Handbook of Psychology, Personality and Social Psychology

Irving B. Weiner, Howard A. Tennen, Jerry M. Suls

• Publisher: Wiley
• Year: 2012
• ISBN:
  9781118281925
• Edition: 2
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

Psychology is of interest to academics from many fields, as well as to the thousands of academic and clinical psychologists and general public who can't help but be interested in learning more about why humans think and behave as they do. This award-winning twelve-volume reference covers every aspect of the ever-fascinating discipline of psychology and represents the most current knowledge in the field. This ten-year revision now covers discoveries based in neuroscience, clinical psychology's new interest in evidence-based practice and mindfulness, and new findings in social, developmental, and forensic psychology.

Part I

Personality

Chapter 1

Genetics of Personality

Susan C. South, Ted Reichborn-Kjennerud, Nicholas R. Eaton, and Robert F. Krueger

Introduction

Behavior Genetics of Personality

Molecular Genetics of Personality

Summary and Future Directions

References

Introduction

In this current chapter, we review findings from an important approach to understanding the etiology of personality—behavior genetics. Behavior genetics encompasses a series of methods for disentangling the relative influence of genes and environment on the variation in a phenotype (observed variable). Personality, or the characteristic ways that people think, feel, and behave, varies considerably within and between different populations. There are people who are more or less aggressive, more or less modest, more or less sociable, more or less humorous, and so on. It is the variation in personality that defines us and makes us distinct from every other individual on the planet. There are certainly those who have questioned the existence of consistent personality dispositions (Mischel, 1968), or theorized as to why any variation in personality exists (Penke, Denissen, & Miller, 2007). Behavior genetics is vitally important in showing that personality does exist, that it does vary in meaningful ways between people, and that this variation is due, in part, to genetic influences.

Indeed, the field of behavior genetics, particularly the use of genetically informative family data, was a driving force in establishing the importance of both nature and nurture in the development of personality. Quantitative behavior genetic methods, which parse out the relative influence of genes, family environment, and unique environmental experiences, have consistently demonstrated the importance of genetic influences. Biometric modeling with twin data reveal one of the most reliable findings in behavior genetics and in the personality literature—heritability, or proportion of variance due to genetic influences, for most major personality traits is approximately 40 to 50%. A robustly replicated finding, this has caused consternation in the molecular genetic field of personality research, where scientists have been largely stymied in their attempts to find measured genes that explain a significant portion of the variance in personality traits. Why, they have asked, is it so difficult to replicate measured gene-measured personality findings, if we know that genetic influences explain 50% of personality variation? In the current chapter, we review what has been found, and how improving technology may partly solve the problem. We also note that part of the difficulty for molecular genetics is the complex interactions that might be occurring, not only between different genes but also between genes and the environment. Gene-environment interaction, or the impact the environment has on the expression of genetic influences, may be important in our search for the etiology of personality. It might be a return to quantitative genetic methods, which can provide estimates of gene-environment interplay, which best informs our understanding of personality development. Thus, the field of behavior genetics comes full circle, from an early reliance on quantitative methods that set the stage for molecular work, which again must now rely on statistical modeling of family data to move forward. In the current chapter, our goal is to present a general conceptual overview of behavior genetics methodology, including both biometric models and molecular genetics techniques, as well as a snapshot of seminal work from the past, current issues and controversies, and recommendations for future research.

Behavior Genetics of Personality

We begin by briefly reviewing the major biometric modeling approaches and the assumptions underlying them, before turning to a broad overview of findings from univariate and multivariate twin studies and adoption studies.

Biometric Modeling Approaches

Much of the classic work in the behavior genetics of personality has been conducted with biometric modeling of twin data. These statistical models are biological in that they utilize known genetic relationships among individuals, and metric because they attempt to provide estimates of genetic and environmental influences based on careful observed measurement of a phenotype. Biometric modeling of personality data is based on certain assumptions about the etiology of personality. As with other types of individual difference phenotypes (e.g., psychopathology), the etiology of personality is probably best explained by a multifactoral polygenic model of inheritance. According to this model, there is a continuum of liability throughout the general population resulting from genetic and environmental influences that act in an additive manner (Falconer, 1965; Gottesman & Shields, 1967). Given that behavior genetic models generally assume there are multiple gene systems, or quantitative trait loci (QTLs), underlying complex behavioral phenotypes like personality, quantitative genetic research (i.e., biometric modeling) is therefore appropriate for examining the etiology of personality. The method most commonly used for examining genetic and environmental influences on the variation in personality is the twin study, therefore much of our review will spotlight this method and the findings from twin studies; adoption methods have also been utilized and we discuss findings from this work later in the chapter.

Twin pairs are a fascinating form of natural experiment, allowing researchers to disentangle the relative influence of genetics and environment on a phenotype, like personality. Identical (monozygotic, MZ) twins are the result of one fertilized egg splitting in two while in utero. Fraternal (dizygotic, DZ) twins are the result of two separate eggs being fertilized at the same time, and are no more alike (genetically) than two nontwin siblings. MZ twins thus share 100% of their genes, while DZ twins share an average of 50% of their segregating genes. When using a sample of MZ and DZ twin pairs in which both members of the pair are raised in the same home, the degree of genetic and shared family environment is known. Biometric modeling can then use the concordance (agreement) between twins on a phenotype of interest to decompose the variance in that phenotype into genetic and environmental components. As a first step, one can estimate correlations between twin pairs and compare differences in the magnitude of correlations between MZ and DZ twins to obtain a general indication of the size of genetic and environmental influences. For instance, when the MZ correlation is greater than the DZ correlation, this indicates the presence of genetic influences on the trait. Formal biometric modeling of twin data is done using structural modeling software (e.g., Mx; Neale, Boker, Xie, & Maes, 2003), by comparing the similarity (i.e., the covariance) within MZ and DZ twin pairs on the phenotype, resulting in estimates for genetic and environmental influences. In this section, we focus on univariate twin studies, which decompose the variance of one phenotype (i.e., one personality trait); however, structural modeling is readily extended to the multivariate case (e.g., the structure of multiple personality traits considered together, discussed below).

A univariate (“one variable”) biometric model with twin pairs is used to separate the variation in a phenotype into three sources that collectively account for the total variance in the population (Plomin, DeFries, McClearn, & McGuffin, 2008). This is an important point, and often lost when discussing biometric models; these models explain variation in the population from which the sample is drawn, not the relative influence of genes and environment on any one person's outcome. The first source of influence on variation is heritability (abbreviated h²), which reflects how much of the variation in a personality phenotype is due to genetic differences between people in a population. The heritability estimate is actually a ratio, or a proportion of genetic variation over total variation (the sum of genetic and environmental variation). Importantly, the heritability statistic and the commensurate estimates for environmental influences are population parameters; when we conclude that the heritability of a personality trait is 50%, what we are actually saying is that genetic differences among people in the sample drawn from that population account for 50% of the variance in that trait; again, we are not saying that genes account for 50% of any one individual's personality. In other words, among any population of individuals, some will be more extroverted and some will be more introverted and most will cluster around the mean, and 50% of the reason for this variation is that genetic influences also vary among people. Most biometric models of personality assume that genetic influences are additive, meaning that personality variation is due to the influence of many genes of small effect size located at different places (loci) on the genome. There is evidence for nonadditive genetic effects on normal personality traits (e.g., Keller, Coventry, Heath, & Martin, 2005) as well.

Beyond genetic influences, a second source of variation in a phenotype is the effect of the shared or common environment, abbreviated c². This component of variance captures the extent to which twins are similar by virtue of growing up in the same household. Examples of the shared environment include neighborhood influences, socioeconomic status, having similar friends or peer groups, customs, habits, and the extent to which siblings have similar interactions with their parents. A final source of phenotypic variance is the unique or nonshared environment, abbreviated e². This component of variance indexes the extent to which twins are different from each other despite having grown up in the same household and sharing genes. Examples of nonshared environmental experience include traumatic events and stressors, having different friends and life experiences from one's sibling, events in utero, and the extent to which each sibling has a unique experience with their parents. It is important to note that measurement errors are included in the estimate of nonshared environmental influences, so any imprecision or bias in measurement among individuals will result in inflated estimates of e². The distinction between the two environmental sources of variance can often be subtle. As noted above, neighborhoods are often thought of as shared environmental influences, working to make siblings within the family more similar to each other; however, one sibling's experience or perception of the environment may be quite unique to him or her and thus work to make siblings growing up within the same family less similar to each other (and this would be accounted for under the nonshared environmental component of variance).

Findings From Univariate Twin Studies of Personality

For decades, researchers, the media, and the lay public debated the relative influence of nature versus nurture on the development of individual differences in human behavior. Thanks to decades of work from the field of behavior genetics, we can now definitely say that both nature and nurture are at work when we consider the etiology of almost any phenotype that differs between people and would be of interest to psychologists. Indeed, virtually every phenotype that can be said to differ in meaningful ways between individuals in the population has a detectable genetic component, a finding so well-replicated that it has been called the “First Law” of behavior genetics (Turkheimer, 2000). This is certainly true for almost every major personality trait or domain, as research across different populations, cultures, and personality measures has found heritability estimates of approximately 50%, with the rest of the variance primarily attributed to nonshared environmental influences. Given the prominence of trait models of personality in general and the Five-Factor Model/Big Five Model in particular (McCrae, Gaines, & Wellington, this volume), it is not surprising that behavior genetic modeling has focused on these personality domains (extraversion, openness, agreeableness, conscientiousness, and neuroticism). The heritability of all five domains range from 40 to 50%, with a majority of the rest of the variance accounted for by nonshared environmental influences (Bouchard & Loehlin, 2001). It is compelling that parameter estimates of genetic and environmental influences are so consistent across these major personality domains, which obviously differ widely in the aspects of human behavior that they capture.

There are several aspects of these key findings that bear further discussion. First, all biometric models with twin data are built on the equal environments assumption (EEA). This assumption derives from the following logic: biometric modeling compares the similarities between MZ twin pairs to the similarities between DZ twin pairs to arrive at estimates of genetic and environmental components of variance. We infer that greater similarities between MZ twin pairs is due to greater genetic similarity, since in both MZ and DZ pairs, twins are raised in the same environment (if they are not reared apart). But what if MZ twin pairs are more similar because their parents impose more similar treatment on them, compared with parents of DZ twin pairs, in ways that do not reflect the contributions of the MZ twins' genes to parental treatment? If this were true, it could result in biased estimates of genetic effects. However, there is now substantial evidence supporting the EEA (Goodman & Stevenson, 1991; Loehlin & Nichols, 1976; Scarr & Carter-Saltzman, 1979); even when the environment does seem biased to making MZ twins more similar (i.e., they are dressed the same by their parents), this does not appear to have a major influence on phenotypic similarity for individual difference phenotypes like personality.

Second, there is a great deal of consistency in the heritability estimates for a w...