What does it mean to say that heritability of height is 90 percent what does that tell us about the contribution of genetics to any one persons height?

3.1 Heritability Levels

The level of heritability has been considered in a large number of twin, adoption, and family studies. The level of heritability is simply the fraction of the population variation in a trait (e.g., BMI) that can be explained by genetic transmission. Such studies use a variety of procedures to control or take into account age and sex differences. Results obtained by a good number of investigators indicate that the heritability level estimates depend on how the study was conducted and on the kinds of relatives upon which they are based (Table 1). For instance, studies conducted with identical twins and fraternal twins or identical twins reared apart have yielded the highest heritability levels with values clustering around 70 percent of the variation in BMI.

Table 1. Overview of the genetic epidemiology of human body fat/obesity (The table is based on the trends in about 50 different studies. In most of these studies, the BMI was the phenotype considered. In some cases, skinfolds or estimates of percentage body fat or fat mass were used. From Bouchard 1994)

In contrast, the adoption studies have generated the lowest heritability estimates of about 30 percent or less. The family studies have generally found levels of heritability intermediate between the twin and the adoption study reports. A few investigations have included all or most of these kinds of relatives in the same analysis. Using analytical techniques developed to use all the information and maximum likelihood procedures, these studies have concluded that the true heritability estimate for BMI in large sample sizes was between 25 and 40 percent. Recent surveys undertaken in Sweden and the USA with the collaboration of severely obese and morbidly obese subjects together with information obtained on their parents, siblings, and spouses confirm that the genetic contribution to obesity may indeed be around 25 to 40 percent of the individual differences in BMI (Bouchard et al. 1998).

3.2 Familial Risk for Obesity

The risk of becoming obese when a first-degree relative is overweight or obese can be quantified using the lambda coefficient (λs) defined as the ratio of the risk of being obese when a biological relative is obese compared with the risk in the population at large, i.e., the prevalence of obesity (Risch 1990). Age- and gender-standardized risk ratios (equivalent to λs) obtained from 2349 first degree relatives of 840 obese probands and 5851 participants of the National Health and Nutrition Examination Survey III (NHANES III) revealed that the prevalence of obesity (BMI⩾30) is twice as high in families of obese individuals than in the population at large (Lee et al. 1997). Moreover, the risk increases with the BMI threshold used to define the obesity. Thus, the risk of extreme obesity (BMI⩾45) is about eight times higher in families of extremely obese subjects.

The familial clustering of morbid obesity was also investigated (Adams et al. 1993) in 221 families (n=1560, subjects ages 18 years and older) in Utah ascertained through a single morbidly obese proband. All probands had to be at least 45.5 kg over their ideal body weight. About 48 percent of the morbidly obese probands had one or more first-degree relatives who were also morbidly obese. Using data from 20,455 families from the general Utah population, the authors estimated that the prevalence of morbid obesity reached 6 percent in the Utah population. Thus, the risk of morbid obesity was about eight times higher (6 percent vs. 48 percent) in families segregating for morbid obesity than in the general population. More recently, using data from 15,245 participants aged 7 to 69 years from the 1981 Canada Fitness Survey, it was shown that the familial risk of obesity was 5 times higher for relatives in the upper 1 percent of the distribution of BMI than in the general Canadian population (Katzmarzyk et al. 1999).

1.1 Univariate Results

Most genetic studies of personality are based on very large twin and family samples. Personality has been assessed on over 50,000 pairs of twins in the national twin registries in Australia, Finland, Norway, Sweden, and The Netherlands, on volunteer twin registries in the USA, and on the 30,000 individuals in the Virginia study of twins and their families. Numerous smaller twin, family, and adoption studies have also been performed. Many of the studies (particularly those in the national twin registries) have assessed personality, using versions of the Eysenck Personality Inventory. However, during the 1980s and 1990s increasing amounts of data have become available concerning other personality inventories, such as the Five-Factor Model of Costa and McCrae (McCrae and Costa 1987) and the Seven-Factor Model of Cloninger (Cloninger 1998).

Remarkably, there is considerable consensus across studies and cultures. Genetic effects account for on average 35–40 percent of the variation in personality traits such as neuroticism and extraversion. (In other words, the heritability of these traits is 0.35–0.40.) Evidence from studies of twins reared apart and adoption studies indicate that shared environmental influences account for another 5–10 percent of the variation in most personality traits, and the remaining variation is individual specific or ‘non-shared’ environmental variation. In other words, people differ for personality. Most of the individual differences we observe are due to individual specific experiences not shared by other family members and about 40 percent of the differences are due to genetic differences among us. The extent to which family members are similar to one another reflects primarily their genetic similarity, and only in part being raised in the same home.

There is some heterogeneity in genetic effects on personality. Traits related to extraversion, such as impulsivity, openness to experience, and sensation seeking tend to have slightly higher heritability estimates than those related to neuroticism and instability. Furthermore, the genetic influences for the extraversion-related traits may in part be operating in a nonadditive manner. That is, there may be interactions within or between genetic loci, such that the effect of one gene may depend on another gene. The latter effect, known as epistasis (see Genetics of Complex Traits Through the Life Cycle), may complicate efforts to identify specific genes that are of importance for personality traits.

Heritability estimates may be on average 40 percent for most personality traits, but there are traits for which genetic effects are of less importance. Examples of these are ‘agreeableness,’ ‘guilt,’ feelings of control over ‘luck,’ and ‘cynic hostility.’ For these measures, familial similarity is primarily due to environmental influences shared by family members rather than genetic similarity.

1.2 Genetic Influences on Stability and Change

There has been a great deal of controversy as to whether personality is stable throughout the lifespan. Some developers of personality inventories argued that finding significant heritability for personality would support hypotheses of stability in personality. Their argument rested on the misconception that if genetic effects are important they are immutable and do not change. Despite the fact that humans are born with a full complement of genes, not all genes may be operating at all phases in life. Some genes turn on and others turn off at various transitional periods, while others may react dependent on external stressors. On the other hand, some genes may have a long-term effect that is stable across the lifespan (see Genetics of Complex Traits Through the Life Cycle). Thus, the issue of genetic contributions to stability for personality is empirically testable.

During infancy, there is less stability in temperament than is typically reported for adulthood. Nevertheless, results from longitudinal twin studies indicate that the moderate stability observed for infant temperament is influenced in part by genetic effects. During childhood, genetic effects are also important contributors to phenotypic (i.e., observable) stability. Longitudinal studies of adult twins confirm findings of substantial phenotypic stability in personality. Furthermore, the genetic effects themselves are very stable, i.e., the same genes appear to be influencing personality throughout adulthood. Nonetheless, genetic and environmental influences are equally important for phenotypic stability. Even though the genetic effects themselves are very stable, they are relatively less important than environmental effects, and thus the net influence on stability is reduced. Environmental effects are the most important influences on change. Thus, genetic influences on personality contribute to stability rather than to change. However not all personality stability can be explained by genetic influences alone.

1.3 Structure of Personality

Developments in theories of personality have been paralleled with developments in statistical analyses of genetically informative data. During the 1980s and 1990s, the three-factor approach to personality was gradually extended to the five-factor theory and the seven-factor theory (McCrae and Costa 1987, Cloninger 1998). The factor analytic approach generally applied at the phenotypic level is also applicable at the level of genetic and environmental influences. The question to be addressed is whether there is a genetic factor structure or an environmental factor structure that is similar to the phenotypic factor structure. Do genetic or environmental effects mediate the associations among various aspects of personality? This line of analysis has only recently been applied and promises to be an important area for future research concerning the structure of personality.

1.4 Genetic Mediation of Associations with Personality

Health psychologists have demonstrated that personality is important for health outcomes (such as cardiovascular disease, cancer, and osteoarthrosis) and response to treatment such as recovery from surgery. Current research efforts are focusing on the extent to which genetic and environmental effects mediate the associations between personality and health outcomes. By evaluating these associations, much can be learned about the mechanisms by which personality affects health. For example, neuroticism in young adulthood is predictive of joint pain and osteoarthritic symptoms some 25 years later in midlife. Genetic influences in common to neuroticism and joint pain and shared familial environmental influences mediate this association. There are no genetic influences unique to joint pain. Because neuroticism preceded the joint pain, it is likely that genetic effects are important for neuroticism which in turn influences joint pain and osteoarthritic symptoms through response to stress, immune, and inflammatory processes (Turk-Charles et al. 1999).

2.1 Genetic Transmission

Anxious children are more likely to have anxious parents and anxious parents are more likely to have anxious children. These familial relationships could indicate either genetic or family environment influences. Evidence confirms that genetic factors do play a part in determining the development of childhood anxiety disorders, but clearly this explanation does not account for many cases of childhood anxiety. The research has found heritability estimates of around 40–50 percent (Thapar and McGuffin 1995), meaning that other factors play an important role, in addition to genetic determination. Although genetic factors are clearly involved in the development of anxiety for some children, it remains to be shown exactly what is inherited. What appears to be inherited is an increased propensity to develop anxiety related problems, rather than a specific anxiety disorder. This propensity may relate to some temperament pattern in the child that increases the risk of developing anxiety disorders.

2.2 Child Temperament

Temperament theorists believe that early child temperament is of etiological significance to the later development of childhood anxiety. ‘Behavioral inhibition’ is the term used to describe one particular pattern of childhood temperament that has been most frequently linked with childhood anxiety problems. It can be defined as a relatively stable temperament style characterized by initial timidity, shyness, and emotional restraint when exposed to unfamiliar people, places, or contexts. This temperament pattern is associated with elevated physiological indices of arousal and has been shown to have a strong genetic component. Most importantly, children exhibiting a temperament style of behavioral inhibition demonstrate an increased likelihood of developing child anxiety (see Kagan 1997 for a review of this area). Other temperament theorists argue for the existence of three stable factors: positive affectivity/surgency (PA/S), negative affectivity/neuroticism (NA/N), and effortful control (EC) (Lonigan and Phillips in press). According to this theory, high NA/N combined with low EC places children at risk for the development of anxiety problems, and there is some tentative evidence to support this proposition. However, as not all children exhibiting an early temperament style of behavioral inhibition, or high NA/N combined with low EC, go on to develop an anxiety disorder, the presence of moderating or mediating variables appears to be likely. In particular, attachment style and parenting characteristics (see Sects. 2.3 and 2.4) are likely to interact with early childhood temperament to determine the development of anxiety problems.

Although the literature regarding childhood temperament is interesting, it tells us little about the exact mechanism of action. It remains to be determined whether temperament impacts upon anxiety through greater susceptibility to conditioning processes, greater emotional and/or physiological arousability to stressful events, or through cognitive processes. For example, it is feasible that ‘at risk’ temperaments have their impact though greater tendencies to detect and attend to threatening stimuli in the environment, or expectations regarding the occurrence of negative outcomes. It has been shown in several studies that anxious children are more likely than others to think about negative events and to expect negative outcomes from situations.

2.3 Parenting Characteristics

The strong family links found in childhood anxiety could also be explained to some degree by parental behavior and the family environments in which the children are brought up. Parenting behavior has been suggested to impact upon child anxiety in a number of ways. From a learning theory perspective, certain forms of parenting behavior may increase the probability that children learn to respond in an anxious manner and fail to acquire the skills needed to cope with the inevitable stressful events that occur during children's lives. Observational studies have demonstrated that parents of anxious children are more likely to model, prompt, and reinforce anxious behavior, such as avoidance and distress in stressful situations. Furthermore, parents of anxious children are more likely to draw their children's attention to the threatening aspects of situations and less likely to encourage ‘brave’ solutions (Rapee in press).

The parents of anxious children are also more likely to engage in behaviors that make it less likely that children will learn how to solve stressful problems themselves. Empirical enquiry has found that parents of anxious children demonstrate higher levels of overcontrolling and overprotective behaviors that disrupt coping skills development. As a group, they are also more likely to be critical of their child's coping attempts, thereby reducing children's confidence in their abilities to solve their own life problems (Dumas, La Freniere and Serketich 1995, Krohne and Hock 1991). These parenting styles may interact with childhood temperament in explaining why some behaviorally inhibited children develop anxiety problems and some do not. For example, parental overprotection and overcontrol appears to be influential in determining the stability of behavioral inhibition in children (Hirshfeld et al. 1997a, 1997b). Parental behavior has also been found to be important in determining the impact of traumatic life events upon childhood psychopathology. Following trauma, children are more likely to develop emotional and behavioral difficulties if their parents react in an overprotective manner after the event (e.g., McFarlane 1987).

It is also important to consider that children have an influence upon parents, and anxious child behavior may cause parents to behave in particular ways. In much of the literature to date, it is not clear whether the overprotective behaviors of parents are definitely a cause of childhood anxiety or whether they could be a consequence of living with an anxious child. Future research needs to clarify these relationships.

2.4 Attachment Style

Recognizing the reciprocal effects of parents and their children, researchers have started to examine the quality of the attachment relationship between children and their caregivers. For example, Warren et al. (1997) found that anxious-resistant attachment at 12 months predicted anxiety disorders in adolescence, even after the effects of maternal anxiety and infant temperament were removed. There is also some evidence that attachment style may interact with infant temperament in the prediction of early markers of anxiety problems (e.g., Fox and Calkins 1993, Nachmias et al. 1996). It appears likely that certain patterns of behavior characteristic of particular forms of early childhood temperament make it difficult for parents and children to form secure attachments. Although this research is in its early stages, it appears to be an area that warrants further investigation. The quality of the parent–child attachment relationship may represent one mechanism through which familial transmission could occur. It is well recognized that parental psychopathology, particularly depression, disrupts parenting skills and interferes in attachment relationships. It may be that high levels of parental anxiety also disrupt effective parenting and attachment relationships, thereby contributing to intergenerational transmission of anxiety.

2.5 Traumatic, Negative and Stressful Life Events

The effect of traumatic, negative, and stressful life events on the development of anxiety in children is another area of etiological investigation. Perhaps not surprisingly, higher rates of anxiety disorders are associated with a range of natural disasters and traumatic life events (Benjamin et al. 1990). However, as not all children experiencing traumatic, negative, or stressful life events go on to develop anxiety disorders, the moderating or mediating influence of parenting behavior has been suggested. Indeed, what emerges from the literature relating to the etiology of childhood anxiety is a complex picture of interacting determinants and multiple pathways through which such problems may develop.

2.6 Protective Factors

Protective factors refer to variables that increase resilience to psychological disorder by reducing the impact of risk factors. Positive social support, particularly from a significant adult, is one such protective factor that has been suggested to provide a buffer against the development of anxiety problems, and indeed against the development of psychopathology in general. For example, a strong negative relationship has been found between child anxiety level and family social support (White et al. 1998). Child coping style is another protective factor suggested to play a role in child anxiety. Coping style is a generic term that relates to the way in which individuals attempt to cope with negative or aversive situations. There is some tentative evidence to suggest that children employing problem-focused strategies are less likely to experience psychopathology, whereas emotion-focused and avoidant coping styles are associated with higher levels of anxiety and depression (Compas et al. 1988).