Genetic Influences on Adolescent Eating Habits

412776 6Beaver et al.Health Education & Behavior

HEBXXX10.1177/109019811141277

Genetic Influences on Adolescent Eating Habits

Health Education & Behavior 39(2) 142–151 © 2012 Society for Public Health Education Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/1090198111412776 http://heb.sagepub.com

Kevin M. Beaver, PhD1,Tori Flores, BS1, Brian B. Boutwell, PhD2, and Chris L. Gibson, PhD3

Abstract Behavioral genetic research shows that variation in eating habits and food consumption is due to genetic and environmental factors. The current study extends this line of research by examining the genetic contribution to adolescent eating habits. Analysis of sibling pairs drawn from the National Longitudinal Study of Adolescent Health (Add Health) revealed significant genetic influences on variance in an unhealthy eating habits scale (h2 = .42), a healthy eating habits scale (h2 = .51), the number of meals eaten at a fast-food restaurant (h2 = .33), and the total number of meals eaten per week (h2 = .26). Most of the remaining variance was due to nonshared environmental factors. Additional analyses conducted separately for males and females revealed a similar pattern of findings. The authors note the limitations of the study and offer suggestions for future research. Keywords Add Health, adolescence, behavioral genetics, eating habits, genetics, sibling pairs The effects that adolescent eating habits have on immediate and later-life outcomes are becoming well known. Unhealthy eating is linked to poorer nutritional profiles, a greater risk for being overweight, and decreased cognitive skills in adolescence (Elias, Elias, Sullivan, Wolf, & D’Agostino, 2003; Huang et al., 2003; Nicklaus, Baranowski, Cullen, & Berenson, 2001). Longitudinally, childhood and adolescent eating habits have been found to be predictive of adult body mass index, risk of heart disease, and likelihood of developing diabetes (Niemeier, Raynor, Lloyd-Richardson, Rodgers, & Wing, 2006; Siega-Riz, Carson, & Popkin, 1998). Even though the negative consequences associated with eating habits are consistently replicated, there remains a substantial percentage of youths who do not practice healthy eating habits. Estimates from recent studies indicate that less than 50% of all youths consume the recommended daily allowance of fruits and vegetables and more than 20% of boys and girls consume extremely high levels of fatty foods and sugary drinks (Enns, Mickle, & Goldman, 2003; Neumark-Sztainer, Wall, Perry, & Story, 2003). Identifying and understanding the factors that promote variation in adolescent eating habits is therefore extremely important in the quest to increase the number of youths who consume the necessary nutritional allotments to support healthy human development. In comparison with the research examining the effects of adolescent healthy eating habits, much less research has examined the factors that are predictive of variation in a number of different eating habits. Much of the research examining the factors associated with eating habits has focused on family-level factors or demographic characteristics. A small pool of behavioral genetic research, however, has begun to

examine whether genetic factors may also be involved in shaping adolescent eating habits. Behavioral geneticists analyze samples of kinship pairs to examine the genetic and environmental underpinnings to phenotypic variation. In doing so, the variance in phenotypes is attributed to three different sources: heritability, the shared environment, and the nonshared environment. Heritability captures the proportion of phenotypic variance that is the result of genetic variance and is frequently denoted as h2. The shared environmental component includes the factors that are the same between siblings and that contribute to sibling similarity and is often symbolized as c2. And, the nonshared environment captures the effects of all environmental factors that are unique to each sibling plus the effects of measurement error and is captured by the symbol e2. Most behavioral geneticists employ samples of twin pairs in their analysis. The use of twin pairs allows for a relatively straightforward way to estimate heritability, shared environmental effects, and nonshared environmental effects. There are two different types of twin pairs: monozygotic (MZ) twin pairs and dizygotic (DZ) twin pairs. MZ twins share 100% of their DNA, whereas DZ twins share, on average, 50% of their distinguishing DNA. The environments for both MZ and DZ twins, however, are the same. As a result, the only reason that 1

Florida State University, Tallahassee, FL, USA Sam Houston State University, Huntsville, TX, USA 3 University of Florida, Gainesville, FL, USA 2

Corresponding Author: Kevin M. Beaver, College of Criminology and Criminal Justice, Florida State University, 634 W. Call Street, Tallahassee, FL 32306-1122, USA Email: [email protected]

143

Beaver et al. MZ twins should be more similar to each other on a phenotype is because they share more genetic material when compared with DZ twins. And, the greater the similarity of MZ twins when compared with DZ twins, the greater the genetic effect. This same logic can be extended to other types of sibling pairs, including regular siblings, half-siblings, and unrelated siblings (i.e., stepsiblings). Genetic effects on phenotypic variance, in short, are detected when the degree of phenotypic similarity increases as a function of the degree of genetic similarity. A vast amount of twin-based research has been conducted to estimate the genetic and environmental effects on virtually every measurable phenotype. Although the precise estimates of heritability, shared environmental effects, and nonshared environmental effects vary across studies and across phenotypes, the common theme across studies is that genes are related to almost all phenotypes (Turkheimer, 2000). Whether these findings would extend to a variety of adolescent eating habits remains to be determined. Extant research, however, does tend to support the joint role genetic and environmental factors play in producing variation in adolescent eating habits. Three studies stand out as particularly noteworthy. In the first study, Breen, Plomin, and Wardle (2006) analyzed a sample of 214 twin pairs between the ages of 4 and 5 years to estimate genetic effects on dessert foods, vegetables, fruits, and protein foods. Results revealed significant genetic influences on all four measures, with the heritability of eating dessert foods = .20, the heritability of eating vegetables = .37, the heritability of eating fruits = .71, and the heritability of eating protein foods = .78. Most of the remaining variance was explained by the shared environment. The second study, carried out by Keskitalo et al. (2008), analyzed a sample of more than 2,000 pairs of young adult twin pairs to estimate genetic influences on food use patterns. Analysis of these data indicated that genetic factors explained approximately 40% to 45% of the variance in measures of healthy foods, high-fat foods, sweet foods, and meat. The remaining variance was attributable to the nonshared environment. Last, Hur, Bouchard, and Eckert (1998) examined genetic and environmental influences on diet by analyzing a sample of reared-apart twins. In line with the other studies, they reported a significant genetic effect with approximately 20% to 30% of the variance in diet being attributable to genetic factors. Once again, the remaining variance was accounted for by nonshared environmental factors. Other studies have detected significant genetic effects on caloric intake (de Castro, 1993), food frequency (van den Bree, Eaves, & Dwyer, 1999), and other measures tapping nutrition (Steinle et al., 2002).

Current Study The current study expands research on the genetics of eating habits in four ways. First, unlike most studies that use twin samples drawn from other countries or from samples that are not nationally representative (Breen et al., 2006; Keskitalo et al., 2008), we employ a sample of kinship pairs that is drawn

from a nationally representative sample of American youths, which allows for a more thorough comparison of the effects across different studies. Second, unlike previous studies that examine samples of young children or adults (Breen et al., 2006; Hur et al., 1998; Keskitalo et al., 2008), our sample includes adolescents. The results of our study will thus provide evidence shedding light on whether genetic effects on eating habits ebb and flow across different developmental time periods. Third, we investigate multiple eating habits, including an unhealthy eating habits scale, a healthy eating habits scale, the number of days each week that a respondent ate fast food, and the total number of meals eaten each week. By using four scales, we are able to examine whether different eating habits are differentially affected by genetic factors. Fourth, following prior research (Keskitalo et al., 2008), we also examine whether heritability varies for males and females.

Method Study Population Data for this study come from the National Longitudinal Study of Adolescent Health (Add Health; Udry, 2003), which is a longitudinal study of a nationally representative sample of American youths. Four waves of data have been collected thus far. Initial data collection began in 1994-1995 when more than 90,000 adolescents in middle and high school were administered self-report surveys during a specified school day (i.e., the Wave 1 in-school survey). To gain more detailed information, a subsample of youths was selected to be reinterviewed at their homes along with their primary caregiver (i.e., the Wave 1 in-home survey). Questions were asked about the adolescents’ involvement in delinquency, their sexual experiences, and their social relationships. A total of 20,745 youths and 17,700 of their primary caregivers participated in the Wave 1 in-home survey (Harris et al., 2003). Approximately 1 to 2 years later, the second wave of data was collected. Because relatively little time had lapsed since Wave 1, most of the respondents were still adolescents. As a result, the survey instruments remained very similar between waves. For example, youths were once again asked about their involvement in risky behaviors, their use of drugs and alcohol, and their eating habits. Overall, 14,738 adolescents were successfully interviewed at Wave 2. Then, during 2001-2002, the third wave of interviews was conducted when most of the 15,197 respondents were young adults. The changing age structure of the sample necessitated a change to the surveys. Participants, for instance, were asked about their employment history, their educational achievements, and their lifetime contact with the criminal justice system (Harris et al., 2003). The fourth and final wave of data was collected in 2008 when 15,701 were interviewed. Respondents were asked a wide range of questions pertaining to adulthood, such as their experiences with childrearing, their highest level of education, and their marital status.

144

Health Education & Behavior 39(2)

Table 1. Descriptive Statistics for Sibling Pair Types and Demographics Frequency (Mean) Sibling type MZ twin DZ twin Full sibling Half-sibling Cousins Unrelated Gender Female Male Race Caucasian Minority Age (at Wave 1)

Percentage (SD)

246 386 981 318 119 267 1,182 1,135 1,412 905 (16.01)

10.6 16.7 42.3 13.7 5.1 11.5 51.0 49.0 60.9 39.1 (1.75)

Note. MZ = monozygotic; DZ = dizygotic.

One of the unique aspects of the Add Health data is that a sample of sibling pairs is embedded within the data. During Wave 1 interviews, respondents were asked whether they currently lived with a co-twin, a half-sibling, a stepsibling, or a cousin. If they responded affirmatively, and if their sibling was between the ages of 11 and 20 years, then they were also added to the study. In addition, a probability sample of full siblings was also included in the data (Jacobson & Rowe, 1999). Overall, more than 3,000 pairs of siblings were included in the data (Harris, Halpern, Smolen, & Haberstick, 2006). Table 1 contains the breakdown of the types of sibling pairs that were employed in the current analysis as well as the demographic characteristics of the siblings.

Measures Unhealthy eating habits. During Wave 2 interviews, respondents were asked a wide range of in-depth questions about their eating habits. We were able to use 20 items to create an unhealthy eating habits scale. Specifically, respondents were asked to think about everything they had to eat and drink the day before the interview. They were then presented with a list of foods and drinks and asked to indicate whether they had eaten that food or drank that beverage the day before. Follow-up questions were then asked to discern the type of food/beverage that was consumed. For example, respondents were asked whether they drank soft drinks or mixers, such as tonic water or club soda. Rather than simply using this item to gauge unhealthy eating habits, we also examined the followup question that asked whether the drinks were regular, diet or sugar free, or both. We opted to dichotomize the measure such that 0 = the food/beverage was diet/sugar-free/low-fat and 1 = the food/beverage was regular (i.e., not a diet food/ drink). In instances where the respondent indicated that the food/beverage was a combination, the response was coded as

Table 2. Descriptive Statistics for the Dietary Habits Scales

Unhealthy eating habits Healthy eating habits Number of days eating fast food Number of meals eaten each week

Mean

SD

Min-Max

4.99 4.79 2.28 15.89

2.74 2.94 1.82 4.74

0-18 0-23 0-7 0-21

a zero (0). This coding scheme allowed for a more conservative estimate of the unhealthy eating habits of adolescents. Responses to the 20 items were then summed together, where higher values represent more unhealthy eating habits (α = .61). Similar types of scales have been used previously to measure eating habits (e.g., Stewart & Menning, 2009; Videon & Manning, 2003). The appendix contains all the individual items that were included in this scale. Table 2 provides descriptive statistics for the unhealthy eating habits scale along with the other dietary habits scales employed in the analyses. Healthy eating habits. During Wave 2 interviews, respondents were also asked a series of questions that were used to create a healthy eating habits scale. Twenty-three items were identified as measuring healthy eating habits and these items were scored in a similar fashion as those that were used to create the unhealthy eating habits scale. In particular, youths were first asked to indicate whether they ate/drank specific foods/beverages the previous day. For example, adolescents were asked whether they drank 100% orange, grapefruit, or tomato juice, and they were also asked whether they had eaten broccoli and carrots. Responses to these items were coded 0 = no and 1 = yes. The items were then summed together to create the healthy eating habits scale, where higher values represent more healthy eating habits (α = .71). See the appendix for a listing of all the items included in this scale. Number of days eating fast food. Prior research has revealed that adolescents who eat extensively at fast-food restaurants are at risk for a range of health-related outcomes both in adolescence and later in life (Niemeier et al., 2006). To examine the extent to which genetic and environmental factors affect eating fast food, we included a one-item variable that measured the number of days that each respondent ate at a fast-food restaurant. Specifically, during Wave 2 interviews, respondents were asked how many days in the last 7 days they ate at a fastfood-type place, such as McDonalds, Kentucky Fried Chicken, Pizza Hut, Taco Bell, and so on. Responses to this item ranged between zero (0) and seven (7), with the value indicating the number of days in the past week that they ate at a fast-food restaurant. Importantly, previous researchers analyzing the Add Health data have used this same variable to measure frequency of fast food consumption (Niemeier et al., 2006; Stewart & Menning, 2009). Number of meals eaten each week. The number of meals eaten each week varies significantly among adolescents (Cusatis & Shannon, 1996); however, research has revealed that maintaining a consistent pattern of meal consumption promotes healthy human development (Siega-Riz et al., 1998). As a

145

Beaver et al. result, we included a variable that measures the number of meals eaten each week. During Wave 2 interviews, adolescents were asked how many days in the last week they ate breakfast, lunch, and dinner. Responses to these three questions ranged from 0 (indicating no days) to 7 (indicating every day of the week). The three items were then summed together to create a variable that measures the total number of meals eaten each week by the adolescent. Previous research has measured the number of meals eaten each week in this same way (Stewart & Menning, 2009).

Statistical Analyses The current study employs DeFries–Fulker (DF) analysis to estimate genetic and environmental influences on adolescent eating habits. DF analysis can be used with samples of kinship pairs to provide estimates of the amount of variance explained by shared environmental factors, nonshared environmental factors, and genetic factors (DeFries & Fulker, 1985). The original DF equation was designed to be used with clinical samples, where one sibling had been diagnosed with a particular disorder. The original DF equation has since been amended so that it can be used with community samples (Rodgers & Kohler, 2005; Rodgers, Rowe, & Li, 1994). The DF equation used in the current study takes the following form: K1 = b0 + b1(K2–Km) + b2[R * (K2– Km)] + e, 1 where K1 is the score for one sibling on one of the eating habits scales/variables, K2 is the score for their sibling on that same eating habits scale/variable, R measures genetic relatedness (R = 1 for MZ twins, .5 for DZ twins and full siblings, .25 for half-siblings, .125 for cousins, and .00 for unrelated siblings), and Km is the mean on the eating habits scale/variable for K2. Equation 1 indicates that K2 is centered to its mean. The interpretation of the coefficients is as follows: b1 = the proportion of variance in the eating habits scale/variable that is attributable to shared environmental effects, b2 = the proportion of variance in the eating habits scale/variable explained by genetic effects, and e = the proportion of variance in the eating habits scale/variable that is explained by nonshared environmental effects plus error. All the models were calculated for the full sample and then separately for sibling pairs where both siblings were females and sibling pairs where both siblings were males. Two reasons informed our decision to calculate sex-specific models. First, difference in means t tests revealed that males, when compared with females, scored significantly higher on the unhealthy eating habits scale (t = 9.4, p < .05), self-reported eating more at fast-food restaurants (t = 2.2, p < .05), and self-reported eating more meals each week (t = 4.3, p < .05). The only nonsignificant difference between males and females was for the healthy eating habits scale. Second, calculating sex-specific models allows for the exploration of potential gender differences

in the genetic and environmental underpinnings to adolescent eating habits.

Results The analysis begins by estimating the proportion of variance in the unhealthy eating habits scale that is accounted for by genetic factors, shared environmental factors, and nonshared environmental factors for the full sample and separately by gender. Figure 1 plots the results of the DF analyses. As can be seen, heritability accounts for 42% of the variance in unhealthy eating habits for the full sample, shared environmental factors account for 0% of the variance, and nonshared environmental factors explain the remaining 58% of the variance. Similar results were garnered for the all-female and the all-male samples. For example, the heritability of unhealthy eating habits for females was .46 and .32 for males, whereas nonshared environmental effects explained the remaining variance; for both female and males, the shared environmental effect was 0. The next sets of models are exact duplicates of those estimated in Figure 1 except that the healthy eating habits scale is employed instead of the unhealthy eating habits scale. Figure 2 shows the results of these models and reveals findings that are consistent with those from the unhealthy eating habits scale. Specifically, the heritability of the healthy eating habits scale was .51 for the full sample, .63 for the all-female sample, and .38 for the all-male sample. Across all three models, the shared environment explained none of the variance in healthy eating habits, whereas the nonshared environment explained the remaining variance that was unaccounted for by heritability. We next turn our attention to the DF analyses that estimated the genetic and environmental effects on variance in the variable measuring the number of days that the adolescent ate fast food each week. Figure 3 shows the results of these models, which indicate that, for the full sample, genetic factors explained 33% of the variance, shared environmental factors explained 9% of the variance, and nonshared environmental factors explained 58% of the variance. For females, a slightly different pattern of results emerged, wherein genetic factors explained 62% of the variance, shared environmental factors explained 0% of the variance, and nonshared environmental factors accounted for 38% of the variance. The DF model that was calculated on the all-male sample revealed that genetic factors explained 43% of the variance, the shared environment explained 0% of the variance, and the nonshared environment explained 57% of the variance. The last series of models, which are depicted in Figure 4, present the results of the DF analyses examining the genetic and environmental effects on variance in the total number of meals eaten each week by the adolescent. The estimates vary considerably across the three models. For the full sample, genetic factors explained 26% of the variance, shared environmental factors accounted for 15% of the variance, and nonshared environmental factors explained the remaining

146

Health Education & Behavior 39(2)

100%

Percentage of Variance Explained

90% 80% 70% 60%

Nonshared Environment Shared Environment Heritability

50% 40% 30% 20% 10% 0%

Full Sample

Females

Males

Figure 1. DeFries–Fulker analysis results for the unhealthy eating habits scale

Note. For the full sample h2 = .42, c2 = .00, and e2 = .58; for the female sample h2 = .46, c2 = .00, and e2 = .54; for the male sample h2 = .32, c2 = .00, and e2 = .68.

100%


90% 80% 70% 60%


50% 40% 30% 20% 10% 0%

Full Sample

Females

Figure 2. DeFries–Fulker analysis results for the healthy eating habits scale

Males


147

Beaver et al.

100%


90% 80% 70% 60%


50% 40% 30% 20% 10% 0%

Full Sample

Females

Males

Figure 3. DeFries–Fulker analysis results for the number of days eating fast food each week


100%


90% 80% 70% 60%


50% 40% 30% 20% 10% 0%

Full Sample

Females

Males

Figure 4. DeFries–Fulker analysis results for total number of meals eaten each week


59% of the variance. In the all-female sample, genetic factors accounted for none of the variance and shared environmental factors explained 39% of the variance, with the nonshared environment accounting for 61% of the variance.

In contrast, for the all-male sample, one half of the variance was attributable to genetic factors and one half of the variance was attributable to nonshared environmental factors.

148

Discussion Findings culled from behavioral genetic research consistently show that genetic factors explain a significant amount of variance in virtually every human phenotype (Turkheimer, 2000). Only recently have behavioral genetic methodologies been employed to examine eating habits and food consumption. The results of these studies have revealed that genetic factors play an integral role in explaining variance in these different types of food intake. We expanded on this line of research by estimating genetic and environmental influences on four different measures of adolescent eating habits by analyzing a sample of sibling pairs. Analysis of the Add Health data produced two main findings. First, genetic factors explained a significant amount of variance in the unhealthy eating habits scale, the healthy eating habits scale, the number of meals eaten at fast-food restaurants variable, and the total number of meals eaten per week variable. Most of the remaining variance was accounted for by nonshared environmental effects. These findings dovetail with those reported in other studies that analyzed different measures of eating habits that were collected from samples of different age ranges (e.g., children and adults; Breen et al., 2006; de Castro, 1993). Given that molecular genetic research has identified certain genes that are related to how the texture of food is sensed and even how food tastes (Tepper, 1998), it is not surprising that genetic effects are detected across eating habits scales and even across different age-groups. In addition, it is not surprising that the remaining variance was accounted for by nonshared environments. Recall that nonshared environments are environments that make siblings different from each other. One of the nonshared environments that is consistently viewed as the most salient during the teenage years is the peer group (Harris, 1998). During adolescence, peers become the dominant socializing agent and peer groups have been shown to affect teenage eating habits (Adams, 1997). Thus, it is quite conceivable that much of the variance accounted for by nonshared environments is the result of the values and norms of certain peer groups that ultimately affect eating habits. Plus, adolescents are able to escape the constant surveillance of their parents and oftentimes eat meals away from home. All these extrafamilial activities that ultimately affect eating habits would be captured by the nonshared component of the DF model. The second key finding to emerge from these analyses is that in general there were not significant differences in the parameter estimates between females and males across the eating habits measures. While heritability estimates did fluctuate, difference in coefficient z tests indicated that these differences were not statistically significant. The one exception to this general finding, however, was with the total number of meals eaten each week variable. For females, none of the variance was accounted for by genetic factors, 39% was accounted for by shared environmental factors, and 61% was the result of

Health Education & Behavior 39(2) nonshared environmental factors. In contrast, for males, none of the variance was accounted for by the shared environment, 50% was accounted for by genetic factors, and 50% was accounted for by nonshared environmental factors. The precise reasons for why the parameter estimates between females and males should only vary for the number of meals eaten each week remains unknown. Perhaps the larger shared environmental effect for females versus males is capturing family-wide pressures instilled in some families that weight and beauty are critically important, especially for females. For males in these families perhaps the emphasis is placed on other outcomes, such as athletics or even academic success. In this way, families may be emphasizing different factors for their sons and daughters. If males are not socialized by their parents with regard to eating habits, then their genetic predispositions will not be dampened and thus heritability estimates will be higher for males than females. Also of importance is that the nonshared environment was greater for females than males (although this difference was not statistically different) for the total meals eaten per week variable. Given that female peer groups often place an extraordinary amount of emphasis on thinness and beauty, while male peer groups place more emphasis on toughness and aggression, the peer group would likely have a larger effect on eating habits for females than for males (Klump, McGue, & Iacono, 2000; Thompson, Coovert, Richards, Johnson, & Cattarin, 1995). While the results of the analyses did not fully support this possibility (as the nonshared estimates were approximately equal between males and females), future research should explore this line of inquiry in greater detail. While the results generated from this study reinforce those found in other studies, the study is host to a number of limitations. First, although the Add Health data are nationally representative, the sibling pairs data that were employed in the current study are not necessarily nationally representative. As a result, the findings reported here may not extend to other American adolescents or to adolescents from other countries. Importantly, however, prior research has failed to detect any significant differences in demographic, behavioral, and personality measures between the full Add Health sample and the sample of sibling pairs (Beaver, 2008; Jacobson & Rowe, 1998). Second, all the eating habits measures were derived from self-reports based on retrospective recollections. If genes affect survey responses, then the heritability estimates for the eating habits scales may actually be measuring genetic influences on survey responses. Third, the data analyzed in this study were collected only at a single time period in adolescence. Although previous studies have detected similar genetic effects on eating habits at different periods of the life course, more research is needed to determine whether genetic factors have relatively consistent effects across the life or whether the effects change. Until these issues are addressed, and replication studies are undertaken, the findings of the current study should be viewed with caution.

149

Beaver et al.

Conclusions In recent months, momentum has been building in the Obama administration’s “Lets Move” initiative to reduce obesity and overweight problems in the United States, especially among children and adolescents (http://www.letsmove.gov/). The administration’s focus on childhood obesity, which emphasizes the importance of increasing healthy eating habits while reducing the intake of unhealthy foods, is timely and long overdue. Over the past several decades, childhood obesity rates have nearly tripled, caloric intake among children has increased substantially, and fatty foods with sugar and artificial sweeteners are more commonly consumed (Ogden, Carroll, Curtin, Lamb, & Flegal, 2010). These trends in food consumption and weight gain, coupled with decreases in physical activity among children in the United States, are particularly troublesome due to the developmental health risks associated with them. Children who are overweight are at risk for developing a host of physical health problems including hypertension, heart disease, type 2 diabetes, gall bladder disease, and sleep disturbance disorders (Daniels et al., 2005; Freedman, Mei, Srinivasan, Berenson, & Dietz, 2007; U.S. Surgeon General, 2001). Realizing the importance of this national problem, research designed to unpack the environmental and genetic contributions to unhealthy and healthy eating behaviors is a first step for preventing obesity and its potential longterm consequences for children in the United States. The current study was an initial step in this direction. As the findings of the current study reveal, a significant amount of variance in eating habits is influenced by genetic factors. This necessarily begs the question of whether policies and programs, such as the state-based programming initiatives funded by the Center Disease Control and the U.S. Department of Health and Human Services that target unhealthy eating habits, obesity, and dietary nutrition enhancement can be successful. Genetic effects have been shown to change over time, and moreover, they can change in response to the environmental conditions (Caspi et al., 2002). Seen in this way, genetic effects, including the genetic effects on adolescent eating habits, can be affected by manipulating the environment. Assessing how children’s genetic propensities may make them more susceptible to environmental efforts to reduce unhealthy eating and obesity is a challenge for future research. Such environmental efforts to be considered should be multilayered and should include a child’s immediate environment at home with parents, neighborhoods and community organizations, peer groups, and school initiatives. Even large governmental policies have been shown to alter genetic effects on health-related outcomes. Boardman (2009), for instance, analyzed a sample of sibling pairs from the Add Health to examine whether the genetic effects on smoking varied across different states. The results of his analyses revealed that genetic effects were attenuated significantly in states that had relatively high taxes on tobacco products.

Similar logic could be used in relation to adolescent eating habits. While eating habits are almost most certainly determined, in part, by genetic factors, these genetic effects can be dampened through different types of policies and public advertisements. The extent to which any policy is effective obviously hinges on a number of different factors, but to a large extent it hinges on our understanding of the target behavior that policies are trying to change. The more that we learn about both the environmental and genetic underpinnings to adolescent eating habits, the better position we will be in to promote healthy eating habits and hopefully reduce unhealthy eating habits.

Appendix Items Included in the Unhealthy and Healthy Eating Habits Scale Unhealthy Eating Habits Scale. Think about everything you had to eat and drink yesterday. Did you eat/drink . . . 1. regular soft drinks or mixers? 2. regular Koolaid, fruit-flavored drinks, or sport drinks? 3. regular breakfast bars or breakfast tarts? 4. regular doughnuts, sweet rolls, muffins, or pastries? 5. regular hot dogs or frankfurters? 6. regular yogurt or cottage cheese? 7. regular cheese, processed cheese, or cheese spreads? 8. regular cookies, brownies, cake, or pie? 9. regular ice cream? 10. regular frozen yogurt? 11. regular butter or margarine? 12. regular salad dressing? 13. regular mayonnaise or sandwich spread? 14. regular ground meat or hamburger? 15. pizza? 16. bacon, sausage, or chorizo? 17. French fries? 18. chocolate bars or candy? 19. fried chicken or turkey? 20. fried fish or seafood?

Healthy Eating Habits Scale. Think about everything you had to eat and drink yesterday. Did you eat/drink . . . 1. 100% orange, grapefruit, or tomato juice? 2. water? 3. apples, applesauce, pears, or pineapple? 4. bananas, plantains, grapes, berries, or cherries? 5. cantaloupes, melons, mangoes, or papayas? 6. oranges, grapefruit, tangerines, or kiwis? 7. peaches, plums, nectarines, or apricots? 8. raisins or dried fruits? (continued)

150

Appendix (continued) 9. mixed vegetables or acorn, hubbard, or winter squash? 10. avocadoes? 11. string beans, green beans, peas, or snow peas? 12. cabbage or bok choy? 13. broccoli? 14. carrots? 15. dried beans, peas, lentils, black beans, or soybeans? 16. field peas, chick peas, or lima beans? 17. kale, beet greens, mustard greens, turnip greens, or collard greens? 18. lettuce or tossed salad? 19. spinach? 20. tomatoes? 21. tofu? 22. yams or sweet potatoes? 23. zucchini, summer squash, eggplants, bell peppers, or asparagus? Authors’ Note This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. Special acknowledgment is due to Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Persons interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123 W. Franklin Street, Chapel Hill, NC27516-2524 ([email protected]). No direct support was received from grant P01-HD31921 for this analysis.

Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The authors received no financial support for the research, authorship, and/or publication of this article.

References Adams, L. B. (1997). An overview of adolescent eating behavior barriers to implementing dietary guidelines. Annals of the New York Academy of Sciences, 817, 36-48. Beaver, K. M. (2008). Nonshared environmental influences on adolescent delinquent involvement and adult criminal behavior. Criminology, 46, 341-369. Boardman, J. D. (2009). State-level moderation of genetic tendencies to smoke. American Journal of Public Health, 99, 480-486. Breen, F. M., Plomin, R., & Wardle, J. (2006). Heritability of food preferences in young children. Physiology and Behavior, 88, 443-447.

Health Education & Behavior 39(2) Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., . . . Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297, 851-854. Cusatis, D. C., & Shannon, B. M. (1996). Influences on adolescent eating behavior. Journal of Adolescent Health, 18, 27-34. Daniels, S. R., Arnett, D. K., Eckel, R. H., Gidding, S. S., Hayman, L. L., Kumanyika, S., . . . Williams, C. L. (2005). Overweight in children and adolescents: Pathophysiology, consequences, prevention, and treatment. Circulation, 111, 1999-2002. de Castro, J. M. (1993). Genetic influences on daily intake and meal patterns of humans. Physiology and Behavior, 4, 777-782. DeFries, J. C., & Fulker, D. W. (1985). Multiple regression analysis of twin data. Behavior Genetics, 15, 467-473. Elias, M. F., Elias, P. K., Sullivan, L. M., Wolf, P. A., & D’Agostino, R. B. (2003). Lower cognitive function in the presence of obesity and hypertension: The Framingham Heart Study. International Journal of Obesity, 27, 260-268. Enns, C. W., Mickle, S. J., & Goldman, J. D. (2003). Trends in food and nutrient intakes by adolescents in the United States. Family Economics and Nutrition Review, 15, 15-28. Freedman, D. S., Mei, Z., Srinivasan, S. R., Berenson, G. S., & Dietz, W. H. (2007). Cardiovascular risk factors and excess adiposity among overweight children and adolescents: The Bogalusa Heart Study. Journal of Pediatrics, 150, 112-117. Harris, J. R. (1998). The nurture assumption: Why children turn out the way they do. New York, NY: The Free Press. Harris, K. M., Florey, F., Tabor, J., Bearman, P. S., Jones, J., & Udry, J. R. (2003). The national longitudinal study of adolescent health: Research design. Retrieved from http://www.cpc .unc.edu/projects/addhealth/design Harris, K. M., Halpern, C. T., Smolen, A., & Haberstick, B. C. (2006). The National Longitudinal Study of Adolescent Health (Add Health) twin data. Twin Research and Human Genetics, 9, 988-997. Huang, T. T. K., Harris, K. J., Lee, R. E., Nazir, N., Born, W., & Kaur, H. (2003). Assessing overweight, obesity, diet, and physical activity in college students. Journal of American College Health, 52, 83-86. Hur, Y.-M., Bouchard, T. J., & Eckert, E. (1998). Genetic and environmental influences on self-reported diet: A reared-apart twin study. Physiology and Behavior, 64, 629-636. Jacobson, K., & Rowe, D. C. (1998). Genetic and shared environment influences on adolescent BMI: Interaction with race and sex. Behavior Genetics, 28, 265-278. Jacobson, K., & Rowe, D. C. (1999). Genetics and environmental influences on the relationship between family connectedness, school connectedness, and adolescent depressed mood: Sex differences. Developmental Psychology, 35, 926-939. Keskitalo, K., Silventoinen, K., Tuorila, H., Perola, M., Pietiläinen, K. H., Rissanen, A., & Kaprio, J. (2008). Genetic and environmental contributions to food use patterns of young adult twins. Physiology and Behavior, 93, 235-242. Klump, K. L., McGue, M., & Iacono, W. G. (2000). Age differences in genetic and environmental influences on eating attitudes and

Beaver et al. behaviors in preadolescent and adolescent female twins. Journal of Abnormal Psychology, 109, 239-251. Neumark-Sztainer, D., Wall, M., Perry, C., & Story, M. (2003). Correlates of fruit and vegetable intake among adolescents: Findings from Project EAT. Preventive Medicine, 37, 198-208. Nicklaus, T. A., Baranowski, T., Cullen, K. W., & Berenson, G. (2001). Eating patterns, dietary quality, and obesity. Journal of the American College of Nutrition, 20, 599-608. Niemeier, H. M., Raynor, H. A., Lloyd-Richardson, E. E., Rodgers, M. L., & Wing, R. R. (2006). Fast food consumption and breakfast skipping: Predictors of weight gain from adolescence to adulthood in a nationally representative sample. Journal of Adolescent Health, 39, 842-849. Ogden, C. L., Carroll, M. D., Curtin, L. R., Lamb, M. M., & Flegal, K. M. (2010). Prevalence of high body mass index in US children and adolescents. Journal of the American Medical Association, 303, 242-249. Rodgers, J. L., & Kohler, H.-P. (2005). Reformulating and simplifying the DF analysis model. Behavior Genetics, 35, 211-217. Rodgers, J. L., Rowe, D. C., & Li, C. (1994). Beyond nature versus nurture: DF analysis of nonshared influences on problem behaviors. Developmental Psychology, 30, 374-384. Siega-Riz, A. M., Carson, T., & Popkin, B. (1998). Three squares or mostly snacks—What do teens really eat? Journal of Adolescent Health, 22, 29-36. Steinle, N. I., Hsueh, W.-C., Snitker, S., Pollin, T. I., Sakul, H., St. Jean, P. L., . . . Shuldiner, A. R. (2002). Eating behavior in the Old Order Amish: Heritability analysis and a genome-

151 wide linkage analysis. American Journal of Clinical Nutrition, 75, 1098-1106. Stewart, S. D., & Menning, C. L. (2009). Family structure, nonresident father involvement, and adolescent eating patterns. Journal of Adolescent Health, 45, 193-201. Tepper, B. J. (1998). 6-n-propylthiouracil: A genetic marker for taste, with implications for food preference and dietary habits. American Journal of Human Genetics, 63, 1271-1276. Thompson, J. K., Coovert, M. D., Richards, K. J., Johnson, S., & Cattarin, J. (1995). Development of body image, eating disturbance, and general psychological functioning in female adolescents: Covariance structure modeling and longitudinal investigations. International Journal of Eating Disorders, 18, 221-236. Turkheimer, E. (2000). Three laws of behavioral genetics and what they mean. Current Directions in Psychological Science, 9, 160-164. Udry, J. R. (2003). The national longitudinal study of adolescent health (Add Health), waves I and II, 1994-1996; wave III, 20012002 [machine-readable data file and documentation]. Chapel Hill, NC: Carolina Population Center, University of North Carolina at Chapel Hill. U.S. Surgeon General. (2001). Overweight and obesity: Health consequences. Rockville, MD: Author. van den Bree, M. B. M., Eaves, L. J., & Dwyer, J. T. (1999). Genetic and environmental influences on eating patterns of twins aged ≥50 y. American Journal of Clinical Nutrition, 70, 456-465. Videon, T. M., & Manning, C. K. (2003). Influences on adolescent eating patterns: The importance of family meals. Journal of Adolescent Health, 32, 365-373.