for example, for predation
risk in mammals (Réale & Festa-Bianchet, 2003), complexity of habitat in fish
(Höjesjö et al., 2004), and food availability in birds (Dingemanse et al., 2004; Quinn
et al., 2009). It is plausible that environmental heterogeneity would be involved in
fitness consequences of personality in humans as well, since the phenomenon seems
to be present in such a wide variety of taxa. Nevertheless, the genetic architecture of
all personality traits, in addition to extraversion, seem to be closer to predictions
58
from mutation–selection balance rather than any form of balancing selection (Penke
et al., 2007). Future research will have to negotiate these contradicting findings.
Genetic variation in personality could also be maintained by stabilising or
disruptive selection. The associations between personality traits and lifetime
reproductive success in all three sub-studies of the present thesis were linear, which
is in line with previous studies in humans and non-human animals (see section 1.1.).
It may be that, for example, even the very extreme end of personality variation in
extraversion-type behaviour, currently defined as a personality disorder, is positively
associated with fitness (Gutierrez et al., 2013). In other words, no signs of stabilising
or disruptive selection have been detected in any studies examining the associations
between personality traits and fitness, including the present thesis. It is therefore
unlikely that such processes would be involved in maintaining genetic variation in
personality.
In sum, the phenotypic associations detected between personality traits and
fitness in the present thesis as well as in numerous previous studies on humans and
non-human animals, indicate that balancing selection is a plausible candidate in
explaining the persistent genetic variation in personality. The mechanisms may
include fluctuating selection through environmental heterogeneity, or other types of
balancing selection. In other words, genetic variation in personality could be
maintained because different ends of personality trait continuums are beneficial in
different environmental circumstances. The one notable exception to this conclusion
is extraversion: it seems to show signs of being under mutation–selection balance,
i.e., possessing one universal optimal value disrupted by mutations.
5.2 Personality and the phenotypic gambit
As outlined in sections 1.1. and 2.2., all previous work on the fitness consequences of
personality in humans, as well as non-human animals, has been conducted on the
phenotypic level. Further, inferences on the evolution of personality are often made
based on the phenotypic findings. If a personality trait is observed to correlate with
fitness, that personality trait is then assumed to undergo evolutionary changes, i.e.,
59
changes in the proportions of underlying genes across generations. As many theories
on the evolution of personality partly rely on conclusions based on this phenotypic
gambit (see e.g., Dingemanse & Wolf, 2010; Nettle, 2006; Penke et al., 2007), the
fact that no previous studies have examined its justification represents a serious
breach in the literature.
In sub-study I, the phenotypic and genetic approaches to (micro)evolution yielded
notably different predictions of the expected selection responses. Firstly,
extraversion was both phenotypically and genetically correlated with lifetime
reproductive success in men. However, the genetic approach of the Robertson–Price
identity yielded decidedly stronger estimates of the selection response than did
either the univariate or the multivariate forms of the breeder’s equation (see Figure
2). This indicates that some – possibly environmental – confounding factors may
attenuate the phenotypically observable associations between extraversion and
lifetime reproductive success. Neuroticism, on the other hand, was not associated
with lifetime reproductive success on either the phenotypic or the genetic level.
Therefore, on the basis of analyses using the univariate breeder’s equation, the
conclusion was that neuroticism was not subject to evolutionary change. However,
the multivariate genetic modelling detected a negative genetic correlation between
neuroticism and extraversion. Thus, the multivariate form of the breeder’s equation
yielded a predicted selection response for neuroticism, due to indirect selection
through extraversion. Furthermore, the Robertson–Price identity yielded a notably
stronger estimate of the predicted selection response on neuroticism. This was
because of carryover effects of the stronger genetic association between extraversion
and lifetime reproductive success compared to the phenotypic one.
Importantly, the results of sub-study I suggest that the associations between
personality traits and lifetime reproductive success are not environmentally
mediated. The presence of genetic covariance between personality and lifetime
reproductive success with the absence of environmental covariance between the two
is surprising, to say the least. Our results suggest that it is not neurotic or extraverted
behaviour per se that leads to differences in fertility, but the genetics underlying the
60
behaviour (see Turkheimer, Pettersson, & Horn, 2014). This is somewhat
counterintuitive because personality traits are associated with reproductive
behaviour. For example, extraversion is associated with a higher number of sexual
partners (Nettle, 2005), sexual risk behaviours such as lack of contraception (Hoyle,
Fejfar, & Miller, 2000), and it also seems to be associated with a higher risk of
unplanned pregnancies (see sub-study III of the present thesis). According to the
results of sub-study I, only insofar as these behavioural tendencies are
manifestations of the underlying genetic variation, they will be associated with
lifetime reproductive success.
Another way of expressing our findings is that personality as a behavioural
tendency is not being selected, but something that correlates genetically with
personality is. The associations between personality and lifetime reproductive
success might, for example, be mediated by common biological factors that underlie
personality variation and reproductive functions, regardless of behaviour. For
example, testosterone – the main male sex hormone – and extraversion seem to be
correlated (e.g., Alvergne, Jokela, Faurie, & Lummaa, 2010). A similar finding in
women has linked traits correlating with higher extraversion and lower neuroticism
with higher oestrogen levels (Ziomkiewicz, Wichary, Bochenek, Pawlowski, &
Jasienska, 2012). Accordingly, men and women with higher extraversion might be
more fecund, resulting in a higher percentage of copulations leading to pregnancy.
However, the evidence for common biological factors explaining the genetic
covariance between personality traits and lifetime reproductive success is too limited
to be evaluated more comprehensively, and further studies are needed both in
humans and other species.
One must bear in mind that the results of sub-study I are preliminary, since it is
the first to examine the differences between phenotypic and genetically informed
approaches to the evolution of personality. The results, however, warrant caution
when making inferences on the evolution of personality in humans (and most likely
in non-human animals, too) based on phenotypic data and analyses only. The
differences between the phenotypic and the genetic approach found here suggest
61
that studies relying solely on phenotypic data may lead not only to misestimation of
the magnitude of selection responses, but to misleading hypotheses on the evolution
of personality.
5.3 Personality and reproductive life-history trade-offs
Sub-study II directly assessed the possibility of personality being involved in a
reproductive quality–quantity trade-off in contemporary humans. Such quality–
quantity trade-offs have been proposed as one of the mechanisms maintaining
genetic variation in personality. Empirically, some findings indicative of such a
trade-off have been reported both in humans and non-human animals (see section
2.3.). However, no signs of a reproductive quality–quantity trade-off were detected
in sub-study II. Personality was associated with number of children and number of
grandchildren
in
a
similar
way.
With
neuroticism,
agreeableness,
and
conscientiousness, there were no statistically significant differences between the
effects of these personality traits on number of children and number of
grandchildren. Openness to experience had a smaller, but parallel, effect on number
of grand-children than on number of children, and extraversion was positively
associated with both, but only through the size of the first generation.
As a measure of long-term fitness, number of grandchildren takes into account
both the quantity of one’s offspring, and their ultimate “quality”, which is the
offspring’s lifetime reproductive success. According to the results of sub-study II,
personality seems not to be involved in quality–quantity trade-offs that would
nullify, let alone reverse, the seemingly directional selection in the long run – at least
not in a contemporary Western society. In other words, associations between
personality traits and fewer children did not improve the offspring’s “quality” in any
way that would have been beneficial for the parents’ long-term fitness. And vice
versa, associations between personality traits and a higher number of children did
not diminish the offspring’s “quality” in a manner that would have been detrimental
for the parents’ long-term fitness.
62
However, the results of sub-study II are not informative of parental behaviour.
Parental investment per child may still be greater in individuals with fewer children.
Previous studies have shown such patterns in 20th century Sweden (Goodman &
Koupil, 2009; Goodman, Koupil, & Lawson, 2012). For example, having fewer
descendants was associated with better socio-economic position of those
descendants throughout four generations, indicating higher parental investment
(Goodman et al., 2012). Nevertheless, having fewer children was also associated with
lower long-term fitness throughout the four generations, indicating that the
increased parental investment did not result in fitness pay-offs. Contemporary
Western societies, with advanced health care systems, good nutritional situation,
and various welfare programs may well have relaxed the quality–quantity trade-off.
However, there is some evidence that during the first phases of the demographic
transition to smaller family sizes, opting for a quality strategy would prove beneficial,
especially for people with a higher socio-economic status (Lawson, Alvergne, &
Gibson, 2012; Lawson & Mace, 2010). Despite the results of sub-study II, it is
plausible that personality traits could be involved in steering reproductive decisions
and parenting towards a “quantity” or a “quality” strategy. For example, parents with
high openness to experience may prefer a quality strategy (see Jokela et al., 2014),
even though the increased parental investment would not translate into better
reproductive success of children in modern environments. Personality variation, at
least in some traits, can thus have been maintained by reproductive quality–quantity
trade-offs in circumstances different from contemporary Western societies.
5.4 Temporal changes in the selection pressures on personality
In sub-study III, the aim was to elaborate the links between personality and lifetime
reproductive success by looking at the types of reproductive behaviour personality is
associated with. Associations between personality and respondents’ reports of
having had planned and non-planned pregnancies were examined and compared.
Overall, the associations were stronger in women than in men (see Figure 5). In
women, all personality traits had differential associations with planned vs. non-
63
planned pregnancies. The probability of non-planned pregnancies was higher among
women with higher extraversion and higher neuroticism, and lower among women
with higher conscientiousness. The probability of planned pregnancies, in turn, was
higher among women with higher agreeableness and lower among women with
higher openness to experience. In men, higher extraversion was associated with a
higher probability of having conceived both non-planned and planned pregnancies,
while conscientiousness was associated with neither. Other personality traits were,
again, differentially associated with planned and non-planned pregnancies. Higher
neuroticism and higher openness to experience were associated with a lower
probability of planned pregnancies. Higher agreeableness, in turn, was associated
with a lower probability of non-planned pregnancies.
We hypothesised that the patterns of associations between different personality
traits and planned and non-planned pregnancies are indicative of novel vs. older
selective pressures on personality. Thus extraversion, showing positive associations
with both non-planned and planned (in men) pregnancies, may be under similar,
positive directional selection in this contemporary Western population, as it has
been in the past. As mentioned above (see section 5.2.), higher extraversion is
associated with higher socio-sexuality (Banai & Pavela, 2015) and a higher
probability of sexually risky behaviour (Hoyle et al., 2000). Such behavioural
tendencies can explain the association between higher extraversion and higher
probability of non-planned pregnancies. In addition, the increase in planned
pregnancies associated with higher extraversion, and high agreeableness in women,
may stem from good mating quality and an ability to attract potential partners.
Individuals whose partners are higher on extraversion and agreeableness are more
satisfied with their relationship (Dyrenforth, Kashy, Donnellan, & Lucas, 2010).
Also, an ideal romantic partner seems to be more extraverted and agreeable than you
yourself are (Figueredo, Sefcek, & Jones, 2006). While these two studies were
conducted in Western contemporary samples, it is possible that such characteristics
would be valued in other socio-cultural environments as well.
64
Openness to experience and conscientiousness, in turn, seem to be involved in
behaviours that may induce novel selective pressures on these personality traits.
Openness to experience and conscientiousness both correlate with educational
motivation and achievement (Poropat, 2009). While the association between
openness and education is partly mediated by intelligence, conscientiousness seems
to have an effect on academic performance that is independent of intelligence
(Poropat, 2009). In addition, openness to experience correlates negatively with
traditional values such as conformity, and positively with more “modern” values,
such as self-directedness and stimulation (Parks-Leduc, Feldman, & Bardi, 2015). In
sub-study III, both women and men scoring high on openness to experience were
less likely to have planned pregnancies by the age of 33, which may indicate that they
prioritise other life goals than family formation. These other life goals could be
features of societal changes brought about by industrialisation, and more recently,
the second demographic transition such as career investments, leisure or selfrealisation (Lesthaeghe, 2010). Notably, such hypotheses are supported by the
finding that these personality traits have become important predictors of fertility
only later in the 20th century (Jokela, 2012).
Conscientiousness is likely to have additional associations with aspects of
contemporary post-industrial societies. This personality trait, besides reflecting a
tendency to plan ahead and be well prepared, has consistently been associated with
less frequent sexual risk behaviours in both sexes, especially lower probability of
unprotected sex and promiscuity (Hoyle et al., 2000; Schmitt & Shackelford, 2008).
Further, in accordance with the results of sub-study III, one previous study on
French women also found conscientiousness to be associated with a lower
probability of unplanned pregnancies (Bouchard, 2005). Thus, while the careful
nature of people high on conscientiousness may have helped to prevent unwanted
pregnancies in the past, this tendency could be amplified with modern contraceptive
methods. Why conscientiousness was not associated with a lower probability of nonplanned pregnancies in men presents a puzzle. Most long-term contraception
methods (e.g., the pill or intra-uterine device) require active behaviour of the
65
woman, which might make the female partner’s conscientiousness a more decisive
factor in preventing non-planned pregnancies. This finding may also reflect the
overall pattern found in sub-study III that at least in this cohort and population,
women’s personality was more strongly associated with reproductive behaviour than
men’s (see Figure 4).
These abovementioned hypotheses are preliminary and speculative, and the
associations between personality traits and family planning behaviour need to be
replicated in other studies. The results of sub-study III do, however, suggest that
similar evolutionary mechanisms and similar selective pressures are not working on
all personality traits. It is also important to note that modern environments may
provide grounds for acceleration as well as deceleration of evolutionary processes
(Hawks, Wang, Cochran, Harpending, & Moyzis, 2007). Human beings have not
ceased to evolve: heritable characteristics, such as personality, affect the way we
make use of novel socio-ecological environments and hence lead to fitness variation
associated with those traits.
5.5 Methodological considerations
All three sub-studies of the present thesis are performed on large, representative,
high quality datasets that had reliable personality measures on previously validated
personality models. Some limitations need to be considered, however. First, the
number of offspring was measured when the participants were relatively old (50–59
years in sub-study I, 55–102 years in sub-study II, and on average 44.8 years in substudy III), giving an almost exhaustive measure of lifetime reproductive success.
However, the reproductive lifespan of some participants may still not have ended –
particularly in sub-study III with the British data. The same applies to a larger extent
to sub-study II, examining the associations between personality and number of
grandchildren. Here, the participants were on average 67 years old, and the
reproductive lifespans of all their children cannot be considered completed. This
may have biased the associations between personality and number of offspring, as
some personality traits may be associated with earlier or later timing of childbearing
66
(e.g., Jokela et al., 2011). As far as personality would only have timing effects on
reproduction, for example, more conscientious people starting reproduction later
but eventually reaching the same quantity as the less conscientious, the results would
be misguided. However, previous studies have found similar associations when
using timing of parenthood and when using number of children as an outcome. For
instance, more extraverted people start having children earlier and end up with more
children (Jokela et al., 2011). In addition, according to the results of sub-study III of
this thesis, the associations between personality and reproduction seem to reach
significant levels early on in the reproductive career, and strengthen in time rather
than reversing at some point (see Figure 5).
Another limitation is the use of self-reported number of offspring in sub-studies
II and III with data form the US and the UK, which may be affected by selective
attrition of participants. Sub-study I with Finnish data, however, had register data
on the number of children, providing a highly accurate estimate of lifetime
reproductive success (the proportion of children without a known biological father
in population registers in Finland is only around 2 per cent; Kartiovaara and
Säkkinen 2007). The results concerning the association between extraversion and
lifetime reproductive success were similar in all three sub-studies, suggesting that
the self-reported number of offspring did not bias the results to a great extent.
A third limitation is that personality was measured after the start of childbearing
in all three sub-studies (although in the genetic analyses of sub-study I only 22 per
cent of women and 12 per cent of men had children at the time of personality
measurement). This introduces the possibility of reverse causality, i.e., that
childbearing affects personality and not vice versa. Parenthood has been associated
with increasing levels of personality tendencies that were present before parenthood,
for example, negative emotionality increasing in individuals with high negative
emotionality before childbearing (Jokela et al., 2009). More comprehensive studies
on parenthood and personality change appear to be lacking, so it is not yet possible
to evaluate which personality dimensions are affected the most, and how lasting
these changes are. However, the rank-order differences in personality have been
67
shown to be fairly stable throughout adulthood (Caspi, Roberts, & Shiner, 2005;
Roberts & DelVecchio, 2000). Also, at least one previous study suggested that the
prospective and retrospective associations between personality and having children
are not substantially different (Jokela et al., 2010). Additional studies with repeated
measurements of personality and family size are needed to examine the potential
bidirectional associations in detail.
One further important caveat relates to sub-study I and characteristics inherent
to the use of twin methodology. The phenotypic (co)variance patterns within and
between traits can arise due to many different kinds of genetic and environmental
influences. To circumvent this problem, the classical twin model makes
assumptions, and if these assumptions are violated, the estimates that the model
yields will be biased (see e.g., Rijsdijk & Sham, 2002). Firstly, there should be no
assortative mating for the trait studied as this could inflate the estimates of shared
environment. This assumption seems to hold for personality (Bouchard & Loehlin,
2001). Secondly, there should be no gene–environment interactions or gene–
environment correlations. If present, different types of gene–environment
interactions and correlations could inflate either the estimates of shared
environmental, genetic, or unique environmental effects (Purcell, 2002). Without
measured environmental influences, these effects are very difficult, if not impossible,
to be disentangled (Rijsdijk & Sham, 2002).
These assumptions may be more problematic in respect to personality traits.
Gene–environment interactions have been reported in some studies of personality
(Badcock et al., 2011; Reiner & Spangler, 2011). In addition, the findings from studies
examining the genetics of personality imply that gene–environment effects may play
a pivotal part in personality (see section 1.2.). Thus, it is impossible to say with the
data at hand whether such factors would inflate the effect of unique environment
and underestimate the genetic covariance between personality and w, or vice versa.
The relatively large point estimate of the genetic covariance between neuroticism
and w in women, with the upper limit of the 95% confidence interval just slightly
above zero (see Table 2), could be a sign of a true genetic covariance between
68
neuroticism and w. This may not have been statistically significantly detected
because of the data and methods at use.
Additionally, the classical twin model cannot separate additive and non-additive
genetic variance from each other. Several studies using extended twin designs have
reported substantial non-additivity in the genetics of personality in humans –
possibly accounting for as much as half of the overall genetic variation (see Chapter
1.2.). Strictly speaking, the Robertson–Price identity states that the selective
response equals the additive genetic covariance between a trait and relative fitness.
As the use of twin modelling only gives an estimate of the overall genetic covariance
between personality traits and fitness, the results of sub-study I must be considered
highly preliminary. Replicative studies on the genetic associations between
personality and fitness are needed, on other human and non-human populations,
and will hopefully shed more light on these matters.
5.6 Conclusions and theoretical implications
The results of the present thesis add to the growing body of empirical and theoretical
literature on the evolution of personality. Inferences can be drawn from two levels of
evidence that are present in all three sub-studies: the phenotypic level (i.e., the
manifest associations between personality traits and lifetime reproductive success)
and the genetic level (i.e., the underlying genetic associations between personality
traits and lifetime reproductive success).
Phenotypically, the results confirm previous findings obtained from preindustrial human populations (Alvergne et al., 2010; Gurven et al., 2014), Western
industrialised societies (e.g., Jokela et al., 2011), and non-human animals (Smith &
Blumstein, 2008) that personality traits are associated with fitness and thus under
natural selection. Sub-study III additionally suggests that modern postindustrialised environments pose novel selective pressures on personality traits,
which may influence the evolutionary future of personality in humans. Further, substudy II showed for the first time in any animal that personality traits are associated
with long-term fitness. This presents more compelling evidence on the evolutionary
69
consequences of the associations between personality traits and reproductive
success than merely showing associations within one generation. The results from
sub-study II are in concordance with a recent finding that there is a strong genetic
correlation between number of children and grandchildren in contemporary
societies (Zietsch, Kuja-Halkola, Walum, & Verweij, 2014). In other words, the
genetic effects that underlie the number of children are to a large extent the same
genetic effects that underlie the number of grandchildren. Thus, “counting babies”
seems to be an evolutionarily relevant fitness measure also in contemporary Western
humans (see also Goodman & Koupil, 2009).
The story of personality evolution on the genetic level is rather much more
complicated. First of all, the results of sub-study I – although very preliminary and
in desperate need of replication – suggest that the phenotypic gambit is not a safe
bet when considering the evolution of personality. The underlying genetic
associations between personality and fitness appear more complex than the
phenotypic correlation may suggest. As noted (see section 1.2.), the state of the art
in the genetics of personality is that we currently know very little. The results of substudy I suggests that it may not actually be personality differences in behaviour that
is under natural selection, but something else correlating genetically with
personality. This idea is further backed up by the finding in sub-study II that
personality traits are similarly associated with number of children and number of
grandchildren. This is most surprising, as one might expect that an individual’s
personality would be more important for the individual’s own fertility than for the
fertility of the individual’s children. A plausible, yet often neglected explanation for
intergenerational continuities would be common genetic factors that affect both
reproductive outcomes and personality (Zietsch et al., 2014). As outlined in section
5.2., personality traits seem to be associated with variation in reproductive hormonal
levels. Thus, the phenotypic associations between at least some personality traits and
lifetime reproductive success may not result from personality variation in behaviour,
but from differences in reproductive hormonal functioning that are related to
fecundity. Accordingly, the higher propensity of non-planned pregnancies in men
70
and women with higher extraversion in sub-study III could result not from
promiscuous behaviour in itself, but from a higher likelihood of an intercourse to
lead to pregnancy.
The idea that genes that affect personality variation may show pleiotropic effects
on other traits, which are the ones affecting fitness rather than personality, has also
been developed by Lukaszewski and von Rueden (2015). They hypothesise that
phenotypic variation in extraversion results from calibration to heritable, conditiondependent psycho-physiological features. Specifically, they state that individuals
who are more attractive, intelligent and able to pose threats on others (e.g., stronger)
would benefit more from extraverted behaviour than individuals lower on such
psycho-physiological traits. Thus, individuals should calibrate their personality to
match their condition. Further, they state that it is this heritable variation in
condition that is under selection, and not extraversion (Lukaszewski & von Rueden,
2015). Accordingly, in studies on laboratory rats and European rabbits, patterns of
higher explorative behaviour and higher boldness have been shown to be positively
associated with higher body mass index (Rödel & Meyer, 2011; Rödel & Monclús,
2011). In humans also, a significant proportion of variation in extraversion is
explained by variation in physical strength and (self-rated) attractiveness
(Lukaszewski & Roney, 2011). Extraversion may thus be under positive directional
selection because of its reactive association with overall phenotypic condition, which
is always positively selected (Lukaszewski & von Rueden, 2015). The hypothesis of
extraversion being calibrated to physiological features, however, fails to take into
account the fact that people simultaneously manifest other personality
characteristics as well. The presence of high neuroticism in an individual showing
high extraversion calibrated to physiological condition could, for instance, prove to
be a deleterious combination. Lukaszewski and von Rueden (2015) also do not seem
to take into account the findings linking reproductive hormonal functioning with
extraversion.
Nevertheless, extraversion seems to manifest fitness associations and features of
genetic architecture that attract intriguing evolutionary theoretical suggestions (see
71
also, e.g., Nettle, 2005), while other traits of the Five Factor Model have received less
attention.
Indeed,
the
results
concerning
neuroticism,
agreeableness,
conscientiousness, and openness to experience are harder to interpret from an
evolutionary point of view. Fitness associations seem to fluctuate quite randomly
across different populations (see e.g., all three sub-studies of the present thesis and
Penke & Jokela, 2016). In addition, corresponding traits especially for openness to
experience and conscientiousness are hard to find in non-human animal personality.
All in all, the results of the present thesis, alongside with other literature on the
subject, suggest that personality traits of the Five Factor Model are quite different in
terms of how their fitness consequences are affected by environments (see also
Penke & Jokela, 2016), or novel selective pressures (sub-study III), or how they are
genetically associated with fitness (sub-study I). Therefore they should probably not
be treated as five different manifestations of the same phenomenon called
“personality” – at least not when considering their evolutionary functions. Nettle
(2006) has proposed that personality traits of the Five Factor Model reflect solutions
to different demands posed by the social and physical environment that involve
trade-off decisions. For example, neuroticism might be beneficial in terms of
avoiding threats, but the increased vigilance of the stress-response system also
increases the probability of anxiety and depression, and poorer physical health. I
further suggest that the personality traits of the Five Factor Model may not only be
qualitatively similar answers to different environmental demands, but qualitatively
different phenomena that do not represent similar functions. For example,
extraversion could be a behavioural response calibrated to better overall phenotypic
condition (Lukaszewski & von Rueden, 2015), and perhaps also to better
reproductive fecundity (Alvergne et al., 2010; Ziomkiewicz et al., 2012), correlating
genetically with fitness (sub-study I) because of these underlying pleiotropic
features. Openness to experience, in turn, could be a manifestation of general
intelligence (Ackerman & Heggestad, 1997), whose by-product is negative fitness in
environments that offer vast amounts of opportunities to pursue intellectual
challenges (sub-study III).
72
This view of personality traits being qualitatively different phenomena is strongly
opposed to that of a general factor of personality. The idea that personality variation
would be driven by one higher order factor with a genetic basis, much like the general
factor of intelligence (the g), has been studied and promoted by Rushton and his
colleagues (e.g., Rushton, Bons, & Hur, 2008; Rushton et al., 2009). They suggest
that high extraversion, agreeableness, conscientiousness, and openness and low
neuroticism correlate to form a general factor of personality which has been under
positive unidirectional selection throughout the evolutionary history of humans
(Rushton et al., 2008). Their view has been greatly disputed, and the general factor
of personality may well be a statistical artefact, or result from reporting bias (e.g.,
Ashton, Lee, Goldberg, & de Vries, 2009; Bäckström, Björklund, & Larsson, 2009).
In addition, the empirical research concerning the fitness effects of personality in
humans or non-human animals does not support the idea of consistent
unidirectional selection. Furthermore, the persistent genetic variation in personality
across the phylogeny is a serious conundrum, if one type of personality should prove
beneficial in all circumstances. I consider the differing patterns of genetic, crossgenerational, and behavioural associations between personality traits and fitness
observed in the present thesis as further evidence against the hypothesis of a general
factor of personality.
In my opinion, future studies on the evolution of personality, be it theoretical or
empirical work, should not resort to averages across different personality traits, such
as reporting the average heritability of personality (see section 1.2.). Instead, they
should target the specific patterns of genetic architecture or fitness consequences of
each personality trait – taking into account the fact that other behavioural
tendencies are also simultaneously present in an individual. In addition,
evolutionary inferences based on phenotypic information on the fitness
consequences of personality should only be drawn with extreme caution, and,
whenever possible, associations should be examined on a genetic level.
Why and how are personality traits associated with fitness? The answer seems to
be that personality is not associated with fitness in any one way, or because of any
73
one evolutionary mechanism. From an evolutionary point of view, there might not
even be a phenomenon called “personality”, but a range of phenomena manifesting
as consistent behavioural tendencies. In their classic book of behaviour genetics,
Falconer and Mackay (1996, p. 348) state that “A complete understanding of how -metric traits are related to fitness -- will be achieved only when we know what the
primary functions of the [underlying] genes are”. With respect to personality, we
know very little, if anything, of the primary functions of the genes that are related to
personality variation. After decades of research on personality, we are currently
taking the very first steps into understanding it.
74
6 References
Ackerman, P. L., & Heggestad, E. D. (1997). Intelligence, personality, and interests: Evidence for
overlapping traits. Psychological Bulletin, 121, 219–245.
Adams, M. J. (2011). Evolutionary genetics of personality in nonhuman primates. In M. InoueMurayama, S. Kawamura, & A. Weiss (Eds.), From Genes to Animal Behavior (pp. 137–164).
Japan: Springer.
Adams, M. J., King, J. E., & Weiss, A. (2012). The majority of genetic variation in orangutan
personality and subjective well-being is nonadditive. Behavior Genetics, 42, 675–686.
Alvergne, A., Jokela, M., Faurie, C., & Lummaa, V. (2010). Personality and testosterone in men from
a high-fertility population. Personality and Individual Differences, 49, 840–844.
Alvergne, A., Jokela, M., & Lummaa, V. (2010). Personality and reproductive success in a highfertility human population. Proceedings of the National Academy of Sciences of the United
States of America, 107, 11745–11750.
Andersen, P. K., & Gill, R. D. (1982). Cox's regression model for counting processes: A large sample
study. The Annals of Statistics, 10, 1100–1120.
Ashton, M. C., Lee, K., Goldberg, L. R., & de Vries, R. E. (2009). Higher order factors of personality:
Do they exist? Personality and Social Psychology Review: An Official Journal of the Society for
Personality and Social Psychology, 13, 79–91.
Atherton, K., Fuller, E., Shepherd, P., Strachan, D. P., & Power, C. (2008). Loss and
representativeness in a biomedical survey at age 45 years: 1958 British birth cohort. Journal of
Epidemiology and Community Health, 62, 216–223.
Avia, M., Sanz, J., Sánchez-Bernardos, M., Martínez-Arias, M., Silva, F., & Graña, J. (1995). The
Five-Factor Model—II. Relations of the NEO-PI with other personality variables. Personality
and Individual Differences, 19, 81–97.
Bäckström, M., Björklund, F., & Larsson, M. R. (2009). Five-factor inventories have a major general
factor related to social desirability which can be reduced by framing items neutrally. Journal of
Research in Personality, 43, 335–344.
Badcock, P. B., Moore, E., Williamson, E., Berk, M., Williams, L. J., Bjerkeset, O., . . . Olsson, C. A.
(2011). Modeling gene-environment interaction in longitudinal data: Risk for neuroticism due to
interaction between maternal care and the dopamine 4 receptor gene (DRD4). Australian
Journal of Psychology, 63, 18–25.
Bailey, D. H., Walker, R. S., Blomquist, G. E., Hill, K. R., Hurtado, A. M., & Geary, D. C. (2013).
Heritability and fitness correlates of personality in the Ache, a natural-fertility population in
Paraguay. PloS One, 8, e59325.
75
Balestri, M., Calati, R., Serretti, A., & De Ronchi, D. (2014). Genetic modulation of personality traits:
A systematic review of the literature. International Clinical Psychopharmacology, 29, 1–15.
Banai, B., & Pavela, I. (2015). Two-dimensional structure of the sociosexual orientation inventory
and its personality correlates. Evolutionary Psychology, 13(3), 1–7.
Barber, I., & Dingemanse, N. J. (2010). Parasitism and the evolutionary ecology of animal
personality. Philosophical Transactions of the Royal Society B-Biological Sciences, 365, 4077–
4088.
Benjamin, J., Liz, L., Pattersonz, C., & Hamer, D. H. (1996). Association between the D4 dopamine
receptor gene and measures of novelty seeking. Nature Genetics, 12, 81–84.
Both, C., Dingemanse, N., Drent, P., & Tinbergen, J. (2005). Pairs of extreme avian personalities
have highest reproductive success. Journal of Animal Ecology, 74, 667–674.
Bouchard, G. (2005). Adult couples facing a planned or an unplanned pregnancy – Two realities.
Journal of Family Issues, 26, 619–637.
Bouchard, T. J., & Loehlin, J. C. (2001). Genes, evolution, and personality. Behavior Genetics, 31,
243–273.
Bouchard, T. J., & McGue, M. (2003). Genetic and environmental influences on human
psychological differences. Journal of Neurobiology, 54, 4–45.
Brent, L. J. N., Semple, S., MacLarnon, A., Ruiz-Lambides, A., Gonzalez-Martinez, J., & Platt, M. L.
(2014). Personality traits in rhesus macaques (Macaca Mulatta) are heritable but do not predict
reproductive output. International Journal of Primatology, 35, 188–209.
Carter, A. J., Feeney, W. E., Marshall, H. H., Cowlishaw, G., & Heinsohn, R. (2013). Animal
personality: What are behavioural ecologists measuring? Biological Reviews, 88, 465–475.
Caspi, A., Roberts, B., & Shiner, R. (2005). Personality development: Stability and change. Annual
Review of Psychology, 56, 453–484.
Chabris, C. F., Lee, J. J., Cesarini, D., Benjamin, D. J., & Laibson, D. I. (2015). The fourth law of
behavior genetics. Current Directions in Psychological Science, 24, 304–312.
Charmantier, A., & Garant, D. (2005). Environmental quality and evolutionary potential: Lessons
from wild populations. Proceedings of the Royal Society B-Biological Sciences, 272, 1415–1425.
Cleves, M. A., Gould, W. W., & Gutierrez, R. G. (2002). An introduction to survival analysis using
Stata. College Station, Texas, the US: Stata Press.
Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temperament
and character. Archives of General Psychiatry, 50, 975–990.
Costa, P. T., & McCrae, R. R. (1992). Four ways five factors are basic. Personality and Individual
Differences, 13, 653–665.
Crawford, C. (2000). Evolutionary psychology: Counting babies or studying informationǦprocessing
mechanisms. Annals of the New York Academy of Sciences, 907, 21–38.
76
de Moor, M. H., Costa, P. T., Terracciano, A., Krueger, R. F., De Geus, E. J., Toshiko, T., ...
Derringer, J. (2012). Meta-analysis of genome-wide association studies for personality.
Molecular Psychiatry, 17, 337–349.
de Moor, M. H., van den Berg, Stephanie M., Verweij, K. J., Krueger, R. F., Luciano, M., Vasquez, A.
A., ... Amin, N. (2015). Meta-analysis of genome-wide association studies for neuroticism, and
the polygenic association with major depressive disorder. JAMA Psychiatry, 72, 642–650.
Del Giudice, M. (2012). Sex ratio dynamics and fluctuating selection on personality. Journal of
Theoretical Biology, 297, 48–60.
Digman, J. (1990). Personality structure – emergence of the 5-factor model. Annual Review of
Psychology, 41, 417–440.
Dijkstra, P., & Barelds, D. P. H. (2009). Women's well-being: The role of individual differences.
Scandinavian Journal of Psychology, 50, 309–315.
Dingemanse, N. J., Both, C., Drent, P. J., & Tinbergen, J. M. (2004). Fitness consequences of avian
personalities in a fluctuating environment. Proceedings of the Royal Society of London BBiological Sciences, 271, 847–852.
Dingemanse, N. J., & Réale, D. (2005). Natural selection and animal personality. Behaviour, 142,
1159-1184.
Dingemanse, N. J., & Wolf, M. (2010). Recent models for adaptive personality differences: A review.
Philosophical Transactions of the Royal Society B-Biological Sciences, 365, 3947–3958.
DiRienzo, N., Niemelä, P. T., Skog, A., Vainikka, A., & Kortet, R. (2015). Juvenile pathogen exposure
affects the presence of personality in adult field crickets. Frontiers in Ecology and Evolution, 3,
1–10.
Dochtermann, N. A., & Roff, D. A. (2010). Applying a quantitative genetics framework to
behavioural syndrome research. Philosophical Transactions of the Royal Society B-Biological
Sciences, 365, 4013–4020.
Dochtermann, N. A., Schwab, T., & Sih, A. (2015). The contribution of additive genetic variation to
personality variation: heritability of personality. Proceedings of the Royal Society of London BBiological Sciences, 282, 20142201.
Draycott, S. G., & Kline, P. (1995). The big three or the big five — the EPQ-R vs the NEO-PI: A
research note, replication and elaboration. Personality and Individual Differences, 18, 801–804.
Dyrenforth, P. S., Kashy, D. A., Donnellan, M. B., & Lucas, R. E. (2010). Predicting relationship and
life satisfaction from personality in nationally representative samples from three countries: The
relative importance of actor, partner, and similarity effects. Journal of Personality and Social
Psychology, 99, 690–702.
Eaves, L. J., Martin, N. G., Heath, A. C., Hewitt, J. K., & Neale, M. C. (1990). Personality and
reproductive fitness. Behavior Genetics, 20, 563–568.
77
Eyre-Walker, A., & Keightley, P. D. (2007). The distribution of fitness effects of new mutations.
Nature Reviews Genetics, 8, 610–618.
Eysenck, H. J. (1967). The Biological Basis of Personality. Springfield, Illinois, the US: Charles C.
Thomas.
Falconer, D. S., & Mackay, T. F. C. (1996). Introduction to Quantitative Genetics (4th ed.). Harlow,
Essex, UK: Pearson Education Ltd.
Fidler, A. E., van Oers, K., Drent, P. J., Kuhn, S., Mueller, J. C., & Kempenaers, B. (2007). DRD4
gene polymorphisms are associated with personality variation in a passerine bird. Proceedings
of the Royal Society B-Biological Sciences, 274, 1685–1691.
Figueredo, A. J., Sefcek, J. A., & Jones, D. N. (2006). The ideal romantic partner personality.
Personality and Individual Differences, 41, 431–441.
Floderus, B. (1974). Psycho-social factors in relation to coronary heart disease and associated risk
factors. Nordisk Hygienisk Tidskrift, Suppl. 6.
Fuller, J. L. (1960). Behavior genetics. Annual Review of Psychology, 11, 41–70.
Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the
lowerlevel facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt & F.
Ostendorf (Eds.), Personality Psychology in Europe, vol. 7 (pp. 7–28). Tilburg, The
Netherlands: Tilburg University Press.
Goodman, A., & Koupil, I. (2009). Social and biological determinants of reproductive success in
Swedish males and females born 1915–1929. Evolution and Human Behavior, 30, 329–341.
Goodman, A., Koupil, I., & Lawson, D. W. (2012). Low fertility increases descendant socioeconomic
position but reduces long-term fitness in a modern post-industrial society. Proceedings of the
Royal Society B-Biological Sciences, 279, 4342–4351.
Gosling, S. D. (2001). From mice to men: What can we learn about personality from animal
research? Psychological Bulletin, 127, 45–86.
Gurven, M., von Rueden, C., Stieglitz, J., Kaplan, H., & Eid Rodriguez, D. (2014). The evolutionary
fitness of personality traits in a small-scale subsistence society. Evolution and Human Behavior,
35, 17–25.
Gurven, M., von Rueden, C., Massenkoff, M., Kaplan, H., & Lero Vie, M. (2013). How universal is
the Big Five? Testing the five-factor model of personality variation among forager–farmers in the
Bolivian Amazon. Journal of Personality and Social Psychology, 104, 354–370.
Gutierrez, F., Garriz, M., Peri, J. M., Ferraz, L., Sol, D., Blas Navarro, J., ... Valdes, M. (2013).
Fitness costs and benefits of personality disorder traits. Evolution and Human Behavior, 34, 41–
48.
78
Hawks, J., Wang, E. T., Cochran, G. M., Harpending, H. C., & Moyzis, R. K. (2007). Recent
acceleration of human adaptive evolution. Proceedings of the National Academy of Sciences of
the United States of America, 104, 20753–20758.
Hill, K., & Kaplan, H. (1999). Life history traits in humans: Theory and empirical studies. Annual
Review of Anthropology, 28, 397–430.
Hill, W. G. (2014). Applications of population genetics to animal breeding, from Wright, Fisher and
Lush to genomic prediction. Genetics, 196, 1–16.
Höjesjö, J., Johnsson, J., & Bohlin, T. (2004). Habitat complexity reduces the growth of aggressive
and dominant brown trout (Salmo Trutta) relative to subordinates. Behavioral Ecology and
Sociobiology, 56, 286–289.
Hollander, F. A., Van Overveld, T., Tokka, I., & Matthysen, E. (2008). Personality and nest defence
in the great tit (Parus Major). Ethology, 114, 405–412.
Hoyle, R. H., Fejfar, M. C., & Miller, J. D. (2000). Personality and sexual risk taking: A quantitative
review. Journal of Personality, 68, 1203–1231.
Jokela, M. (2012). Birth-cohort effects in the association between personality and fertility.
Psychological Science, 23, 835–841.
Jokela, M., Alvergne, A., Pollet, T. V., & Lummaa, V. (2011). Reproductive behavior and personality
traits of the Five factor model. European Journal of Personality, 25, 487–500.
Jokela, M., Alvergne, A., Rotkirch, A., Rickard, I. J., & Lummaa, V. (2014). Associations between
family size and offspring education depend on aspects of parental personality. Personality and
Individual Differences, 58, 95–100.
Jokela, M., Hintsa, T., Hintsanen, M., & Keltikangas-Järvinen, L. (2010). Adult temperament and
childbearing over the life course. European Journal of Personality, 24, 151–166.
Jokela, M., & Keltikangas-Järvinen, L. (2009). Adolescent leadership and adulthood fertility:
Revisiting the “central theoretical problem of human sociobiology”. Journal of Personality, 77,
213–230.
Jokela, M., Kivimäki, M., Elovainio, M., & Keltikangas-Järvinen, L. (2009). Personality and having
children: A two-way relationship. Journal of Personality and Social Psychology, 96, 218-230.
Jokela, M., Pulkki-Råback, L., Elovainio, M., & Kivimäki, M. (2014). Personality traits as risk factors
for stroke and coronary heart disease mortality: Pooled analysis of three cohort studies. Journal
of Behavioral Medicine, 37, 881–889.
Jones, J. H., & Bird, R. B. (2014). The marginal valuation of fertility. Evolution and Human
Behavior, 35, 65–71.
Juster, F. T., & Suzman, R. (1995). An overview of the Health and Retirement Study. Journal of
Human Resources, 30(XXX), S7–S56.
79
Kaprio, J., & Koskenvuo, M. (2002). Genetic and environmental factors in complex diseases: The
Older Finnish Twin Cohort. Twin Research, 5, 358–365.
Kendler, K. S., Gatz, M., Gardner, C. O., & Pedersen, N. L. (2006). Personality and major depression
– A Swedish longitudinal, population-based twin study. Archives of General Psychiatry, 63,
1113–1120.
Korsten, P., Mueller, J. C., Hermannstädter, C., Bouwman, K. M., Dingemanse, N. J., Drent, P. J., ...
van Overveld, T. (2010). Association between DRD4 gene polymorphism and personality
variation in great tits: A test across four wild populations. Molecular Ecology, 19, 832–843.
Kortet, R., Hedrick, A. V., & Vainikka, A. (2010). Parasitism, predation and the evolution of animal
personalities. Ecology Letters, 13, 1449–1458.
Koski, S. E. (2011). How to measure animal personality and why does it matter? Integrating the
psychological and biological approaches to animal personality. In M. Inoue-Murayama, S.
Kawamura, & A. Weiss (eds.), From Genes to Animal Behavior (pp. 115–136). Japan: Springer.
Krueger, R. F., South, S., Johnson, W., & Iacono, W. (2008). The heritability of personality is not
always 50%: Gene–environment interactions and correlations between personality and
parenting. Journal of Personality, 76, 1485–1521.
Kruuk, L. E. B., Slate, J., Pemberton, J. M., Brotherstone, S., Guinness, F., & Clutton-Brock, T.
(2002). Antler size in red deer: Heritability and selection but no evolution. Evolution, 56, 1683–
1695.
Lachman, M. E., & Weaver, S. L. (1997). The midlife development inventory (MIDI) personality
scales: Scale construction and scoring. Waltham, MA, US: Brandeis University, Psychology
Department.
Lande, R., & Arnold, S. J. (1983). The measurement of selection on correlated characters. Evolution,
37, 1210–1226.
Lawson, D. W., Alvergne, A., & Gibson, M. A. (2012). The life-history trade-off between fertility and
child survival. Proceedings of the Royal Society B-Biological Sciences, 279, 4755–4764.
Lawson, D. W., & Mace, R. (2010). Optimizing modern family size. Human Nature – An
Interdisciplinary Biosocial Perspective, 21, 39–61.
Lee, R. (2003). The demographic transition: three centuries of fundamental change. The Journal of
Economic Perspectives, 17, 167–190.
Lee, W., Wadsworth, M., & Hotopf, M. (2006). The protective role of trait anxiety: A longitudinal
cohort study. Psychological Medicine, 36, 345–351.
Lesch, K. P., Bengel, D., Heils, A., Sabol, S. Z., Greenberg, B. D., Petri, S., ... Murphy, D. L. (1996).
Association of anxiety-related traits with a polymorphism in the serotonin transporter gene
regulatory region. Science, 274, 1527–1531.
80
Lesthaeghe, R. (2010). The unfolding story of the second demographic transition. Population and
Development Review, 36, 211–251.
Lukaszewski, A. W., & Roney, J. R. (2011). The origins of extraversion: Joint effects of facultative
calibration and genetic polymorphism. Personality and Social Psychology Bulletin, 37, 409–
421.
Lukaszewski, A. W., & von Rueden, C. R. (2015). The extraversion continuum in evolutionary
perspective: A review of recent theory and evidence. Personality and Individual Differences, 77,
186–192.
MacDonald, M. E., Ambrose, C. M., Duyao, M. P., Myers, R. H., Lin, C., Srinidhi, L., ... Groot, N.
(1993). A novel gene containing a trinucleotide repeat that is expanded and unstable on
Huntington's disease chromosomes. Cell, 72, 971–983.
Malouff, J., Thorsteinsson, E., & Schutte, N. (2005). The relationship between the Five-Factor
Model of personality and symptoms of clinical disorders: A meta-analysis. Journal of
Psychopathology and Behavioral Assessment, 27, 101–114.
Massen, J. J., Antonides, A., Arnold, A. K., Bionda, T., & Koski, S. E. (2013). A behavioral view on
chimpanzee personality: Exploration tendency, persistence, boldness, and toolǦorientation
measured with group experiments. American Journal of Primatology, 75, 947–958.
McCrae, R. R. (2002). Cross-cultural research on the Five-Factor Model of personality. Online
Readings in Psychology and Culture, 4(4).
Meagher, R. K., Bechard, A., & Mason, G. J. (2012). Mink with divergent activity levels have
divergent reproductive strategies. Ethology, 118, 543–554.
Mealey, L., & Segal, N. L. (1993). Heritable and environmental variables affect reproduction-related
behaviors, but not ultimate reproductive success. Personality and Individual Differences, 14,
783–794.
Merilä, J., Kruuk, L., & Sheldon, B. (2001). Cryptic evolution in a wild bird population. Nature, 412,
76–79.
Merilä, J., Sheldon, B., & Kruuk, L. (2001). Explaining stasis: Microevolutionary studies in natural
populations. Genetica, 112, 199–222.
Merilä, J., & Sheldon, B. C. (1999). Genetic architecture of fitness and nonfitness traits: Empirical
patterns and development of ideas. Heredity, 83, 103–109.
Montiglio, P. O., Ferrari, C., & Réale, D. (2013). Social niche specialization under constraints:
Personality, social interactions and environmental heterogeneity. Philosophical Transactions of
the Royal Society of London B - Biological Sciences, 368, 20120343.
Morrissey, M., Kruuk, L., & Wilson, A. (2010). The danger of applying the breeder's equation in
observational studies of natural populations. Journal of Evolutionary Biology, 23, 2277–2288.
81
Munafò, M. R., & Flint, J. (2011). Dissecting the genetic architecture of human personality. Trends
in Cognitive Sciences, 15, 395–400.
Munafò, M. R., Freimer, N. B., Ng, W., Ophoff, R., Veijola, J., Miettunen, J., . . . Flint, J. (2009). 5Ǧ
HTTLPR genotype and anxietyǦrelated personality traits: A metaǦanalysis and new data.
American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 150, 271–281.
Munafò, M. R., Yalcin, B., Willis-Owen, S. A., & Flint, J. (2008). Association of the dopamine D4
receptor (DRD4) gene and approach-related personality traits: Meta-analysis and new data.
Biological Psychiatry, 63, 197–206.
Neale, M. C., & Maes, H. H. M. (2003). Methodology for genetic studies of twins and families.
Dordrecht, The Netherlands: Kluwer, Academic Publishers B. V.
Nettle, D. (2005). An evolutionary approach to the extraversion continuum. Evolution and Human
Behavior, 26, 363–373.
Nettle, D. (2006). The evolution of personality variation in humans and other animals. American
Psychologist, 61, 622–631.
Nettle, D., & Penke, L. (2010). Personality: Bridging the literatures from human psychology and
behavioural ecology. Philosophical Transactions of the Royal Society of London B - Biological
Sciences, 365, 4043–4050.
NHGRI (2005). National Human Genome Resource Institute: “International Consortium Completes
Map of Human Genetic Variation,” National Institutes of Health News, October 2005, available
at <http://www.genome.gov/17015412> (visited March 29, 2016).
Niemelä, P., Lehtinen, P., Rauramo, L., Hermansson, R., Karjalainen, R., Mäki, H., & StorÅ, C.
(1981). The first abortion – and the last? A study of the personality factors underlying repeated
failure of contraception. International Journal of Gynecology & Obstetrics, 19, 193–200.
Parks-Leduc, L., Feldman, G., & Bardi, A. (2015). Personality traits and personal values: A metaanalysis. Personality and Social Psychology Review, 19, 3–29.
Penke, L., & Jokela, M. (2016). The evolutionary genetics of personality revisited. Current Opinion
in Psychology, 7, 104-109.
Penke, L., Denissen, J. J. A., & Miller, G. F. (2007). Evolution, genes, and inter-disciplinary
personality research. European Journal of Personality, 21, 639-665.
Perkins, A. M., Cserjesi, R., Ettinger, U., Kumari, V., Martin, N. G., & Arden, R. (2013). Personality
and occupational markers of ‘solid citizenship’ are associated with having fewer children.
Personality and Individual Differences, 55, 871–876.
Polderman, T. J., Benyamin, B., de Leeuw, C. A., Sullivan, P. F., van Bochoven, A., Visscher, P. M., &
Posthuma, D. (2015). Meta-analysis of the heritability of human traits based on fifty years of
twin studies. Nature Genetics, 47, 702–709.
82
Poropat, A. E. (2009). A meta-analysis of the Five-Factor Model of personality and academic
performance. Psychological Bulletin, 135, 322–338.
Poropat, A. E. (2014). A meta-analysis of adult-rated child personality and academic performance in
primary education. British Journal of Educational Psychology, 84, 239–252.
Posthuma, D., & Boomsma, D. I. (2000). A note on the statistical power in extended twin designs.
Behavior Genetics, 30, 147–158.
Power, R., & Pluess, M. (2015). Heritability estimates of the Big Five personality traits based on
common genetic variants. Translational Psychiatry, 5, e604.
Power, C., & Elliott, J. (2006). Cohort profile: 1958 British birth cohort (National Child
Development Study). International Journal of Epidemiology, 35, 34–41.
Purcell, S. (2002). Variance components models for gene-environment interaction in twin analysis.
Twin Research, 5, 554–571.
Quinn, J. L., Patrick, S. C., Bouwhuis, S., Wilkin, T. A., & Sheldon, B. C. (2009). Heterogeneous
selection on a heritable temperament trait in a variable environment. Journal of Animal
Ecology, 78, 1203–1215.
Rausher, M. D. (1992). The measurement of selection on quantitative traits: Biases due to
environmental covariances between traits and fitness. Evolution, 46, 616–626.
Réale, D., Reader, S. M., Sol, D., McDougall, P. T., & Dingemanse, N. J. (2007). Integrating animal
temperament within ecology and evolution. Biological Reviews, 82, 291–318.
Réale, D., & Festa-Bianchet, M. (2003). Predator-induced natural selection on temperament in
bighorn ewes. Animal Behaviour, 65, 463–470.
Reiner, I., & Spangler, G. (2011). Dopamine D4 receptor exon III polymorphism, adverse life events
and personality traits in a nonclinical German adult sample. Neuropsychobiology, 63, 52–58.
Reis, O., Doernte, M., & von der Lippe, H. (2011). Neuroticism, social support, and the timing of
first parenthood: A prospective study. Personality and Individual Differences, 50, 381–386.
Richardson, M., Abraham, C., & Bond, R. (2012). Psychological correlates of university students'
academic performance: A systematic review and meta-analysis. Psychological Bulletin, 138,
353–387.
Rijsdijk, F. V., & Sham, P. C. (2002). Analytic approaches to twin data using structural equation
models. Briefings in Bioinformatics, 3, 119–133.
Roberts, B., & Del Vecchio, W. (2000). The rank-order consistency of personality traits from
childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126,
3–25.
Rödel, H. G., & Meyer, S. (2011). Early development influences ontogeny of personality types in
young laboratory rats. Developmental Psychobiology, 53, 601–613.
83
Rödel, H. G., & Monclús, R. (2011). Long-term consequences of early development on personality
traits: A study in European rabbits. Behavioral Ecology, 22, 1123–1130.
Rolland, J. (2002). The cross-cultural generalizability of the Five-Factor Model of personality. In R.
R. McCrae & J. Allik (Eds.), The Five Factor Model of Personality Across Cultures (pp.7–28).
New York, the US: Kluwer Academic/Publishers.
Rose, R. J., Kaprio, J., & Koskenvuo, M. (1988). Genetic contributions to behavioral consistency and
change – data from a 6-year follow-up of the Finnish Twin Cohort. Behavior Genetics, 18, 730–
730.
Rotkirch, A. (2007). All that she wants is a(nother) baby’? Longing for children as a fertility
incentive of growing importance. Journal of Evolutionary Psychology, 5, 89–104.
Rushton, J. P., Bons, T. A., Ando, J., Hur, Y., Irwing, P., Vernon, P. A., ... Barbaranelli, C. (2009). A
general factor of personality from multitrait–multimethod data and cross-national twins. Twin
Research and Human Genetics, 12, 356–365.
Rushton, J. P., Bons, T. A., & Hur, Y. (2008). The genetics and evolution of the general factor of
personality. Journal of Research in Personality, 42, 1173–1185.
Sandefur, G. D., Martin, M., Eggerling-Boeck, J., Mannon, S. E., & Meier, A. M. (2001). An overview
of racial and ethnic demographic trends. In N. J. Smelser, W. J. Wilson & F. Mitchell (Eds.),
America becoming: Racial trends and their consequences, volume 1 (pp. 40–102). Washington,
D.C., US: National Academy Press.
Sarna, S., Kaprio, J., Sistonen, P., & Koskenvuo, M. (1978). Diagnosis of twin zygosity by mailed
questionnaire. Human Heredity, 28, 241–254.
Schmitt, D. P., & Shackelford, T. K. (2008). Big Five traits related to short-term mating: From
personality to promiscuity across 46 nations. Evolutionary Psychology, 6, 246–282.
Service, S., Verweij, K., Lahti, J., Congdon, E., Ekelund, J., Hintsanen, M., ... Widen, E. (2012). A
genome-wide meta-analysis of association studies of Cloninger's temperament scales.
Translational Psychiatry, 2, e116.
Shipley, B. A., Weiss, A., Der, G., Taylor, M. D., & Deary, I. J. (2007). Neuroticism, extraversion,
and mortality in the UK health and lifestyle survey: A 21-year prospective cohort study.
Psychosomatic Medicine, 69, 923–931.
Sih, A., Bell, A. M., Johnson, J. C., & Ziemba, R. E. (2004). Behavioral syndromes: An integrative
overview. Quarterly Review of Biology, 79, 241–277.
Sinn, D. L., Apiolaza, L. A., & Moltschaniwskyj, N. A. (2006). Heritability and fitness-related
consequences of squid personality traits. Journal of Evolutionary Biology, 19, 1437–1447.
Skirbekk, V. (2008). Fertility trends by social status. Demographic Research, 18, 145–180.
Skirbekk, V., & Blekesaune, M. (2014). Personality traits increasingly important for male fertility:
Evidence from Norway. European Journal of Personality, 28, 521–529.
84
Smith, B. R., & Blumstein, D. T. (2008). Fitness consequences of personality: A meta-analysis.
Behavioral Ecology, 19, 448–455.
Spengler, M., Luedtke, O., Martin, R., & Brunner, M. (2013). Personality is related to educational
outcomes in late adolescence: Evidence from two large-scale achievement studies. Journal of
Research in Personality, 47, 613–625.
Stearns, S. C. (1986). Natural selection and fitness, adaptation and constraint. In D. M. Raup & D.
Jablonski (Eds.), Patterns and Processes in the History of Life (pp. 23–44). Berlin, Germany:
Springer.
Stearns, S. C., Byars, S. G., Govindaraju, D. R., & Ewbank, D. (2010). Measuring selection in
contemporary human populations. Nature Reviews Genetics, 11, 611–622.
Stelmack, R. M. (1991). Advances in personality theory and research. Journal of Psychiatry &
Neuroscience, 16, 131–138.
Stone, E. A., Shackelford, T. K., & Buss, D. M. (2012). Is variability in mate choice similar for
intelligence and personality traits? Testing a hypothesis about the evolutionary genetics of
personality. Intelligence, 40, 33–37.
Tooby, J., & Cosmides, L. (1990). On the universality of human-nature and the uniqueness of the
individual – the role of genetics and adaptation. Journal of Personality, 58, 17–67.
Trivers, R. (1972). Parental investment and sexual selection. In B. Campbell (Ed.), Sexual Selection
and the Descent of Man 1871–1971, (pp. 136–179). Chicago, the US: Aldine Publishing Company.
Turiano, N. A., Chapman, B. P., Gruenewald, T. L., & Mroczek, D. K. (2015). Personality and the
leading behavioral contributors of mortality. Health Psychology, 34, 51–60.
Turkheimer, E., Pettersson, E., & Horn, E. E. (2014). A phenotypic null hypothesis for the genetics
of personality. Annual Review of Psychology, 65, 515–540.
van den Berg, S. M, de Moor, M. H., McGue, M., Pettersson, E., Terracciano, A., Verweij, K. J., ...
van Grootheest, G. (2014). Harmonization of neuroticism and extraversion phenotypes across
inventories and cohorts in the genetics of personality consortium: An application of item
response theory. Behavior Genetics, 44, 295-313.
van den Berg, S. M., de Moor, M. H., Verweij, K. J., Krueger, R. F., Luciano, M., Vasquez, A. A., ... &
Gordon, S. D. (2016). Meta-analysis of Genome-Wide Association Studies for Extraversion:
Findings from the Genetics of Personality Consortium. Behavior genetics, 46, 170-182.
van der Linden, D., Figueredo, A. J., de Leeuw, R. N. H., Scholte, R. H. J., & Engels, R. C. M. E.
(2012). The general factor of personality (GFP) and parental support: Testing a prediction from
life history theory. Evolution and Human Behavior, 33, 537–546.
van Oers, K., Drent, P., de Jong, G., & van Noordwijk, A. (2004). Additive and nonadditive genetic
variation in avian personality traits. Heredity, 93, 496–503.
85
van Oers, K., & Sinn, D. L. (2011). Toward a basis for the phenotypic gambit: Advances in the
evolutionary genetics of animal personality. In M. Inoue-Murayama, S. Kawamura, & A. Weiss
(eds.), From Genes to Animal Behavior (pp. 165–183). Japan: Springer.
van Tienderen, P. H., & de Jong, G. (1994). A general-model of the relation between phenotypic
selection and genetic response. Journal of Evolutionary Biology, 7, 1–12.
Verweij, K. J. H., Yang, J., Lahti, J., Veijola, J., Hintsanen, M., Pulkki-Råback, L., . . . Zietsch, B. P.
(2012). Maintenance of genetic variation in human personality: Testing evolutionary models by
estimating heritability due to common causal variants and investigating the effect of distant
inbreeding. Evolution, 66, 3238–3251.
Vinkhuyzen, A. A., Pedersen, N. L., Yang, J., Lee, S. H., Magnusson, P. K., Iacono, W. G., . . .
Luciano, M. (2012). Common SNPs explain some of the variation in the personality dimensions
of neuroticism and extraversion. Translational Psychiatry, 2, e102.
Vukasovic, T., & Bratko, D. (2015). Heritability of personality: A meta-analysis of behavior genetic
studies. Psychological Bulletin, 141, 769–785.
Williams, R. L. (2000). A note on robust variance estimation for cluster-correlated data. Biometrics,
56, 645–646.
Wolf, M., van Doorn, G. S., Leimar, O., & Weissing, F. J. (2007). Life-history trade-offs favour the
evolution of animal personalities. Nature, 447, 581–584.
Wolf, M., & McNamara, J. M. (2012). On the evolution of personalities via frequency-dependent
selection. American Naturalist, 179, 679–692.
Yamagata, S., Suzuki, A., Ando, J., Ono, Y., Kijima, N., Yoshimura, K., . . . Spinath, F. M. (2006). Is
the genetic structure of human personality universal? A cross-cultural twin study from North
America, Europe, and Asia. Journal of Personality and Social Psychology, 90, 987–998.
Zietsch, B. P., Kuja-Halkola, R., Walum, H., & Verweij, K. J. H. (2014). Perfect genetic correlation
between number of offspring and grandoffspring in an industrialized human population.
Proceedings of the National Academy of Sciences of the United States of America, 111, 1032–
1036.
Ziomkiewicz, A., Wichary, S., Bochenek, D., Pawlowski, B., & Jasienska, G. (2012). Temperament
and ovarian reproductive hormones in women: Evidence from a study during the entire
menstrual cycle. Hormones and Behavior, 61, 535–540.
86