QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gibson, D. Search for Related Content PubMed PubMed Citation Articles by Gibson, D.

---

Community and International Nutrition

Long-Term Food Stamp Program Participation is Differentially Related to Overweight in Young Girls and Boys^1,2

School of Public Affairs, Baruch College, City University of New York, New York, NY 10010

³To whom correspondence should be addressed. E-mail: diane_gibson{at}baruch.cuny.edu.

	ABSTRACT

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
This paper examines the relation between long-term Food Stamp Program (FSP) participation and overweight in children using data on children from the National Longitudinal Survey of Youth 1979 Child Sample. A child was categorized as overweight if his or her BMI was the 95th percentile of sex- and age-specific BMI. The data were arranged as a panel with multiple observations per child, and the preferred models of overweight included long-term FSP participation, additional demographic, socioeconomic, and environmental characteristics, and child fixed effects. Child fixed effects were used to take into account unobserved differences across children that did not vary over time. The models were estimated separately for younger (5–11 y old) and older (12–18 y old) children. In Ordinary Least Squares models, long-term FSP participation was positively and significantly related to overweight in young girls (P = 0.048), and negatively and significantly related to overweight in young boys (P = 0.100). Compared with girls and boys whose families did not participate in the FSP during the previous 5 y, FSP participation during all of the previous 5 y was associated with a 42.8% increase for young girls and a 28.8% decrease for young boys in the predicted probability of overweight. Long-term FSP participation was not significantly related to overweight in older children. Although these models did not control for food insecurity, the potential role of food insecurity in FSP participation was considered in the interpretation of the relation between FSP participation and child weight.

KEY WORDS: • overweight • food stamps • children

This paper considers the relation between long-term participation in the Food Stamp Program (FSP)⁴ and the weight of children. This question is of interest given that 8.82 million children 18 y old participated in the FSP during 2001 (51% of total participants), and previous research found a positive relation between FSP participation and weight in women (1–3). The adverse health consequences of overweight in childhood include an increased likelihood of hypertension, hyperlipidemia, glucose intolerance, and obesity in adulthood [previous research on this topic is summarized in Dietz (4)]. Obesity in adulthood has been found to increase the likelihood of heart disease, high blood pressure, cancer, and diabetes (5–7).

Child overweight may be influenced by long-term FSP participation if long-term FSP participation affects food insecurity or health behaviors related to child weight such as the quantity, quality, or timing of food consumption (8–10). Alternatively, long-term FSP participation may not play a causal role in child weight but may instead be related to child weight as a result of the influence of household food insecurity on both child weight and long-term FSP participation (11–14).

Previous research that has considered the relation between FSP participation and the weight of children used cross-sectional data. Bhattacharya and Currie (15) examined the relation between FSP participation and obesity (BMI > 27.3 kg/m² for girls and > 27.8 kg/m² for boys) in 12- to 16-y-old adolescents using data from the Third National Health and Nutritional Examination Survey (NHANES III). They found that current family FSP participation was not significantly related to adolescent obesity in linear probability models that also controlled for current family income, parent’s education, and other individual and family characteristics that were expected to influence youth obesity. The models did not include variables that measured an adolescent’s history of poverty or social program participation. Gibson (16) examined the relation between FSP participation and the obesity (BMI 30 kg/m²) of 12- to 16-y-old adolescents using y 1 of data from the National Longitudinal Survey of Youth 1997 (NLSY97). She found that current family FSP participation was not significantly related to adolescent obesity in logistic regression models that controlled for long-term poverty and social program participation in addition to the variables used by Bhattacharya and Currie. These models did not include long-term FSP participation because this information was not available in y 1 of the NLSY97. Jones et al. (17) considered whether the relation between food program participation and the risk of overweight (BMI 85th percentile) differed with a child’s food security status. The food programs of interest were the FSP, School Lunch Program, and School Breakfast Program. Using data on low-income children aged 5–12 y from the 1997 Panel Study of Income Dynamics Child Development Supplement (PSID CDS), they estimated logistic regression models of risk of overweight that included two sets of food program participation variables, one for food-insecure children and another for food-secure children. Their models also included gender, race or ethnicity, age, total family income, and the education level of the household head. The models did not include a separate variable for food security status or variables that measured a child’s history of poverty or social program participation. They found that participation in all types of food programs was negatively and significantly related to risk of overweight in food-insecure girls, and FSP participation without participation in other food programs was negatively and significantly related to risk of overweight in food-secure girls. They also found that participation in all types of food programs was not significantly related to overweight in both food-insecure and food-secure boys.

The empirical analyses in this paper attempted to build on the previous research by using panel data and specifically considering the relation between long-term FSP participation and child overweight. The panel structure of the data allowed family fixed effects or child fixed effects to be included in the empirical models of child overweight. Family fixed effects took into account unobserved family characteristics that affected child weight and did not change over time. Child fixed effects took into account unobserved child characteristics that affected child weight and did not change over time.

	SUBJECTS AND METHODS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
    Sample. Data from the National Longitudinal Survey of Youth 1979 (NLSY79) on mothers and their children were used to examine the relation between FSP participation and the weight of children. The initial wave of the NLSY79 was administered in 1979 to 6283 females and 6403 males aged 14–22 y in 8770 households. Follow-up interviews were conducted annually until 1994 and biennially thereafter. The children born to the women in the NLSY79 sample were surveyed in 1986. Surveys of these children have been conducted biennially since then, and children born after 1986 were added to the sample of children in subsequent survey years. All of the children in a family are surveyed. With these data, it is possible to link together detailed longitudinal information on mothers and their children.

The data were arranged as a panel so that there were multiple observations per child and the unit of analysis was a child-year. The sample used in the empirical analyses included observations on children from the 1986 through the 2000 waves of the survey, although data from earlier years of the NLSY79 were used to create the variables that measured long-term family resources. An observation on a child was included in the sample every survey year in which the child was between the ages of 5 and 18 y and for which there was information on the child’s current weight, height, FSP participation status, and family income. The final sample contained 12,801 observations on 3831 girls from 2656 families and 13,303 observations on 4012 boys from 2707 families.

    Conceptual model. The conceptual model of child weight (Fig. 1) that guided the empirical analyses is similar to that used in previous research on child weight (14,18) and on the relation between FSP participation and adult weight (2–3). The model assumes that a child’s weight is influenced by current and past child, family, and environmental characteristics that may affect the child’s weight directly or indirectly through their influence on health behaviors related to weight or food insecurity. Examples of these characteristics are a child’s Food Stamp (FS) family resources, non-FS family resources, age, race or ethnicity, gender, pregnancy status, birth order, birthweight, whether breast-fed, mother’s age at birth of child, family composition, parental education, parental employment, region and urbanicity of residence, and underlying genetic factors. Some of these personal characteristics may vary over time and some may not.

View larger version (20K): [in this window] [in a new window]   FIGURE 1 Conceptual framework for the relation between current child weight and current and past child, family, and environmental characteristics.

The conceptual model presents the direction of influence as flowing from FSP participation to food insecurity. However, as Gibson (2) and Frongillo (11) point out with respect to a similar conceptual model of the weight of adults, FSP participation could also be influenced by food insecurity. Both of these possibilities are illustrated by previous research that found that household FSP participation was associated with a reduction in household food insecurity (12,15), but households that participated in the FSP were still more likely to be food insecure than nonparticipating households (12,13). The empirical analyses in this paper take the approach depicted by the conceptual model and estimate the relation between child weight and child, family, and environmental characteristics without controlling for food insecurity. However, the potential role of food insecurity in FSP participation will be considered in the interpretation of the relation between FSP participation and child weight.

    Dependent variables. Child weight and height in a survey year were combined to calculate a child’s BMI in that survey year. A child was categorized as overweight if his or her BMI was the 95th percentile of sex- and age-in-months–specific BMI from the revised BMI percentiles issued by the Centers for Disease Control in 2000 (24). At risk of overweight was defined as a BMI the 85th percentile and < the 95th percentile of sex- and age-in-months–specific BMI.

    Independent variables. The empirical models of overweight included long-term FSP participation. This variable measures the percentage of time a child’s family participated in the FSP over the 5 calendar years preceding the survey year (hereafter referred to as the "previous 5 y"). This variable was calculated using the available years of data for the 3096 observations without data on family FSP participation in each of the previous 5 y.

The models also included a measure of long-term family resources other than long-term FSP participation. In every survey year, information was collected on total family income in the calendar year preceding the survey year. Total family income included income from possible sources such as wages, income from a business, and income from social programs such as Aid to Families with Dependent Children (AFDC), Temporary Aid for Needy Families (TANF), FS, and other public assistance (25).

FS eligibility income was defined as total family income minus FS benefits. The FS eligibility income-to-needs ratio for a survey year was calculated by dividing the FS eligibility income of the child’s family by the poverty threshold appropriate for the family’s size (26). Long-term family resources other than long-term FSP participation were approximated with the mean of the child’s family FS eligibility income-to-needs ratios over the previous 5 y. This variable was calculated using the available years of data for children without data on FS eligibility income in each of the previous 5 y.

In the models of child overweight without family or child fixed effects, other control variables included a child’s age, race or ethnicity, pregnancy status, birth order, birthweight, whether breast-fed, mother’s age at birth of the child, number of children in family, mother’s highest grade completed, mother’s AFQT (Armed Forces Qualifying Test) score, mother’s marital status, mother’s employment status, region and urbanicity of residence, and indicators for mother-reported or self-reported child weight or height. Child weight and height were mother-reported, self-reported, or measured with a scale and tape measure. The models also included indicators for survey year.

The empirical analyses used family fixed effects or child fixed effects in the models of child overweight to deal with the possibility that unmeasured characteristics were related to both long-term FSP participation and child weight. Family fixed effects take into account unobserved family characteristics that affect child weight and do not vary over time. Because girls and boys were analyzed separately, family fixed effects could vary by gender for children from the same family. A disadvantage of family fixed effects is that information on children who were in the sample in only one survey year and who did not have any siblings of the same gender could not be used to estimate the coefficients on the other variables in the model, in effect shrinking the sample size (there were 393 girls and 410 boys in the sample for whom this was the case). This is a consequence of the fact that models of overweight with family fixed effects use within-family variation in personal characteristics to explain within-family variation in child weight. Family characteristics that do not vary over time can be omitted from these models (27). The variables that were excluded were the child’s race or ethnicity and mother’s AFQT score. Race or ethnicity did not vary between children in a family because maternal race or ethnicity was used to identify a child’s race or ethnicity.

Family fixed effects may result in biased estimates if children with multiple observations in the sample or with a sibling of the same gender in the sample differ systematically from children who were in the sample in only one survey year and who did not have any siblings of the same gender in the sample. This bias may not be large for girls because the unweighted prevalence of overweight between the two groups did not differ significantly for girl-year observations aged 5–11 y (15.6 vs. 16.4%, Pearson ² = 0.22, P = 0.64) or girl-year observations aged 12–18 y (15.6 vs. 16.4%, Pearson ² = 0.19, P = 0.66). However, this bias may be more of a concern for boys because the unweighted prevalence of overweight between the two groups differed significantly for boy-year observations aged 5–11 y (17.0 vs. 20.1%, Pearson ² = 3.54, P = 0.06) and boy-year observations aged 12–18 y (16.9 vs. 23.7%, Pearson ² = 14.0, P < 0.01).

Child fixed effects control for unobserved child characteristics that affect child weight and do not vary over time. Child fixed effects may be an improvement over family fixed effects given that previous research has found that parenting choices often differ for children in the same family [see for example (9,10)]. With child fixed effects, information about children who were in the sample in only one survey year could not be used to estimate the coefficients on the other variables in the model, again shrinking the effective sample size (there were 708 girls and 784 boys in the sample for whom this was the case). The time-invariant controls excluded from the empirical models of overweight with child fixed effects were the child’s race or ethnicity, birth order, birthweight, whether breast-fed, mother’s age at birth of the child, and mother’s AFQT score. Age was also excluded from the child fixed effects models because it could not be identified separately from the time indicators (2,28).

Child fixed effects may result in biased estimates if children with multiple observations in the sample differ systematically from children without multiple observations in the sample. This bias may not be large for girls and younger boys because the unweighted prevalence of overweight between the two groups did not differ significantly for girl-year observations aged 5–11 y (15.5 vs. 16.7%, Pearson ² = 0.95, P = 0.33), girl-year observations aged 12–18 y (15.5 vs. 17.6%, Pearson ² = 2.39, P = 0.12) or boy-year observations aged 5–11 y (17.1 vs. 17.4%, Pearson ² = 0.07, P = 0.80). However, the unweighted prevalence of overweight between the two groups differed significantly for boy-year observations aged 12–18 y (16.8 vs. 22.5%, Pearson ² = 16.02, P < 0.01).

    Statistical analysis. As in Alaimo et al. (14) and Jones et al. (17), girls and boys were analyzed separately. Differences in the unweighted prevalence of overweight and at-risk of overweight among FSP participation categories were examined with Pearson ² tests using a significance level of P < 0.10. Differences in the mean weight among FSP participation categories were examined with two-sample t tests with the assumption of unequal variance using a significance level of P < 0.10.

Data on children in three age ranges were used in Ordinary Least Squares (OLS) and logistic regression models of overweight. The full sample included children between the ages of 5 and 18 y, a younger sample included children between the ages of 5 and 11 y, and an older sample included children between the ages of 12 and 18 y. Three models were estimated for each age range: model (1) did not include fixed effects, model (2) included family fixed effects, and model (3) included child fixed effects.

Following the NSLY79 study recommendations, the regression models used statistical controls to specify group membership rather than using sample weights. To that end, the models were estimated separately by gender and included indicator variables for urbanicity, region of residence, and race or ethnicity (only in the models without fixed effects) (25). Huber-White standard errors were calculated in the OLS models without fixed effects and with family fixed effects, with clustering on the child in both cases. This method assumed that observations within a cluster were not independent. The standard errors were not adjusted further to account for design effects; however, the design effects in longitudinal data are expected to decline over time due to the mobility of respondents (25). Statistical analyses were completed with Stata, Version 7 (Stata Corporation).

	RESULTS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
In the sample, 15.6% of girl-year observations and 17.1% of boy-year observations were overweight, and 22.7% of girl-year observations and 22.7% of boy-year observations were current FSP participants. The prevalence of overweight by FSP participation (Table 1 for girls and Table 2 for boys) was significantly different for girl-year observations (Pearson ² = 15.65, P < 0.01) but not boy-year observations (Pearson ² = 0.42, P = 0.52).

The relation between long-term FSP participation and overweight in girls varied across models (Table 3). In model (1) (without fixed effects), long-term FSP participation was positively and significantly related to overweight in girls in the full sample and the older sample. In model (2) (with family fixed effects) and model (3) (with child fixed effects), long-term FSP participation was positively and significantly related to overweight in the younger sample but was not significant for the other age ranges. The change in the magnitude and significance of the coefficients on long-term FSP participation with the introduction of fixed effects suggests that there were unobserved family characteristics related to both long-term FSP participation and overweight in girls. Furthermore, the change in the coefficients on long-term FSP participation with the inclusion of child fixed effects rather than family fixed effects indicates that there were unobserved child-specific characteristics related to both long-term FSP participation and child weight.

The relation between long-term FSP participation and overweight in boys also varied across models (Table 4). In model (1) (without fixed effects), long-term FSP participation was not significantly related to overweight in boys. In model (2) (with family fixed effects), long-term FSP participation was negatively and significantly related to overweight in younger boys, but was not significant for the other age ranges. In model (3) (with child fixed effects), long-term FSP participation was also negatively and significantly related to overweight in younger boys, although the coefficient was barely significant at the 10% level. As with girls, the change in the coefficient estimates with the introduction of fixed effects suggests that there were unobserved family and child characteristics related to both long-term FSP participation and overweight in boys.

To test whether the relation between long-term FSP participation and child weight was sensitive to the definition of "long-term," models (1)–(3) were estimated using the percentage of a child’s lifetime spent on the FSP and the child’s lifetime mean FS eligibility income-to-needs ratio instead of the variables calculated using the previous 5 y. The significance, magnitude, and direction of these estimates were very similar to those in the models that used the variables for the previous 5 y, although the P-values were higher for younger girls and lower for younger boys (results not shown).

	DISCUSSION

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
The empirical models of child overweight controlled for a large set of child, family, and environmental characteristics. However, the results presented so far do not establish whether the relation between long-term FSP participation and child overweight was due specifically to long-term FSP participation or due instead to long-term FSP participation as a proxy for long-term poverty.

To explore this possibility, long-term FSP participation was removed from models (2) and (3) and replaced with long-term eligibility for the FSP. Following previous research, a child’s family was assumed to be eligible for the FSP in survey years in which the family’s FS eligibility income corresponded to a FS eligibility income-to-needs ratio that was <1.3 (15,29,30). Long-term eligibility for the FSP was defined as the percentage of time a child’s family was eligible for the FSP over the previous 5 y. The inclusion of both the long-term mean FS eligibility income-to-needs ratio and long-term eligibility accounts for the mean and the variance of long-term non-FS family resources. Neither the long-term mean FS eligibility income-to-needs ratio nor long-term eligibility were significantly related to overweight for girls or boys (results not shown). This suggests that the significance of the relation between child overweight and long-term FSP participation was not due to long-term poverty.

An additional question is whether the significance of the relation between long-term FSP participation and child overweight was due to FSP participation in particular, other types of social program participation, or social program participation in general. Information on long-term AFDC or TANF participation is available in the NLSY79, and long-term AFDC or TANF participation was defined as the percentage of time a child’s family participated in AFDC or TANF over the previous 5 y. When long-term FSP participation was removed from models (2) and (3) and replaced with long-term AFDC or TANF participation, this variable was not significantly related to overweight in either girls or boys (results not shown). These results support the argument that long-term AFDC or TANF participation or social program participation in general were not responsible for the relation between long-term FSP participation and child overweight. Given the limited information available in the NLSY79 on benefits from the Special Supplemental Nutrition Program for Women, Infants and Children (WIC), it was not possible to consider the relation between long-term WIC participation and child weight.

Even with family or child fixed effects, the estimates of the relation between long-term FSP participation and child overweight may be biased as a result of reverse causality from child overweight to FSP participation or selection bias. Reverse causality is a possibility if a child’s weight status influenced a family’s decision to participate in the FSP. Selection bias remains a potential problem if there were unobserved child, family, or environmental characteristics of children that varied over time and were related to both child overweight and long-term FSP participation.

Research by Gleason et al. (31) suggests possible sources of selection bias. They found that events such as recent changes in income, household composition, or the receipt of other public assistance were more likely among households that started participating in the FSP than among households that remained nonparticipants. To take into account the possibility that these types of events also influenced a child’s overweight status, controls for the change in a family’s income eligibility for the FSP and maternal marital status in the previous calendar year were added to models (2) and (3). Long-term FSP participation remained positively and significantly related to overweight in younger girls and negatively and significantly related to overweight in younger boys in these models, with magnitudes and significance levels very similar to the results of the models excluding the change variables (results not shown).

As mentioned earlier, another potential source of bias arises if food insecurity or other intermediate variables in the conceptual model influence a family’s participation in the FSP as well as child overweight. Unfortunately, the NLSY79 does not contain information on a family’s food security status.

Four studies have looked at the relation between child weight and food insufficiency or food insecurity. Using data from the 1994–1996 Continuing Survey of Food Intakes by Individuals, Casey et al. (32) found that the prevalence of overweight (BMI 85th percentile) in children aged 0–17 y old did not differ between children in low-income food-insufficient families and children in low-income food-sufficient families. Using data from the NHANES III, Alaimo et al. (14) estimated logistic regression models of overweight (BMI 85th percentile) that included a large set of controls for child and family background characteristics. They found that food insufficiency was negatively and significantly related to overweight for younger (aged 2–7 y) girls, positively and significantly related to overweight for older (aged 8–16 y) non-Hispanic Caucasian girls, and not significantly related to overweight for other older girls of boys of any age.

With data from the 1997 PSID CDS, Jones et al. (17) found that the prevalence of risk of overweight (BMI 85th percentile) in children aged 5–12 y was significantly lower among low-income food-insecure boys and girls than low-income food-secure boys and girls. Also using data from the NHANES III, Bhattacharya et al. (33) combined girls and boys and found that family food insecurity was not significantly related to obesity (BMI 95th percentile) for all age groups (aged 0–5 y, 6–11 y, and 12–17 y) in logistic regression models that included a somewhat more limited set of controls for child and family background characteristics than those of Alaimo et al. (14).

As mentioned earlier, Jones et al. (17) found that FSP participation was negatively and significantly related to risk of overweight in both food-insecure and food-secure low-income girls, but was not significantly related to risk of overweight in low-income boys. Their results suggest that they would have found a negative and significant relation between FSP participation and risk of overweight in girls and no relation in boys if their models had included food program participation variables without food security status. This does not match the relations found in this paper. The models in this paper contained a more detailed set of variables for child, family, and environmental characteristics than those used by Jones et al. (17); yet there were still substantial changes in the coefficients on long-term FSP participation and other variables with the addition of family or child fixed effects. This suggests that there were unobserved characteristics related to food program participation and child overweight, and that the estimates of Jones et al. (17) are therefore likely to suffer from omitted variable bias.

Given the findings of the previous research, there is a possibility that the coefficients on long-term FSP participation were biased as a result of the exclusion of food insecurity from the empirical models of child overweight; however, given the inconsistent findings across previous studies, the direction of the bias is unclear. Additional research that examines the relation between food insecurity and weight separately for girls and boys is required to determine the strength of the relation between food insecurity and child weight, and whether the omission of food insecurity is likely to bias the estimates of the relation between child overweight and long-term FSP participation.

An important issue is whether there is a plausible mechanism to account for the relation between long-term FSP participation and overweight, beyond the role of long-term FSP participation as a potential proxy for food insecurity. The mechanism will have to explain why the relation between long-term FSP participation and overweight was positive in young girls, negative in young boys and not significant in older children. Possible mechanisms could be based on differences by age and gender in food choices, energy intake, and/or energy expenditures as a result of FSP participation. Previous research has considered the relation between FSP participation and nutrient intakes, but has not considered the relation between FSP participation and activity levels.

Estimates of the relation between FSP participation and the intake of nutrients by children have varied considerably across studies [previous research on this topic is summarized in Currie (34)]. The previous research did not examine girls and boys separately; thus, there is no direct information available on whether girls and boys make different food or energy intake choices in response to the FSP.

Research on the effect of the family food environment on child eating and child weight suggests that girls and boys respond differently to similarities in the family food environment and also encounter differences in the family food environment. For example, role models such as parents, siblings, and peers heavily influence the food choices of children, but there is substantial evidence that mothers are particularly strong role models for daughters [previous research on this topic is summarized in Birch and Davison (9)]. For another example, previous research suggests that girls and boys respond somewhat differently to restrictive child-feeding practices that limit access to certain types of food. A number of studies by Birch and colleagues found that restrictive child-feeding practices were related to a compromised ability to regulate energy intake for both girls and boys (35–37) but more so for girls (37). Additionally, a positive relation between restrictive child-feeding practices and overweight was found for girls (37–39) but not for boys (37).

This research suggests that changes in the family environment could have a different effect on weight for girls and boys. However, previous research does not offer much insight into how FSP participation might change the family food environment. For example, it is not clear how parental eating patterns change with FSP participation. Almost all of the available empirical evidence suggests that participation in the FSP increased family food expenditures (40–44), yet the estimates of the relation between participation in the FSP and the intake of nutrients by adults vary widely across studies [summarized in Currie (34) and Fraker (41)]. The relation between FSP participation and child feeding practices has not been examined.

A possible yet somewhat unlikely mechanism is that maternal obesity may mediate the relation between long-term FSP participation and overweight in younger girls. Previous research considered whether FSP participation could lead to overweight and obesity in adults (2,3). Current and long-term FSP participation were positively associated with obesity in low-income women in models that also included extensive controls for demographic, socioeconomic, and environmental characteristics as well as individual fixed effects (2). Additionally, current FSP participation was positively related to overweight in women in models that also controlled for current food insecurity as well as other demographic, socioeconomic, and environmental characteristics (3). Maternal obesity in turn was positively related to restrictive child-feeding practices (38,39) and maternal disinhibited eating (45), both of which were associated with an increase in overweight in girls but not boys (37,45). Further research is required to determine whether changes in maternal obesity result in changes in child-feeding practices.

In the empirical analyses in this paper, long-term FSP participation was not significantly related to overweight in older children. The research that connects the family food environment to the weight of girls used samples composed primarily of younger children (6 y old). These findings may hold for older girls, yet additional exposure to the FSP may not increase the likelihood of overweight if these girls were already overweight as a result of exposure to the FSP at younger ages. Alternatively, these findings may not hold for older children because older children have more control over their food and activity choices and may therefore not be affected in the same way as younger children by family participation in the FSP.

Gibson (2) argued that the FSP offered a way not only to reach low-income women with a high likelihood of obesity, but also to reach these women at a time when they were more likely to be obese. The empirical analyses presented here support a similar conclusion with respect to overweight in young girls. The results do not establish that long-term FSP participation caused an increase in the weight of girls. However, targeting healthy eating and weight reduction policies through the FSP offers a way not only to reach families that may have multiple members at risk of overweight, but also to reach these families at a time when family members are more likely to be overweight.

	ACKNOWLEDGMENTS

 
Marcellus Andrews, Nancy Aries, Gerald Cubbin and Sanders Korenman provided extremely helpful comments on earlier versions of this paper. Thanks are also due to Petros Petrou for excellent research assistance.

	FOOTNOTES

 
¹ Presented in preliminary form at the Department of Health Policy at Mt. Sinai Hospital, September 16, 2003 [Gibson, D. Long-term food stamp program participation is related to overweight in young girls and boys].

² Supported by a grant from the PSC-CUNY Research Awards Program.

⁴ Abbreviations used: AFDC, Aid to Families with Dependent Children; AFQT, Armed Forces Qualifying Test; FS, Food Stamps; FSP, Food Stamp Program; NHANES III, Third National Health and Nutrition Examination Survey; NLSY79, National Longitudinal Survey of Youth 1979; NLSY97, National Longitudinal Survey of Youth 1997; OLS, Ordinary Least Squares; PSID CDS, Panel Study of Income Dynamics Child Development Supplement; TANF, Temporary Aid for Needy Families; WIC, Special Supplemental Nutrition Program for Women, Infants and Children.

Manuscript received 18 August 2003. Initial review completed 21 October 2003. Revision accepted 12 November 2003.

	LITERATURE CITED

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 

1. U.S. Department of Agriculture, Food and Nutrition Service (2001) Characteristics of Food Stamp Households: Fiscal Year (Advance Report). 2001 http://www.fns.usda.gov/oane/menu/Published/fsp/FILES/Participation/2001advrtp.pdf. (accessed August 28, 2002).

2. Gibson, D. (2003) Food stamp program participation is positively related to obesity in low income women. J. Nutr. 133:2225-2231.[Abstract/Free Full Text]

3. Townsend, M., Peerson, J., Love, B., Achterberg, C. & Murphy, S. (2001) Food insecurity is positively related to overweight in women. J. Nutr. 131:1738-1745.[Abstract/Free Full Text]

4. Dietz, W. (1998) Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics 101:518-525.[Abstract/Free Full Text]

5. Stevens, J., Jianwen, C., Pamuk, E., Williamson, D., Thun, M. & Wood, J. (1998) The effects of age on the association between body-mass index and mortality. N. Engl. J. Med. 338:1-7.[Abstract/Free Full Text]

6. Woolf, A. & Colditz, G. (1998) Current estimates of the economic cost of obesity in the United States. Obes. Res. 6:97-106.[Medline]

7. Allison, D., Fontaine, K., Manson, J., Stevens, J. & VanItallie, T. (1999) Annual deaths attributable to obesity in the United States. J. Am. Med. Assoc. 282:1530-1538.[Abstract/Free Full Text]

8. Sobal, J. & Stunkard, A. (1989) Socioeconomic status and obesity: a review of the literature. Psychol. Bull. 105:260-275.[Medline]

9. Birch, L. & Davison, K. (2001) Family environmental factors influencing the developing behavioral controls of food intake and childhood overweight. Pediatr. Clin. N. Am. 48:893-907.[Medline]

10. Birch, L. & Fisher, J. (1998) Development of eating behaviors among children and adolescents. Pediatrics 101:539-549.[Abstract/Free Full Text]

11. Frongillo, E. (2003) Understanding obesity and program participation in the context of poverty and food insecurity. J. Nutr. 133:2117-2118.[Free Full Text]

12. Borjas, G. (2002) Food Insecurity and Public Assistance. Working Paper. 2002 Kennedy School of Government Harvard University, Boston, MA.

13. Nord, M. (2001) Food Stamp Program participation and food security. Food Rev 24:13-19.

14. Alaimo, K., Olson, C. & Frongillo, E., Jr (2001) Low family income and food insufficiency in relation to overweight in US children. Arch. Pediatr. Adolesc. Med. 155:1161-1167.[Abstract/Free Full Text]

15. Bhattacharya, J. & Currie, J. (2002) Youths at nutrition risk: malnourished or misnourished?. Gruber, J. eds. Risky Behavior among Youths: An Economic Analysis 2002 University of Chicago Press Chicago, IL. .

16. Gibson, D. (2001) Food Stamp Program participation and health: estimates from the NLSY97. Michael, R. eds. Social Awakening: Adolescent Behavior as Adulthood Approaches 2001 Russell Sage Foundation New York, NY. .

17. Jones, S., Jahns, L., Laraia, B. & Haughton, B. (2003) Lower risk of overweight in school-aged food insecure girls who participate in food assistance. Arch. Pediatr. Adolesc. Med. 157:780-784.[Abstract/Free Full Text]

18. Olson, C. (1999) Nutrition and health outcomes associated with food insecurity and hunger. J. Nutr. 129:521S-524S.[Medline]

19. Fogelholm, M., Nuutinen, O., Pasanen, M., Myohanen, E. & Saatela, T. (1999) Parent-child relationship of physical activity patterns and obesity. Int. J. Obes. 23:1262-1268.[Medline]

20. Dowda, M., Ainsworth, B., Addy, C., Saunders, R. & Riner, W. (2001) Environmental influences, physical activity, and weight status in 8- to 16-year-olds. Arch. Pediatr. Adolesc. Med. 155:711-717.[Abstract/Free Full Text]

21. Frongillo, E., Jr, Olson, C., Rauschenbach, B. & Kendall, A. (1997) Nutritional Consequences of Food Insecurity in a Rural New York State County. 1997 Institute for Research on Poverty, University of Wisconsin Madison, WI Discussion Paper no. 1120–97.

22. Sarlio-Lahteenkorva, S. & Lahelma, E. (2001) Food insecurity is associated with past and present economic disadvantage and body mass index. J. Nutr. 131:2880-2884.[Abstract/Free Full Text]

23. Dietz, W. (1995) Does hunger cause obesity?. Pediatrics 95:766-767.[Abstract/Free Full Text]

24. Kuczmarski, R., Ogden, C. & Grummer-Strawn, L. (2000) CDC Growth Charts: United States. 2000 National Center for Health Statistics Hyattsville, MD Advance Data Number 314.

25. Center for Human Resource Research (1999) NLSY79 User’s Guide, 1999. 1999 U.S. Government Printing Office Washington, DC.

26. U.S. Department of Health and Human Services. (2002) Summary Historical Figures and Federal Register References for the HHS Poverty Guidelines since 1982. 2002 http://www.aspe.hhs.gov/poverty/figures-fed-reg.htm (accessed December 12, 2002).

27. Greene, W. (1997) Econometric Analysis 3rd ed. 1997 Prentice-Hall Upper Saddle River, NJ .

28. Lakdawalla, D. & Philipson, T. (2001) The Economics of Obesity: A Theoretical and Empirical Examination. Presented at the annual meetings of the American Economics Association 2001 New Orleans, LA January 5-7.

29. Mathematica Policy Research Institute (2000) Dietary Intake and Dietary Attitudes Among Food Stamp Participants and Other Low-Income Individuals. 2000 U.S. Department of Agriculture, Food and Nutrition Service Washington, DC Report 8370–025.

30. Blank, R. & Ruggles, P. (1996) When do women use aid to families with children and food stamps? The dynamics of eligibility versus participation. J. Hum. Resour. 31:57-89.

31. Gleason, P., Schochet, P. & Moffitt, R. (1998) The Dynamics of Food Stamp Program Participation in the Early 1990s 1998 U.S. Department of Agriculture, Food and Nutrition Service Washington, DC .

32. Casey, P., Szeto, K., Lensing, S., Bogle, M. & Weber, J. (2001) Children in food-insufficient, low-income families: prevalence, health and nutrition status. Arch. Pediatr. Adolesc. Med. 155:508-514.[Abstract/Free Full Text]

33. Bhattacharya, J., Currie, J. & Haider, S. (2002) Food Insecurity or Poverty? Measuring Need-Related Dietary Adequacy 2002 National Bureau of Economic Research Cambridge, MA NBER working paper 9003.

34. Currie, J. (2000) U.S. Food and Nutrition Programs. Working Paper. 2000 Department of Economics, University of California at Los Angeles Los Angeles, CA.

35. Fisher, J. & Birch, L. (1999) Restricting access to foods and children’s eating. Appetite 32:405-419.[Medline]

36. Fisher, J. & Birch, L. (1999) Restricting access to a palatable food affects children’s behavioral response, food selection and intake. Am. J. Clin. Nutr. 69:1264-1272.[Abstract/Free Full Text]

37. Johnson, S. & Birch, L. (1994) Parent’s and children’s adiposity and eating style. Pediatrics 94:653-661.[Abstract/Free Full Text]

38. Birch, L. & Fisher, J. (2000) Mothers’ child-feeding practices influence daughters’ eating and weight. Am. J. Clin. Nutr. 71:1054-1061.[Abstract/Free Full Text]

39. Francis, L., Hofer, S. & Birch, L. (2001) Predictors of maternal child-feeding style: maternal and child characteristics. Appetite 37:231-243.[Medline]

40. Devaney, B. & Fraker, T. (1989) The effect of food stamps on food expenditures: an assessment of findings from the nationwide food consumption survey. Am. J. Agric. Econ. 71:99-104.

41. Fraker, T. (1990) The Effects of Food Stamps on Food Consumption: A Review of the Literature 1990 U.S. Department of Agriculture,Food and Nutrition Service Alexandria, VA.

42. Levedahl, J. (1991) The Effect of Food Stamps and Income on Household Food Expenditures 1991 United States Department of Agriculture, Economic Research Service Washington, DC.

43. Levedahl, J. (1995) A theoretical and empirical evaluation of the functional forms used to estimate the food expenditure equation of food stamp recipients. Am. J. Agric. Econ. 77:960-968.

44. Fraker, T., Martini, A. & Ohls, J. (1995) The effect of food stamp cashout on food expenditures: an assessment of the findings from four demonstrations. J. Hum. Resour. 30:633-649.

45. Cutting, T., Fisher, J., Grimm-Thomas, K. & Birch, L. (1999) Like mother, like daughter: familial patterns of overweight are mediated by mothers’ dietary disinhibition. Am. J. Clin. Nutr. 69:608-613.[Abstract/Free Full Text]

This article has been cited by other articles:

				K. Kim and E. A. Frongillo Participation in Food Assistance Programs Modifies the Relation of Food Insecurity with Weight and Depression in Elders J. Nutr., April 1, 2007; 137(4): 1005 - 1010. [Abstract] [Full Text] [PDF]

				D. Gibson Long-Term Food Stamp Program Participation Is Positively Related to Simultaneous Overweight in Young Daughters and Obesity in Mothers J. Nutr., April 1, 2006; 136(4): 1081 - 1085. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gibson, D. Search for Related Content PubMed PubMed Citation Articles by Gibson, D.