Examining Relationships Among Obesity, Food Insecurity, Stress, and Emotional Eating in Low-Income Caregivers of Head Start Children

Year: 2007

Research Center: Southern Rural Development Center, Mississippi State University

Investigator: Lofton, Kristi, and Carol Connell

Institution: University of Southern Mississippi

Project Contact:
Kristi Lofton
The University of Southern Mississippi
118 College Drive #5172
Hattiesburg, MS 39406
Phone: 601-266-5377


Amidst a food environment that is perceived as highly accessible and affordable, research indicates that food insecurity and obesity occur paradoxically. Socioeconomic similarities exist between food insecurity and obesity, such as the highest prevalence occurring among women and ethnic minorities and a higher prevalence of food insecurity and obesity among people of low socioeconomic status. Food insecurity and obesity have also been reported to occur simultaneously in the same individuals as well as in members within the same household. However, other studies have found no association between food insecurity and obesity when other factors, such as income, education, ethnicity, and so forth, are controlled or when food insecurity is persistent or severe. This result suggests that other factors, in conjunction with food insecurity, may be playing a role in the food-insecurity-obesity paradox.

The food insecurity/obesity relationship may be explained by food insecurity experiences associated with stress and, in turn, the impact of stress on eating behavior. These stress-related food insecurity experiences may prompt individuals to develop food-related behaviors, such as emotional eating, when food is in adequate supply that contribute directly or indirectly to overweight and obesity. It is theoretically plausible that, in situations of low food security when low-cost energy-dense food is available, emotional eating occurs in response to the stress of uncertain resources.

Because the literature suggests that stress or other psychosocial factors influence eating behaviors associated with obesity, and stress has been related to food insecurity, investigating whether obesity, stress, and food insecurity are linked is important. Therefore, the primary objective of this study was to examine relationships among food insecurity, stress, emotional eating, and obesity by using a moderation analysis model. Likewise, the literature indicates that the food-insecurity-obesity relationship exists primarily among women, so two sets of analyses were conducted. The first set examined the research questions among an entire sample of caregivers of Head Start children, whereas the second set of analyses examined the research questions among the females in the sample. A secondary purpose of the study was to assess low-income women’s perceptions of their own weight status compared with measured weight status to inform educational efforts of Federal food assistance programs.

The survey instrument used in this cross-sectional study consisted of previously validated instruments to measure household food security, perceived stress, and three components of emotional eating: anger/frustration, anxiety, and depression. The study surveyed 690 participants with children attending two area Head Start agencies or with children receiving benefits from the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) or food stamps were surveyed and measured the children’s heights and weights. Initial analyses consisted of 634 participants with complete data for the dependent and independent variables, and the final analyses of the women’s data consisted of 627 participants with complete data. Due to the very low number of men in the sample, only the women’s data are reported in this study.

The majority of the participants were African-American (83.7 percent) and living without a partner (62.8 percent). Over 75 percent of the women were overweight or obese. Using the traditional U.S. Department of Agriculture definition of food security based on the number of affirmative responses to the food security items, only 64 percent of the respondents were food secure. Another 26 percent were classified as low food secure and 9.6 percent were classified as very low food secure. These rates were twice as high as national prevalence rates for households with children under the age of 6. Only 39 percent of this population participated in the Food Stamp Program.

Results revealed no statistically significant differences in body mass index (BMI, a measure of weight status) based on the food security status of the household in this group of low-income women. This result is likely due to the high rates of overweight/obesity across all levels of food security attenuating any differences that may have existed. The study revealed that perceived stress increased from food secure to marginally food secure to low food secure. However, perceived stress scores of the women in the low food secure and very low food secure groups showed no difference. Mean scores for the emotional eating subscales at various levels of food security did not increase in a linear fashion. The emotional eating scores were not different between the fully food secure and marginally food secure groups. The very low food secure group was not different from the marginally food secure group on these three subscales (anger/frustration, anxiety, depression). The low food secure group was different from the fully food secure group on all emotional eating subscales, but this finding was not a consistent when comparing the two groups with the other two food security groups and is probably due to large standard deviations around mean scores among these groups.

Likewise, regression analysis did not reveal an association between BMI and food security status when marginally food secure, low food secure, and very low food secure (represented by dummy variables) were used as predictors of BMI. Furthermore, the moderation analysis revealed no association among stress or emotional eating subscales or their interaction with each other and BMI. Due to the high rates of overweight/obesity already present in these adult women, further research of a longitudinal design is needed to clarify whether the development of eating habits in response to food security could contribute to development of overweight and obesity in adulthood.

Finally, over 75 percent of the women who were obese misclassified themselves as overweight, and 45 percent of overweight women misclassified themselves as normal weight. It is possible that the term “obese” has such negative connotations that few women would classify themselves as such. Also, cultural perceptions of acceptable body weights among African-American women may vary from those of the majority population resulting in acceptance of larger body sizes as the norm. Directors of agencies involved in providing nutrition or health education, such as WIC, Head Start, or the Family Nutrition Program, need to be aware of these factors in order to provide culturally appropriate and sensitive weight management programs.

Direct inquiries about this study to the Project Contact listed above.