AJCN 19th International Congress of Nutrition
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Speakman, J. R
Right arrow Articles by Jackson, D. M
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Speakman, J. R
Right arrow Articles by Jackson, D. M
Agricola
Right arrow Articles by Speakman, J. R
Right arrow Articles by Jackson, D. M
American Journal of Clinical Nutrition, Vol. 86, No. 2, 316-323, August 2007
© 2007 American Society for Nutrition

ORIGINAL RESEARCH COMMUNICATION

Assortative mating for obesity1,2,3

John R Speakman, Kurosh Djafarian, Joanne Stewart and Diane M Jackson

1 From the Aberdeen Centre for Energy Regulation and Obesity (ACERO), Division of Obesity and Metabolic Health, Rowett Research Institute, Aberdeen, Scotland, United Kingdom (JRS, KD, JS, and DMJ), and the ACERO, School of Biological Sciences, University of Aberdeen, Scotland, United Kingdom (JRS)

2 Supported by the Scottish Executive Environmental and Rural Affairs Department.

3 Reprints not available. Address correspondence to JR Speakman, Aberdeen Centre for Energy Regulation and Obesity (ACERO), Division of Obesity and Metabolic Health, Rowett Research Institute, Aberdeen, Scotland, AB21 9SB, United Kingdom. E-mail: j.speakman{at}abdn.ac.uk.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 
Background: Assortative mating is the nonrandom mating of individuals with respect to phenotype and cultural factors. Previous studies of assortative mating for obesity have indicated that it may have contributed to the obesity epidemic. However, those studies all used body mass index or skinfold thicknesses to measure obesity and did not always account for potential confounding factors.

Objective: We aimed to assess the level of assortative mating for obesity by using dual-energy X-ray absorptiometry to characterize body composition.

Design: This was a cross-sectional study of 42 couples.

Results: Raw spousal correlations showed assortative mating for age, weight, body mass index, lean mass, and fat mass. Removing the effect of age on fat mass strengthened the spousal correlation (r = 0.405). Social homogamy did not appear to be important, because in this sample there was no significant effect of area of origin on age-corrected fat and lean tissue masses for either sex. Regional body-composition analysis showed that subjects with disproportionately large arms (both fat and lean) assortatively mated with partners with the same trait. However, both men and women with high lean tissue in their arms assortatively mated with partners that had a disproportionately low fat content in their legs.

Conclusions: These data confirm that assortative mating for obesity exists when dual-energy X-ray absorptiometry is used to evaluate adiposity. We hypothesize that assortative mating may have contributed to the obesity epidemic because the time course of obesity development has shifted progressively earlier, allowing singles in their late teens and early twenties to more easily distinguish partners with obese and lean phenotypes.

Key Words: Assortative mating • obesity • body mass index • dual-energy X-ray absorptiometry • body composition


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 
Assortative mating occurs when men and women do not mate at random with respect to phenotypic and cultural traits. Assortative mating has been shown for height (1, 2), attitude domain of personality (3), education (4-6), religion (4, 6), politics (7), age (7-9), smoking habits (10, 11), and antisocial behavior (12). Assortative mating for obesity has also been inferred in previous studies (6, 9, 13-17), although other studies failed to detect it (18, 19). Assortative mating may have contributed to the obesity epidemic because it increases homozygosity and the likelihood of exposing harmful recessive traits (9). It is impossible to rigorously quantify this contribution unless the exact genetic contribution to the trait in question and the proportion of contributing genes that are recessive are known. These parameters are not yet available for obesity. Nevertheless, we can make some crude estimates of its theoretical effect (Appendix A), which suggests that all things being equal, a switch from random to complete assortative mating would increase the prevalence of obesity in a baseline population from 5% to between 6.6% and 12.6% in just 2 generations. Additionally, assortative mating can result in spurious indications of the role of shared environmental factors in models that attempt to partition the causality of variability in phenotype (13, 16, 20, 21).

There are, however, several problems with previous assessments of assortative mating for obesity. First, almost all previous studies used body mass index (BMI) as a measure of obesity. An exception is the study of Ginsburg et al (15), which used skinfold thicknesses and circumferences. The inadequacies of BMI as a measure of obesity at the individual level are well known (22-27), and these problems may lead either to spurious inferences of assortative mating or underestimates of its significance. Second, the main method for inferring assortative mating is to examine the correlation of given traits between mates (spousal correlation). The spousal correlation, however, is influenced by several other factors that need to be accounted for. Spurious positive associations may occur because obesity is related to other traits that are themselves subject to assortative mating. The most obvious of these is age (13). Because people generally get fatter as they get older (28), and people tend to form relationships with people of their own age (7, 8), taking a cross-section of couples of various ages will generate a spousal correlation in obesity because older persons are generally more obese. Another problem is social homogamy, that is, the tendency for persons to assortatively mate with partners from their own social setting. Because obesity also correlates with social class (29-31), assortative mating may be primarily a consequence of social homogamy. Not all previous studies accounted for these effects. A further difficulty is that couples generally share a common environment for much of their lives. Any spousal correlation in obesity-related traits may then reflect a shared environmental effect (17, 19, 32).

In the present study, we examined assortative mating for obesity and attempted to overcome some of these difficulties. We used dual-energy X-ray absorptiometry (DXA) to measure body composition, thereby overcoming the inadequacies of using BMI. Moreover we accounted for the effects brought about by assortative mating for age, and social homogamy, and the similarity between couples because of the duration occupying a shared environment.


   SUBJECTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 
Study subjects
The project was part of a larger cohort observation (Rascal-Rowett assessment of childhood appetite and metabolism) in which families were recruited from the northeast of Scotland, United Kingdom (Aberdeenshire), through poster and newspaper advertisements to take part in a study investigating familial nutrition. As part of the study, body composition was assessed in 42 couples by use of DXA (Norland XR-26, Mark II high-speed pencil beam scanner; Norland Corporation, Fort Atkinson, WI, equipped with dynamic filtration, with version 2.5.2 of the Norland software). Height was measured with a stadiometer (Holtain Ltd, Crymych, Dyfed, United Kingdom). Weight was measured, with the subjects wearing light clothes, to the nearest 0.1 kg, by using a digital weighing scale (Ohaus, Pine Brook, NJ). All measurements of the family were conducted in the morning after the subjects had fasted overnight. Weekly quality-control checks performed with a phantom over a period of 7 mo indicated CVs of 0.94% and 1.52% for bone mineral density and bone mineral content, respectively. The subjects gave verbal and written consent, and the project was approved by the Grampian local research ethics committee.

Statistical analysis
All data were analyzed by using MINITAB 14 (Minitab Inc, State College, PA). To explore the association between phenotypes of men and women, we used the square of the spousal correlation (r2). We assessed the significance of the associations between men and women for each phenotype by fitting ordinary least-squares regressions and deriving the resultant significance from the F test in the regression analysis of variance. We also derived the gradients of reduced major axis regressions, which are more appropriate for these data for which the x and y traits have equal error variance. In our data set, there was strong assortative mating for age. Because body composition potentially varies with age, spousal associations for body composition might emerge because of the correlation of fatness with age. We therefore explored the effects of age and sex on body composition by using generalized linear modeling. This showed that age was not a significant associate of body composition in our sample. Nevertheless, a weak but not significant association of fatness with age might contribute to a spousal correlation when there is strong assortative mating for age. We therefore fitted regressions of body fatness against age for each sex and derived the residual body fatness corrected for the age of the person. We then examined the spousal correlation in the body fatness corrected for age to establish whether the spousal correlation had been generated because of the strong assortative mating for age in our sample.

In the United Kingdom, geographical areas are classified into postal codes, very much like zip codes in the United States but covering smaller areas. The UK census, which provides comprehensive social information, has been analyzed for each postal code region and these data are available in the public domain. These data show enormous differences in social class between different postal codes over even very small spatial distances. We therefore used postal codes to indicate the social class of the couples. The sample of couples in this study was drawn from 4 different postal code areas. To explore the possibility that we had detected associations because of social homogamy, we examined whether couples varied significantly across these areas of origin with respect to the phenotypes of interest by using one-way analysis of variance with area of origin as the factor.

Finally, we examined the regional body-composition data derived from DXA to explore whether assortative mating for regional body fatness and leanness existed. To explore this possibility, we used the regional lean and fat masses by DXA for the trunk, abdomen, arms, and legs, thus generating 8 traits for each partner. Obviously, subjects that have high total body fat tend to also have high levels of fat in each of their body regions (with such correlations generally having r values around 0.8–0.95), and the same is true of total lean tissue and regional lean tissue masses. Spurious spousal correlations might then emerge because of the overall spousal correlations for total body fat and total-body lean tissue. To eliminate these potential artifacts, we regressed the lean tissue of each region onto the total lean tissue for subjects of each sex and then calculated the residual values to the fitted regression. We repeated this for all 4 body regions in each sex, for both lean and fat tissue (regressing fat tissue mass of each region on total fat tissue mass). These residual values express the extent to which a subject has disproportionate fat or lean tissue amounts in the region of interest. For example, a person with a positive value for leg fat on this scale would have disproportionately large amounts of fat tissue in their legs relative to their overall body fatness. We used these variables to assess whether a person with each characteristic might be involved assortatively in relationships with partners who had similar (or other) regional composition traits. All effects were considered significant when P < 0.05.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
DISCUSSION
 APPENDIX A
 REFERENCES
 
Total body composition
The mean characteristics of the men and women in the sample are shown in Table 1Go. The subjects were typical of the UK population and Western societies in general and included many overweight and obese subjects [the mean BMI (in kg/m2) of the men was 26.2 (range: 20.3–45) and that of the women was 24.6 (range: 20.6–43)]. Simple spousal correlations for the raw traits showed significant positive associations in age, weight, BMI, and total-body lean tissue content and total-body fat tissue content by DXA but not in height or bone mineral content by DXA (Table 2Go and Figure 1Go). The relation between ages included an outlier (male age of 55 y and female age of 30 y; Figure 1Goa); when this point was excluded, the r was increased considerably to 0.84. Even including this outlier, there was strong assortative mating with respect to age in this population. Because weight, BMI, and lean and fat mass by DXA are all traits that potentially increase with age, the spousal correlations in these instances may have occurred because of the associated age effect. In fact, however, for weight and lean tissue mass by DXA, there were significant effects of sex but not of age and no significant sex-by-age interactions; for BMI, neither age or sex nor their interaction was significant (Table 3Go).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Subject characteristics of 42 couples in long-term partnerships recruited into the present study1

 

View this table:
[in this window]
[in a new window]

  		TABLE 2. Raw spousal correlations (r) and significance (P) with the gradient of the reduced major axis regression (bRMA) for significant relations of various phenotypic traits measured in 42 couples from northeast of Scotland1

 

Figure 1
View larger version (24K):
[in this window]
[in a new window]

  		FIGURE 1.. Spousal relations for raw traits: age (r = 0.613, P < 0.001), weight (r = 0.425, P = 0.005), height (r = 0.187, P > 0.05), BMI (r = 0.332, P = 0.032), fat mass (r = 0.386, P = 0.012), lean mass (r = 0.370, P = 0.016), and total bone mineral content (TBMC; r = 0.129, P > 0.05) in 42 couples from northeast Scotland. Descriptive statistics for all traits are shown in Table 2Go.

 

View this table:
[in this window]
[in a new window]

  		TABLE 3. Generalized linear model effects of age and sex on weight, BMI, and lean and fat mass by dual-energy X-ray absorptiometry in 42 couples from northeast Scotland1

 
Although the relation between fat mass and age was not statistically significant, the weak positive association could still contribute to the spousal correlation of fat mass (Figure 1God and Table 2Go). To assess this contribution, we calculated the residual fat mass taking into account the effect of age by fitting a regression to the relation between age and fat mass for each sex. We then examined the spousal correlation of the residual fatness to these fitted regressions (Figure 2Go). With the age effect removed from fat mass, the spousal correlation became stronger (r = 0.405; df = 1,40; P < 0.008). Persons who were fat for their age tended to be in relationships with other persons who were fat for their age.


Figure 2
View larger version (20K):
[in this window]
[in a new window]

  		FIGURE 2.. Residual fat mass in men with the effects of age removed plotted against the same trait in their partners. The association was positive and significant.

 
To explore the possible role of social homogamy, we examined whether the lean and fat masses corrected for age were significantly associated with the area of origin (postal code area) of the couples. In this sample, there were no significant effects of the area of origin of the subjects on the age-corrected fat masses of either the women (df = 3,38; P = 0.704) or the men (df = 3,38, P = 0.467) or in the age-corrected lean masses of the women (df = 3,38; P = 0.264). An effect of area of origin on the age-corrected lean masses of the men approached significance (df = 3,38; P = 0.061).

To assess the extent to which couples might converge because of their shared environment, we examined the difference between their residual fat masses, accounting for the effects of age, against the time that they had been in a relationship and the time that they had been living together. If convergence occurred, the difference between partners would be expected to decline as these durations increased. In both cases, these relations were not significant (time together in the relationship: df = 1,30 and P = 0.432; time living together: df = 1,30 and P = 0.210). Because only 32 of the 42 couples provided information on the durations of their relationships, the failure to detect such an effect might have been a consequence of reduced power because of the lowered sample size. There were strong correlations for these 32 couples between their average ages and both the time they had been in a relationship (r = 0.847) and the time they had been living together (r = 0.886). In the entire data set of 42, therefore, we examined whether there was a relation between their average age and the difference in their residual fat masses. As their average age increased, we would predict the difference in their residual fat masses to get smaller if environmental factors were important. There was a significant relation (Figure 3Go; df = 1,40; P = 0.011), but the trend was positive rather than negative. However, one data point (indicated by an arrow in Figure 3Go) had a high leverage on the regression (Cook's distance = 0.45), and eliminating this point resulted in the regression losing significance (P = 0.099).


Figure 3
View larger version (11K):
[in this window]
[in a new window]

  		FIGURE 3.. The absolute difference between male residual fat mass with age effects removed and the same trait in their partners, plotted against the average age of the couple. The relation was significant and positive, but significance was lost if the arrowed point (which had a high Cooks distance) was removed.

 
Regional body composition
Having shown that persons assortatively mate for total body fatness, even after age effects were allowed for, we used the regional analysis by DXA to explore whether there is additional assortative mating based on particular body regions. Regional body composition did not vary significantly with age for any trait included in the analysis (P > 0.05). For trunk and abdomen with the use of both lean and fat masses, there were no significant spousal correlations, and these traits were also not related to the residual lean and fat masses of the limbs. However, when examining the residual limb values, we found several significant correlations (Table 4Go and Figure 4Go). In particular, we found that there were positive associations between both the residual arm fat tissue and lean tissue contents between men and women. In other words, men and women with arms that contained disproportionately large amounts of fat or lean tissue, relative to total body fat and lean tissue levels, were assortatively mated with partners who had similar disproportionately increased arm fat and lean tissue levels relative to their total body fat and lean tissue levels. That is, persons with disproportionately heavy arms—whether lean or fat—tend to assortatively mate with one another.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Spousal correlations between fat and lean tissue amounts in the limbs (corrected for total body fat and lean tissue, respectively) across 42 couples from northeast Scotland

 

Figure 4
View larger version (30K):
[in this window]
[in a new window]

  		FIGURE 4.. Interrelations between arm fat tissue mass relative to the expectation from total body fat and arm lean tissue mass relative to the expectation from total lean tissue in both men and women in long-term partnerships. All 4 relations were significant. Men and women with disproportionately large arms assortatively mate with partners with similarly large arms. (Note that the scale for female residual arm fatness is expanded to reflect the greater variance relative to that in men.)

 
There were also significant negative spousal correlations between the residual amounts of lean tissue in the arms and the residual amounts of fat tissue in a partner's legs. This was true for both men and women. Hence, if a man or a woman had disproportionately large amounts of lean tissue in his or her arms, he or she tended to be assortatively mated with a partner who had disproportionately low amounts of fat in his or her legs (Table 4Go and Figure 5Go). Clearly, multiple testing is involved in this matrix of correlations. If the Bonferroni correction is applied, only the correlation between disproportionate arm fatness in both men and women remained significant at P < 0.05, but this may be an overly conservative correction.


Figure 5
View larger version (19K):
[in this window]
[in a new window]

  		FIGURE 5.. Interrelations between arm lean tissue content relative to total lean tissue of both men and women and the leg fat tissue of their partners relative to their total body fatness. For both sexes, persons with disproportionately large amounts of lean tissue in their arms were assortatively mated with persons who had disproportionately low amounts of fat in their legs.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
APPENDIX A
 REFERENCES
 
Significant raw spousal correlations were found for age, BMI, and body weight in accord with many previous studies. Despite previous indications that humans also assortatively mate for height (1, 2), we did not confirm this relation in our sample. Previous studies based on relatively small samples like our own also failed to confirm assortative mating for height (4, 9). The strong assortative mating for age occurs primarily because most persons form their first long-term relationships during their late teens and early twenties. At this stage, the potential pool of unattached partners is dominated by persons of the same age. Because we selected for this study couples who were in long-term relationships, the couples involved predominantly met at similar ages and then grew older together, which led to a spousal correlation for age. Undoubtedly, there is probably also assortative mating with respect to age, even in persons who become available during later years after divorce or spousal death, which maintains a strong spousal correlation for age, despite a relatively high divorce rate in Western societies. Clearly, however, these trends can be broken, as illustrated by the outlying couple in our own sample (Figure 1Goa) and the reduced degree of age correlation among divorced and remarried couples (8). Some degree of assortative mating for age in humans is clearly evolutionarily advantageous because it allows significant biparental care over the protracted developmental period of offspring.

Our data also confirm previous suggestions of assortative mating for body weight and BMI on the basis of the raw spousal correlations (6, 9, 13-17). Consequently, despite the limitations of using BMI as an index of obesity, measures of lean body tissue and fat tissue by DXA support the suggestion that assortative mating for obesity exists. Because, in general, older persons get heavier and fatter, the highly significant assortative mating for age can generate spurious illusions of assortative mating for these other traits. Some (13) but not all previous studies accounted for age effects. In our sample, there were no significant effects of age on weight, BMI, or lean or fat mass, so the raw spousal correlations we observed could not have been due to an age artifact. Indeed, when we corrected for age with respect to fat mass, which was the only characteristic with a relation to age approaching significance, the spousal correlation was slightly strengthened. In our sample, there was also no evidence for any social homogamy with respect to obesity, because the age-corrected fat and lean masses of both sexes were not associated with the region of origin of the subjects.

A second potential artifact generating an illusion of assortative mating is the potential for convergence in body composition because of shared environmental effects. If this effect was important the longer a couple lived together, the more one might anticipate they would converge if shared environmental effects were important. In our data, however, we found no evidence for convergence in body composition of body fatness (corrected for age) with the time the couples had been in a relationship and the time they had been living together. This is consistent with many previous studies that indicated a relatively trivial role for shared environmental effects in the etiology of body-composition differences (9, 13, 18). Indeed, Jacobsen et al (14) found that spousal correlations in BMI were strongest among couples with the shortest duration of cohabitation, which is consistent with our own observation of a trend for couples to diverge in body fatness with time together (Figure 3Go: P = 0.011, before eliminating the data point that had a large leverage on relation, P = 0.099 after removal). However, our conclusion differs from that of Salces et al (17), who indicated that shared environment could not be eliminated as a factor influencing spousal correlation for body composition in a population from Spain.

These data therefore confirm the indication from previous studies based on BMI and skinfold thicknesses that a degree of assortative mating for body fatness exists. Why this occurs is not immediately clear from our data. Although some evidence suggests that ratings of physical attractiveness depend on the BMI of the rater, these effects are most pronounced for persons with eating disorders (anorexia nervosa and bulimia nervosa; 33). Ratings of the health status of subjects, which may contribute to attractiveness, are negatively correlated with subject BMI (34) but are independent of the rater's BMI (35). This suggests that couples do not assortatively mate with respect to obesity because lean and obese persons find each other inherently attractive. One possible interpretation is that persons who are physically attractive (on average leaner) pair off with each other first, diminishing the pool of available attractive persons and leaving the remainder (on average more obese) to pair off with each other. Supporting this idea is an enormous literature showing that leanness is correlated with ratings of physical attractiveness (33-42) and that men with greater BMIs are less likely to approach women that they rate as attractive (43). However, it is also clear that decisions concerning life partnerships and friendships are not based on ratings of physical attraction alone (44-48), because physical attractiveness is not always seen in a positive light. Physically attractive men, for example, are more likely to be rated as "likely to carry a sexually transmitted disease" (49), whereas attractive women are more likely to be rated as "likely to be unfaithful" (50) and "neurotically preoccupied with body weight" (51). A direct consequence, for example, is that obese adolescent girls are significantly less likely to be rated as pretty, but this does not appear to affect their popularity among their peers (52).

Nevertheless, accepting the complexity of the situation, if the negative association between physical attractiveness and obesity does contribute to the mechanism for assortative mating with respect to obesity, it provides an explanation for why assortative mating may have contributed to the emerging obesity epidemic. Many years ago, when obesity was less of an issue, the progress toward development of obesity as a function of age was probably much slower. Couples in their late teens and early twenties would therefore have had problems separating currently thin but incipiently obese partners from "real" thin partners, who would remain thin throughout their lives. It is difficult to imagine how assortative mating for obesity could occur in these conditions. However, as the epidemic has progressed, an increasing proportion of the population develops obesity at younger ages (53), while at the same time people are getting married at older ages (54). This makes discriminating obese from lean partners much easier and allows assortative mating to occur. Elevated assortative mating rates may then contribute to the progression of the epidemic (9).

With respect to regional adiposity, the effects we found with respect to positive assortative mating for disproportionate arm lean and fat tissue but negative assortative mating with respect to comparisons of arm lean tissue with leg fat tissue were completely unexpected and have not been previously described. Why such trends exist is uncertain. It will be interesting to see whether such effects will be repeated in future studies. It would be particularly interesting to know whether ratings of physical attractiveness are dependent on both arm and leg sizes of the person being rated and the person doing the rating. Because it is widely suggested that visceral adiposity is a more serious problem than subcutaneous adiposity (55-58), it is perhaps reassuring that the overall effect of assortative mating for obesity does not appear to be magnified by an additional contribution of assortative mating with respect to disproportional abdominal fat tissue.


   APPENDIX A
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
REFERENCES
 
Theoretical quantification of assortative mating effects on the prevalence of obesity
In theory, assortative mating for any polygenically determined trait will have an effect on the prevalence of that trait because bringing persons with the trait together will increase the likelihood that recessive genes of large effect are exposed as homozygotes (9). However, it is impossible to rigorously quantify this effect unless 1) the exact genetic contributions to the trait and 2) the proportion of genes contributing to the trait that are recessive and their individual effects are known. These parameters are not yet available for obesity. We can, however, make some very crude estimates of its theoretical effect in the limiting situations where mating is completely random or mating is completely assortative and where contributory genes are predominantly recessive or dominant.

Consider obesity as a binary trait: persons are either obese or nonobese. The proportion that is nonobese in generation x is Pnonobx, and the proportion that is obese (Pobx) is by definition 1 – Pnonobx. In the situation where mating is random, the proportion of obese persons in generation x + 1 can be defined as

Formula A1(A1)
Where Poff1 is the probability of an offspring from a mating of 2 nonobese persons being obese, Poff2 is the probability of an offspring from a mating of a nonobese person with an obese person being obese, and Poff3 is the probability of an offspring from a mating of 2 obese persons being obese. When mating is completely assortative, the proportion of matings reflects only the proportion of each type in the population. Hence, the proportion of obese persons in generation x + 1 can be defined as

Formula A2(A2)
Consider a starting population (generation x = 0) with an obesity prevalence Pobx of 0.05, and hence Pnonobx = 0.95. For the scenario where most genes contributing to obesity are recessive, we will set the probability of producing obese offspring (Poff1 and Poff2) at 0.05. We will set the probability of producing an obese offspring resulting from an obese person mating with another obese person (Poff3) at 0.75 because of exposure of recessive obesity genes. The prevalence of obesity over 3 generations 1 to 3 in a random mating situation using Equation A1 would be 0.0517 in generation 1, 0.0519 in generation 2, and 0.0519 in generation 3. In contrast, if there was complete assortative mating, the prevalence using equation A2 would increase from 0.05 over the 3 generations to 0.085, 0.109, and 0.126. In the 2 generations that have spanned the current obesity epidemic, a large assortative mating effect of this type could result in a doubling of the prevalence of obesity. In contrast, when most genes are dominant, the respective probabilities of producing obese offspring can be set at Poff1 = 0.03, Poff2 = 0.23, and Poff3 = 0.60 (ie, obese x nonobese matings have a probability of producing obese offspring midway between the obese x obese and nonobese x nonobese matings). In this situation in the population, where there is random mating, obesity prevalence over the 3 generations from equation A1 is 0.0504, 0.0506, and 0.0507, whereas the increase in the assortatively mated population is 0.0585, 0.063, and 0.066.


   ACKNOWLEDGMENTS
 
We thank all the families who took part in the study and the Scottish Executive Environmental and Rural Affairs Department for funding. We are grateful to the anonymous referees for their helpful comments.

All authors provided intellectual input. The other contributions of the authors were as follows—JRS, DMJ, and KD: study design; DMJ, KD, and JS: data collection; and JRS: statistical analysis and writing the first draft of the manuscript. None of the authors had a conflict of interest.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 

  1. Ginsburg E, Livshits G, Yakovenko K, Kobyliansky E. Major gene control of human body height, weight and BMI in five ethnically different populations. Ann Hum Genet 1998;62:307–22.[Medline]
  2. Pawlowski B. Variable preferences for sexual dimorphism in height as a strategy for increasing the pool of potential partners in humans. Proc Biol Sci 2003;270:709–12.[Medline]
  3. Luo SH, Klohnen EC. Assortative mating and marital quality in newlyweds: a couple-centered approach. J Pers Soc Psyc 2005;88:304–26.
  4. Hur YM. Assortative mating for personality traits, educational level, religious affiliation, height, weight, and body mass index in parents of a Korean twin sample. Twin Res 2003;6:467–70.[Medline]
  5. Halpin B, Chan TW. Educational homogamy in Ireland and Britain: trends and patterns. Br J Sociol 2003;54:473–95.[Medline]
  6. al Kandari Y, Crews DE, Poirier FE. Consanguinity and spousal concordance in Kuwait. Collegium Antropologicum 2002;26:1–13.[Medline]
  7. Watson D, Klohnen EC, Casillas A, Simms EN, Haig J, Berry DS. Match makers and deal breakers: analyses of assortative mating in newlywed couples. J Personality 2004;72:1029–68.[Medline]
  8. Gelissen J. Assortative mating after divorce: a test of two competing hypotheses using marginal models. Soc Sci Res 2004;33:361–84.
  9. Hebebrand J, Wulftange H, Goerg T, et al. Epidemic obesity: are genetic factors involved via increased rates of assortative mating? Int J Obes 2000;24:345–53.[Medline]
  10. Sutton GC. Assortative marriage for smoking-habits. Ann Hum Biol 1980;7:449–56.[Medline]
  11. Wilson SE. The health capital of families: an investigation of the inter-spousal correlation in health status. Soc Sci Med 2002;55:1157–72.[Medline]
  12. Smith CA, Farrington DP. Continuities in antisocial behavior and parenting across three generations. J Child Psych Psychiatry 2004;45:230–47.
  13. Allison DB, Neale MC, Kezis MI, Alfonso VC, Heshka S, Heymsfield SB. Assortative mating for relative weight: genetic implications. Behav Genet 1996;26:103–11.[Medline]
  14. Jacobson P, Torgerson JS, Sjostrom L, Bouchard C. Spouse resemblance in body mass index: effects on adult obesity prevalence in the offspring generation. Am J Epidemiol 2007;165:101–8.[Abstract/Free Full Text]
  15. Ginsburg E, Livshits G, Yakovenko K, Kobyliansky E. Genetics of human body size and shape: evidence for an oligogenic control of adiposity. Ann Hum Biol 1999;26:79–87.[Medline]
  16. Silventoinen K, Kaprio J, Lahelma E, Viken RJ, Rose RJ. Assortative mating by body height and BMI: Finnish twins and their spouses. Am J Hum Biol 2003;15:620–7.[Medline]
  17. Salces I, Rebato E, Susanne C. Evidence of phenotypic and social assortative mating for anthropometric and physiological traits in couples from the Basque Country (Spain). J Biosocial Sci 2004;36:235–50.[Medline]
  18. Vogler GP, Sorensen TIA, Stunkard AJ, Srinivasan MR, Rao DC. Influences of genes and shared family environment on adult body-mass index assessed in an adoption study by a comprehensive path model. Int J Obes 1995;19:40–5.[Medline]
  19. Hur YM, Bouchard TJ, Eckert E. Genetic and environmental influences on self-reported diet: a reared-apart twin study. Phys Behav 1998;64:629–36.[Medline]
  20. Rice T, Perusse L, Bouchard C, Rao DC. Familial clustering of abdominal visceral fat and total fat mass: The Quebec Family Study. Obes Res 1996;4:253–61.[Medline]
  21. Magnusson PKE, Rasmussen F. Familial resemblance of body mass index and familial risk of high and low body mass index. A study of young men in Sweden. Int J Obes 2002;26:1225–31.
  22. Pietrobelli A, Faith MS, Heymsfield SB. BMI: does it really reflect body fat mass? Reply. J Pediatr 1999;134:522–3.
  23. Widhalm K, Schonegger K. BMI: does it really reflect body fat mass? J Pediatr 1999;134:522–3.[Medline]
  24. Widhalm K, Schonegger K, Huemer C, Auterith A. Does the BMI reflect body fat in obese children and adolescents? A study using the TOBEC method. Int J Obes 2001;25:279–85.
  25. BMI poor indicator of body fat in individual kids. J Am Diet Assoc 2000;100:628.
  26. Fernandez JR, Heo M, Heymsfield SB, et al. Is the prediction of body fat from BMI in Hispanics different from African Americans and European Americans? Obes Res 2001;9:139S.
  27. Adams T, Lamonte M, Gress R, Hunt S. Gender differences in percent body fat for a given BMI extend into the severely obese. Obes Res 2003;11:A131.
  28. Movahed MR, Ahmadi-Kashani M, Seyed SSA, Kasravi B. The prevalence of increased body mass index and obesity based on age and gender using a large database. Int J Obes 2004;28:S73.
  29. Moore ME, Stunkard A, Srole L. Obesity, social class, and mental illness. Obes Res 1997;5:503–8.[Medline]
  30. Gerald LB, Anderson A, Johnson GD, Hoff C, Trimm RF. Social-class, social support and obesity risk in children. Child Care Health Dev 1994;20:145–63.[Medline]
  31. Silverstone JT. Obesity and social class. Psychother Psychosom 1970;18:226–30.[Medline]
  32. An P, Rice T, Gagnon J, et al. Familial aggregation of resting blood pressure and heart rate in a sedentary population—The HERITAGE Family Study. Am J Hypertens 1999;12:264–70.[Medline]
  33. Tovee MJ, Hancock PJB, Mahmoodi S, Singleton BRR, Cornelissen PL. Human female attractiveness: waveform analysis of body shape. Proc Biol Sci 2002;269:2205–13.[Medline]
  34. Tovee MJ, Maisey DS, Emery JL, Cornelissen PL. Visual cues to female physical attractiveness. Proc Biol Sci 1999;266:211–8.[Medline]
  35. Han TS, Morrison CE, Lean MEJ. Age and health indications assessed by silhouette photographs. Eur J Clin Nut 1999;53:606–11.[Medline]
  36. Fan J, Liu F, Wu J, Dai W. Visual perception of female physical attractiveness. Proc Biol Sci 2004;271:347–52.[Medline]
  37. Forestell CA, Humphrey TM, Stewart SH. Involvement of body weight and shape factors in ratings of attractiveness by women: a replication and extension of Tassinary and Hansen (1998). Personality Ind Diff 2004;36:295–305.
  38. Tovee MJ, Cornelissen PL. Female and male perceptions of female physical attractiveness in front-view and profile. Br J Psychol 2001;92:391–402.[Medline]
  39. Furnham A, Moutafi J, Baguma P. A cross-cultural study on the role of weight and waist-to-hip ratio on female attractiveness. Personality Ind Diff 2002;32:729–45.
  40. Haavio-Mannila E, Purhonen S. Slimness and self-rated sexual attractiveness: comparisons of men and women in two cultures. J Sex Res 2001;38:102–10.
  41. Jackson LA, McGill OD. Body type preferences and body characteristics associated with attractive and unattractive bodies by African Americans and Anglo Americans. Sex Roles 1996;35:295–307.
  42. Singh D. Ideal female body shape—role of body-weight and waist-to-hip ratio. Int J Eat Disord 1994;16:283–8.[Medline]
  43. Clayson DE, Klassen ML. Perception of attractiveness by obesity and hair color. Percep Motor Skills 1989;68:199–202.[Medline]
  44. Brase GL, Walker G. Male sexual strategies modify ratings of female models with specific waist-to-hip ratios. Human Nature-An Interdisciplinary Biosocial Perspective 2004;15:209–24.
  45. Greitemeyer T, Fischer P. Tell me how beautiful and dominant my partner is: the effects of evaluations of a friendly person regarding physical attractiveness and dominance of a potential partner on perceived contact-willingness. Z Sozialpsychologie 2004;35:231–9.
  46. Urbaniak GC, Kilmann PR. Physical attractiveness and the "nice guy paradox": do nice guys really finish last? Sex Roles 2003;49:413–26.
  47. Townsend JM, Wasserman T. Sexual attractiveness: sex differences in assessment and criteria. Evol Hum Behav 1998;19:171–91.
  48. Lundy DE, Tan J, Cunningham MR. Heterosexual romantic preferences: the importance of humor and physical attractiveness for different types of relationships. Personal Relationships 1998;5:311–25.
  49. Smith CA, Stillman S. What do women want? The effects of gender and sexual orientation on the desirability of physical attributes in the personal ads of women. Sex Roles 2002;46:337–42.
  50. Zaromatidis K, Carlo R, Racanello D. Sex, perceptions of attractiveness, and sensation seeking and ratings of the likelihood of having sexually transmitted diseases. Psychol Rep 2004;94:633–6.[Medline]
  51. Singh D. Mating strategies of young women: role of physical attractiveness. J Sex Res 2004;41:43–54.[Medline]
  52. Davis C, Claridge G, Fox J. Not just a pretty face: physical attractiveness and perfectionism in the risk for eating disorders. Int J Eat Disord 2000;27:67–73.[Medline]
  53. Phillips RG, Hill AJ. Fat, plain, but not friendless: self-esteem and peer acceptance of obese pre-adolescent girls. Int J Obes 1998;22:287–93.[Medline]
  54. Frelut ML, Cathelineau L, Bihain BE, Navarro J. [Trends in prevalence in childhood obesity throughout the world—which perspectives. ] Arch Pediatr 1995;2:1124–5 (in French).[Medline]
  55. Schoen R, Weinick RM. The slowing metabolism of marriage: figures from 1988 U.S. marital status life tables. Demography 1993;30:737–46.
  56. Kabir M, Catalano KJ, Ananthnarayan S, et al. Molecular evidence supporting the portal theory: a causative link between visceral adiposity and hepatic insulin resistance. Am J Physiol 2005;288:E454–61.
  57. You TJ, Ryan AS, Nicklas BJ. The metabolic syndrome in obese postmenopausal women: relationship to body composition, visceral fat, and inflammation. J Clin Endocrinol Metab 2004;89:5517–22.[Abstract/Free Full Text]
  58. Onat A, Avci GS, Barlan MM, Uyarel H, Uzunlar B, Sansoy V. Measures of abdominal obesity assessed for visceral adiposity and relation to coronary risk. Int J Obes 2004;28:1018–25.[Medline]
  59. Hayashi T, Boyko EJ, Leonetti DL, et al. Visceral adiposity is an independent predictor of incident hypertension in Japanese Americans. Ann Intern Med 2004;140:992–1000.[Abstract/Free Full Text]
Received for publication March 9, 2007. Accepted for publication April 9, 2007.





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Speakman, J. R
Right arrow Articles by Jackson, D. M
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Speakman, J. R
Right arrow Articles by Jackson, D. M
Agricola
Right arrow Articles by Speakman, J. R
Right arrow Articles by Jackson, D. M

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS