Professional Documents
Culture Documents
Knafonoam 2017
Knafonoam 2017
PII: S2352-250X(17)30162-8
DOI: http://dx.doi.org/doi:10.1016/j.copsyc.2017.08.013
Reference: COPSYC 538
To appear in:
Please cite this article as: A. Knafo-Noam, D. Vertsberger, S. Israel, Genetic and
Environmental Contributions to Children’s Prosocial Behavior: Brief Review and
New Evidence from a Reanalysis of Experimental Twin Data, COPSYC (2017),
http://dx.doi.org/10.1016/j.copsyc.2017.08.013
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
• Twin research shows, a genetic contribution to variability in children’s prosocial behavior
• Genetic effects are found across measures and cultures
• The association between sharing and comforting reflects overlapping genetic effects
• Molecular genetic evidence using genome-wide and polygenic measures is needed
• Gene-environment interaction and gene-environment correlations are reviewed
t
ip
cr
us
an
M
d
p te
ce
Ac
Page 1 of 25
Genetic and Environmental Contributions to Children’s Prosocial Behavior: Brief
t
ip
Ariel Knafo-Noam
cr
Dana Vertsberger
us
Salomon Israel
an
Psychology Department, The Hebrew University of Jerusalem
M
e d
pt
ce
Ac
Page 2 of 25
Abstract
Children’s prosocial behaviors show considerable variability. Here we discuss the genetic
t
and environmental contributions to individual differences in children’s prosocial
ip
behavior. Twin research systematically shows, at least from the age of 3 years, a genetic
cr
contribution to individual differences in prosocial behavior, both questionnaire-based and
observed. This finding is demonstrated across a wide variety of cultures. We discuss the
us
possibility that different prosocial behaviors have different genetic etiologies. A re-
an
analysis of past twin data shows that sharing and comforting are affected by overlapping
genetic factors at age 3.5 years. In contrast, the association between helping and
M
comforting is attributed to environmental factors. The few molecular genetic studies of
children’s prosocial behavior are reviewed, and we point out genome-wide and polygenic
d
methods as a key future direction. Finally, we discuss the interplay of genetic and
e
environment correlations.
ce
Ac
Page 3 of 25
Introduction
Children start helping, sharing with, and comforting others at a very early age[1,
2]. When voluntary and intended to benefit others, such behaviors are referred to as
t
prosocial behaviors[3]. Despite this early emergence, prosocial behaviors also show
ip
considerable variation; some children willingly sacrifice personal resources and help
cr
others while some children do not. Studies have linked individual differences to many
us
factors, including parenting, peers, school variables, and temperament[3,4]. Here we
(see [6] for a broader discussion of morality genetics in children and adults).
M
Quantitative Genetic Designs
share virtually 100% of their genes, and dizygotic (DZ) twins, who share on average 50%
pt
of the genetic variance. Assuming both twin types received equally similar environments,
ce
while twin differences not due to genetic differences indicate the contribution of non-
shared environmental effects and measurement error. Importantly, these estimates are
contingent upon the context, culture, age and population from which they are derived.
Twin studies in infancy and early childhood show relatively low heritability
Page 4 of 25
example, 19-25 month-olds’ observed prosocial helping toward their mothers in a
effect[7].
Starting at age three, research shows more clearly the importance of genetics to
t
ip
prosocial behavior. This was replicated with questionnaire (parent, teacher, and self-
report) data from samples of diverse cultural backgrounds (Israel[8], Nigeria[9], South
cr
Korea[10], the USA[11], and the United Kingdom[12]). Research using observational
us
measures is rarer. One study[13] found only environmental contributions to
for compliant (following request) (34%) and self-initiated (without request) prosocial
M
behaviors (43%), with nonshared environment accounting for the remaining variance[14],
Longitudinal designs enable study of the changing roles of genetics and the
e
environment within the same individuals. A study of observed prosocial acts[15] found
pt
that the magnitude of genetic and environmental effects varied across ages among
ce
American 14 to 36 month-olds. While most of the variance was due to the nonshared
environment, genetic effects dropped first and then tended to rise and shared environment
Ac
effects tended to fall during early development. This general pattern of increasing
heritability and dropping shared environment has also been observed throughout
found that heritability increased from 26-37% at age two years to 60-62% at age seven
years. At the same time, shared environment effects became less important with age. A
Page 5 of 25
similar pattern was found[16] (cross-sectionally) from later childhood to early-mid
this developmental process. For potential explanations for the pattern of increasing
t
ip
heritability and decreasing shared environment effects, which has also been reported for
empathy[15] , general cognitive abilities[17] and other traits, see previous review[18].
cr
Complexity of Prosocial Behavior
us
Up until now, we have largely treated prosocial behavior as a unitary concept. Most
of the research cited above has used broad questionnaires, assessing a global prosocial
an
behavior dimension. However, prosocial behavior should be seen as a multifaceted,
complex trait[4, 19]. First, at the behavioral level sharing, helping and comforting do not
M
always correlate, suggesting that the behaviors reflect different psychological factors[20].
Second, different aspects of prosocial behavior have different social correlates, such as
d
social concern, kindness, helping, and empathic concern) in seven-year old Israeli twin
children all inter-correlated positively, consistent with the notion of an overall underlying
prosociality trait. Importantly, each aspect of prosociality also had substantial variance
not accounted for by the global prosociality factor. The common prosociality factor was
highly heritable (69%), while there were also genetic effects specific to each of the facets.
Page 6 of 25
This last finding suggested that one reason for which different prosocial behaviors only
partially overlap is that there are unique genetic factors that separately contribute to
t
ip
Methodological details appear elsewhere[24, 25]. Six experimental situations tested 3.5 year-
cr
old Israeli twins’ helping, sharing, and comforting[26].
us
(Figure 1). In the case of helping, similar MZ and DZ correlations (note to Figure 1) indicated
an
shared environment effects (when analyzed separately from self-initiated helping, compliant
helping did show higher MZ correlations). The positive correlation was larger among MZ
M
twins in both sharing and comforting, suggesting a genetic influence. In the case of
comforting, MZ twins correlated positively, while DZ twins did not correlate at all, which
d
Comforting correlated with both sharing, r = .20, p < .05, and helping, r = .23,
pt
p<.01, while sharing and helping did not correlate with each other, r = .04, ns. A bivariate
ce
results show the complex nature of prosociality and its genetic etiology.
Page 7 of 25
While twin studies demonstrate a phenotype’s heritability, they are uninformative
regarding the specific biological processes involved. For this, we turn to molecular genetic
studies. Research on the associations between specific genetic polymorphisms and prosocial
behavior has largely taken a candidate gene approach, selecting polymorphisms for analysis
t
ip
based on putative links between the behavior and the biological mechanisms which give rise
to it[18]. Most candidate gene studies of social phenotypes have focused on a few
cr
neurotransmitter systems in adults[18, 27]: the dopaminergic system, involved with executive
us
function, learning and reward[28]; the oxytocinergic and vasopressinergic systems, involved
in social bonding and affiliative behaviors[29]; and the serotonergic system, involved in
nucleotide polymorphism (SNP, variation in a single “letter” of the genetic code) of the
e
oxytocin receptor gene (OXTR) was associated with Chinese children’s higher helping and
pt
comforting, but not sharing, behaviors towards experimenters[32]. Importantly, this SNP was
OXTR SNP, rs2254298, was associated with self-reported prosocial tendencies, an effect
Ac
mediated via moral evaluation of behaviors[34]. Variation in the promoter region of the
AVPR1a gene which codes for the arginine vasopressin 1a receptor has been associated with
observed sharing behavior in Israeli 3.5 year old twin children[35]. Using parent reports of
older children’s prosocial behavior, an American study marginally replicated this finding[36].
Page 8 of 25
Although intriguing, candidate gene findings have been criticized for several
reasons[18], including relatively small samples, not controlling for differences in allele
frequencies in subsamples of different ancestry, and lack of replication. Although it has not
been tested directly, the genetic architecture of prosocial behavior, like other complex traits,
t
ip
genes[37] is likely not based on a few genes with large effects, but rather on additive and
interactive small effects across many genes. A more recent approach includes the hypothesis-
cr
free study of genetic variation in hundreds of thousands of loci across the genome. Naturally,
us
such genome-wide research necessitates very large samples, typically tens of thousands of
Gene-Environment Interactions
pt
Although until now we have discussed the role of genetics and the environment
separately, the two do not operate independently. Gene x Environment interactions (GxE)
ce
are indicated when individuals with certain genotypes are affected by the environment
differently than other individuals, and vice versa[40, 41]. Several molecular genetic
Ac
studies have demonstrated evidence for GxE, but only a few studied children. For
example, variation in a functional variable number tandem repeat in the third exon of the
across studies and different environmental factors; in one study the DRD4 polymorphism
Page 9 of 25
was associated with sharing with peers, but did not interact with parenting in predicting
sharing[43]). In another study, DRD4 did not interact with being preferred by peers to
t
ip
genetic effects. Heritability of parent-rated prosocial behavior in 3-year old Israeli twins
was estimated as much lower when twins had no additional siblings than when they
cr
did[8]. In a Nigerian study the relative contribution of environmental factors to
us
adolescents’ prosocial behavior was higher in families low in cohesiveness than in
leadership when children had many friends, suggesting that the quality of the social
M
environment moderates genes’ influence[45]. Importantly, the differential heritability
observed in different ages[6] could also indicate that the role of genes becomes different
d
as children’s social environment changes, for example as they move to larger social
e
Gene-Environment Correlations
ce
Gene-environment correlations (rGE) are another major way in which genetic and
environmental influences are interrelated[46, 47]. Passive rGE occurs when one family
Ac
factors that overlap with genetic effects on the child’s behavior[48]. Such association
between children’s genes and the environment they receive indicates that some
correlations observed between parenting and children’s behavior may reflect the shared
genetics of family members[49]. Although studies have yet to test this in the context of
Page 10 of 25
prosocial behavior, the AVPR1A gene, which has been associated with preschoolers'
sharing[35] has also been associated with mothers’ parenting[50], suggesting it may be a
Children also affect their environment in a way corresponding with their own
t
ip
genotype, a process termed evocative rGE. One way to address this is by seeing whether
siblings with different genotypes (e.g., DZ twins) receive different parenting that is
cr
associated with differences in their own, genetically-influenced behavior. For example,
us
using a twin design, the associations between children’s prosocial behavior and the
parenting they received were largely accounted for by genetic overlap between parenting
an
and genetic differences between twins in prosocial behavior[51] . Moreover, children’s
Children are also active agents in their prosocial development. Their genetically-
d
influenced prosocial behavior can lead them to select environments (e.g., peer networks)
e
that will suit their own behavior; a process called active rGE. One study demonstrating
pt
active rGE reported that a common genetic factor related both to children’s tendency to
ce
reciprocated friendships, evocative rGE might be at play too.) Both active and evocative
Ac
rGE affect the input children receive from their environment. To the extent that such
experiences (e.g., parenting or peer reaction) also affect the developing child,
accumulate and amplify the initial differences in prosociality. This process of genetically
Page 11 of 25
mediated selection into specific environments could also be one explanation for the
Conclusions
t
ip
the role of genetics in individual differences in prosocial behavior, although genetic
effects are dependent on age, content of behavior (sharing, helping or comforting), and
cr
additional, environmental variables. However, research addressing the complex network
us
of rGE and GxE[55] is still needed. Epigenetics, modifications in the DNA that do not
affect the DNA sequence but alter gene expression[56, 57, 58] are another potentially
an
important way for future research on how environments and genes work together in
prosocial development.
M
Prosocial behavior is a complex trait, likely to be influenced by many genes with
small effects[37]. Only recently has research begun to unravel the ways by which the
d
interplay of genes and environments account for variation in prosocial behavior, delineate
e
different aspects of prosocial behavior, and link prosocial behavior with other
pt
psychological processes, such as theory of mind, empathy, personality traits and values.
ce
The way all of these variables relate and interact at the psychological and brain level is a
fascinating quest for our understanding of the noble side of human nature.
Ac
Page 12 of 25
Acknowledgments
This research was supported by grants from the Israel Science Foundation (31/06,
1670/13). The authors thank all the families who participated in the Longitudinal Israeli
Study of Twins.
t
ip
References
[1] A. Dahl, “The developing social context of infant helping in two US samples,”
cr
Child Dev., vol. 86, no. 4, pp. 1080–1093, 2015.
us
* This study shows that very young infants (from around the first birthday) help at home,
linking this help to encouragement, thanking, or praising by family members.
NJ, US: John Wiley & Sons Inc, 2015, pp. 610–656.
pt
** A comprehensive review that describes prosocial development from infancy and on,
ce
[4] G. Padilla-Walker, L. M., & Carlo, “Prosocial behavior: Past, present, and future,”
[5] P. Plomin, R., DeFries, J. C., McClearn, G. E., & McGuffin, Behavior Genetics.
[6] A. Knafo and R. Plomin, “Prosocial behavior from early to middle childhood:
Page 13 of 25
vol. 42, no. 5, pp. 771–786, 2006.
Familial Link Between Positive Affect and Empathy Development in the Second
t
ip
[8] A. Knafo and S. Israel, “Genetic and Environmental Influences on Prosocial
Behavior Ariel,” in Prosocial motives, emotions, and behavior: The better angels
cr
of our nature, 2009, pp. 149–167.
us
[9] Y.-M. Hur, J. Taylor, H.-U. Jeong, M.-S. Park, and B. C. Haberstick, “Perceived
2017.
M
*The only behavioral genetic study of prosocial behavior with an African sample, also
prosocial behaviour in 2-to 9-year-old South Korean twins,” Biol. Lett, vol. 3, no.
pt
* A twin study demonstrating that although prosocial behavior in young USA children
Page 14 of 25
“Behavioral genetic analyses of prosocial behavior in adolescents,” Dev. Sci., vol.
t
ip
cr
us
“In defence of situational morality: genetic, dispositional and situational
determinants of children’s donating to charity,” J. Moral Educ., vol. 39, no. 1, pp.
[14]
1–20, 2010. an
A. Knafo, S. Israel, and R. P. Ebstein, “Heritability of children’s prosocial
M
behavior and differential susceptibility to parenting by variation in the dopamine
53–67, 2011.
ce
Page 15 of 25
Kovas, R. P. Corley, J. C. DeFries, J. K. Hewitt, R. K. Olson, S.-A. Rhea, S. J.
linearly from childhood to young adulthood,” Mol. Psychiatry, vol. 15, no. 11, pp.
t
ip
1112–1120, 2010.
cr
prosociality,” Curr. Opin. Psychol., vol. 6, pp. 55–59, 2015.
us
* a review of quantitative and molecular genetic research on diverse aspects of morality
adults. an
[19] A. Knafo and S. Israel, “Empathy, prosocial behavior, and other aspects of
M
kindness,” in Handbook of temperament, 2012, pp. 168–182.
comforting subtypes,” Front. Psychol., vol. 5, no. AUG, pp. 1–13, 2014.
e
prosocial behavior in children: Links with attachment theory and research,” Soc.
ce
* In this study of USA children aged 18–30 months diverse prosocial behaviors
Page 16 of 25
(empathic and instrumental helping, sharing), related differently to third variabels
behavior did not relate to temperament, the authors emphasize the role of
socialization.
t
ip
[23] A. Knafo-Noam, F. Uzefovsky, S. Israel, M. Davidov, and C. Zahn-Waxler, “The
cr
mother-reported behavior of 7-year-old twins,” Front. Psychol., vol. 6, no. FEB,
us
pp. 1–9, 2015.
[24] R. Avinun and A. Knafo, “The Longitudinal Israeli Study of Twins (LIST)-an
an
integrative view of social development.,” Twin Res. Hum. Genet., vol. 16, no. 1,
*this twin study of 7-year old Israeli children asks whether prosocial behavior can be
pt
behavior, such as helping and sharing, all load on a single, substantially heritable
factor, but each has additional independent variance with separate etiology,
Ac
taking ability is associated with low self-initiated pro-sociality.,” Emotion, vol. 11,
Page 17 of 25
Development,” in Oxford Handbook of Parenting and Moral Development, .
social brain,” Trends Cogn. Sci., vol. 13, no. 1, pp. 27–35, 2009.
t
ip
Genes: Evolutionary Ancient System Impacting on Human Affiliation, Sociality,
and Psychopathology,” Biol. Psychiatry, vol. 79, no. 3, pp. 174–184, 2016.
cr
[30] A. A. Marsh, S. L. Crowe, H. H. Yu, E. K. Gorodetsky, D. Goldman, and R. J. R.
us
Blair, “Serotonin transporter genotype (5-HTTLPR) predicts utilitarian moral
[31] an
M. J. Crockett, L. Clark, M. D. Hauser, and T. W. Robbins, “Serotonin selectively
influences moral judgment and behavior through effects on harm aversion,” Proc.
M
Natl. Acad. Sci., vol. 107, no. 40, pp. 17433–17438, 2010.
[32] Y. Wu, N., & Su, “Oxytocin receptor gene relates to theory of mind and prosocial
d
behavior in children,” J. Cogn. Dev., vol. 16, no. 2, pp. 302–313, 2015.
e
[34] S. Shang, N. Wu, and Y. Su, “How oxytocin receptor (OXTR) single nucleotide
Ac
**this study of Chinese adolescents tries to unravel the role in prosocial behavior of the
oxytocin receptor gene, a gene that has been implicated in many social variables.
The authors test a model in which moral evaluation of positive and negative
Page 18 of 25
behaviors mediates the effect of the gene on prosocial behavior.
t
ip
[36] S. E. Meek, “Gene-Environment Interplay and Prosocial Behavior An Analysis of
cr
[37] D. J. Benjamin, D. Cesarini, M. J. H. M. van der Loos, C. T. Dawes, P. D.
us
Koellinger, P. K. E. Magnusson, C. F. Chabris, D. Conley, D. Laibson, M.
2012.
M
[38] V. Warrier, R. Toro, B. Chakrabarti, N. Litterman, D. Hinds, T. Bourgeron, and S.
and correlates with sex, education, and psychiatric risk,” bioRxiv, vol. 44, no. 0, p.
e
50682, 2016.
pt
[39] D. W. Belsky and I. S, “Integrating Genetics and Social Science: Genetic Risk
ce
Scores,” Biodemography Soc. Biol., vol. 60, no. 2, pp. 137–155, 2014.
[40] A. Avinun, Reut, and Knafo-Noam, “Socialization, genetics, and their interplay in
Ac
neurodevelopmental theory.,” Dev. Psychopathol., vol. 23, no. 1, pp. 7–28, 2011.
Page 19 of 25
susceptibility to rearing environment depending on dopamine-related genes: New
evidence and a meta-analysis,” Dev. Psychopathol., vol. 23, no. 1, pp. 39–52,
2011.
t
ip
interaction effects on preschoolers’ social behaviors,” Dev. Psychobiol., vol. 51,
cr
[44] J. M. Buil, H. M. Koot, T. Olthof, K. A. Nelson, and P. A. C. van Lier, “DRD4
us
Genotype and the Developmental Link of Peer Social Preference with Conduct
Problems and Prosocial Behavior Across Ages 9–12 Years,” J. Youth Adolesc.,
and correlation in the analysis of human behavior.,” Psychol. Bull., vol. 84, no. 2,
ce
[47] S. Scarr and K. McCartney, “How people make their own environments: a theory
Ac
of genotype greater than environment effects.,” Child Dev., vol. 54, no. 2, pp. 424–
435, 1983.
[48] A. Rice, F., Lewis, G., Harold, G. T., & Thapar, “Examining the role of passive
sensitive design,” Dev. Psychopathol., vol. 25, no. 1, pp. 37–50, 2013.
Page 20 of 25
[49] A. Knafo and S. R. Jaffee, “Gene–environment correlation in developmental
with arginine vasopressin receptor 1A gene.,” Biol. Lett., vol. 8, no. 5, pp. 894–6,
t
ip
2012.
[51] A. Knafo and R. Plomin, “Parental discipline and affection and children’s
cr
prosocial behavior: Genetic and environmental links.,” J. Pers. Soc. Psychol., vol.
us
90, no. 1, pp. 147–164, 2006.
longitudinal twin models,” Dev. Psychopathol., vol. 25, no. 1, pp. 7–16, 2013.
d
Special Section,” Child Dev., vol. 87, no. 6, pp. 1655–1667, 2016.
Page 21 of 25
a meta-analysis of children-as-twins studies.,” Personal. Soc. Psychol. Rev., vol.
t
ip
Natl. Acad. Sci., vol. 113, no. 27, pp. E3816–E3823, 2016.
cr
of the oxytocin receptor gene, as well as brain and social variables. Methylation
us
related to brain function and structure as well as to secure attachment styles and
Page 22 of 25
Ac
ce
pt
ed
M
an
us
cr
ip
t
Page 23 of 25
cr
us
Figure 1.
an
Genetic and Environmental Variance Component Estimates in Different Prosocial Behaviors
M
ed
pt
ce
Ac
Note: Sharing, helping and comforting were each tested twice, once in a self-initiated setting, and once following a request, and
standardized scores were then averaged separately for each of pair of tasks. When possible, we used finer coding (e.g., not only
presence of sharing, but also amount shared) than the original study[26]. Twin correlations for helping, sharing, and comforting, were
.23, .24, and .57, respectively for MZ twins (N=61 pairs), and .22, .15, and .04 for DZ twins (N=91 pairs). The genetic effect on
helping and the shared environment effects on sharing and comforting could be dropped off the model without affecting model fit. All
effects portrayed in the figure had 95% confidence intervals above zero. Additional information on the analyses can be obtained from
the authors.
Page 24 of 25
Conflict of interest: None
t
ip
cr
us
an
M
d
p te
ce
Ac
Page 25 of 25