Background Ethnic differences in obesity have been well documented, but comparatively little is known about whether these differences vary according to the degree of exposure to the new environment following migration.

Objectives To investigate the extent of generational differences in adult health-related lifestyles and socioeconomic circumstances, and explore whether these differences might explain changing patterns of obesity in ethnic minorities in England.

Method Seven ethnic minority groups were selected from the ethnically boosted 1999 and 2004 Health Survey for England (Indian n=887; Pakistani n=603; Bangladeshi n=275; Black Caribbean n=762; Black African n=147; Chinese n=413; and Irish n=1438). A White group was used as a reference population (n=5899). Age and sex adjusted logistic regression estimated the odds of having a poor health behaviour in the second generation compared to the first. Age- and sex-adjusted odds of being obese in the second generation compared to the first were estimated before and after adjusting for generational differences in health related behaviours (snacking, eating cakes and fried foods, having low levels of physical exercise, any drinking including binges, current smoking status) and socioeconomic factors (social class, equivalised income and highest qualification).

Results Overall, second generation ethnic minority men and women were significantly more likely than the first to have low vegetable consumption (<one portion/day), snack daily on chocolate, crisps, biscuits and cakes, be current drinkers and binge drink. Second generation women were more likely to currently smoke than the first generation. Conversely, men and women were significantly less likely to have low levels of physical activity in the second generation. However, there were considerable variations in the uptake of these behaviours within individual groups. Indian (OR: 1.76; 1.14, 2.71) and Chinese (OR: 3.65; 1.37, 9.78) groups were more likely to be obese in the second generation than the first after adjusting for age and sex, with no significant differences observed in all other groups. Adjusting for health behaviours in each ethnic minority group had a negligible impact on the risk of second generation obesity. However, the risk of obesity increased in all groups after adjusting for the better socioeconomic circumstances of the second generation.

Conclusions Socioeconomic shifts determine generational differences in obesity risk to a greater extent than acculturative changes in behaviours. Findings suggest that generational variation in obesity rates for ethnic minorities may be more effectively controlled through reductions in wider socioeconomic inequalities rather than targeting individual health related behaviours.