Factors contributing to educational differences in obesity among women: evidence from South Korea

Abstract

Background: Obesity is more prevalent among less-educated women than highly-educated women around the world. However, little is known about the factors which cause this difference in obesity, and almost nothing is known about how the individual factors which explain differences in education among women alone contribute to obesity. In this study, we identified the factors which help explain the relationship between education and obesity in women, and quantified their separate contributions to obesity.

Methods: We analyzed information on 14,577 women aged 25 years or over using datasets from the Korea National Health and Nutrition Examination Survey (2010-2014). We divided the women into two education groups: women who had, at most, finished high school (less-educated women), and women who had college degrees and beyond (highly-educated women). Using an extended Oaxaca-Blinder method, we decomposed the difference in obesity prevalence between the two education groups into the contributions (%) due to two effects: composition effect and association effect.

Results: Obesity was more than twice as prevalent among the less-educated women (34.3%) than it was among the highly-educated women (16.0%). The composition effect-contribution of differences in the distribution of observed characteristics compared to that of the difference in obesity prevalence between the two education groups-was 38.2%. The association effect-contributions of differences in the estimated coefficients of characteristics compared to that of the difference in obesity prevalence between the two education groups-was 55.8%, of which lifestyle factors were the most important contributor (43.6%). Of the separate contributions of each factor, the association effect of the factor related to women's stress exhibited the largest contribution (23.0%).

Conclusion: We suggest that to effectively mitigate the high prevalence of obesity among less-educated women, it may be necessary to help low-educated women who do not feel stressful develop strategies to combat their higher risk of obesity. We also suggest the need to conduct decomposition studies in countries which show significant relationships between education and obesity among women, and to create targeted policies to reduce this population's overall risk of obesity.