Regional and district peer effects in obesity: A UK panel data analysis.

Abstract

Most research concerning the causes of obesity has focused on macroeconomic trends or individual-level determinants relating to socioeconomic status. There is now growing interest in the psychosocial determinants of obesity. Little research has been conducted into wide-scale peer effects and so the aim of this study was to determine whether regional and district peer effects determine individual-level obesity in the UK. Data from the British Household Panel Survey (for the years 2004 and 2006) were used to model individual-level BMI and probability of obesity as a function of regional and district BMI, controlling for a variety of other variables. Instrumental variable techniques were used to investigate simultaneity bias. We found no statistically significant regional peer effect in the UK. For men, we found that a 1 point increase in average district BMI increased individual BMI by 0.640 and probability of being obese by 1.9% points; for women, BMI increased by 0.230 and probability of being obese by 1.6% points. Men appear to be more susceptible to these effects than women. Peer effects at the district level may have played a significant role in the rise of obesity in the UK. Further research must aim to disentangle the two routes by which these processes are theorised to function.