Patterns of response by sociodemographic characteristics and recruitment methods for women in UK population surveys and cohort studies.

Women are an important public health focus, because they are more likely to experience some social determinants of disease, and they influence family health. Little research has explored the sociodemographic representativeness of women in research studies. We examined the representativeness of female respondents across four sociodemographic factors in UK population surveys and cohort studies. Six UK population-based health surveys (from 2009-2013) and eight Medical Research Council cohort studies (from 1991 to 2014) were included. Percentages of women respondents by age, income/occupation, education status, and ethnicity were compared against contemporary population estimates. Women aged <35 years were under-represented. The oldest women were under-represented in four of nine studies. Within income/occupation, at the highest deprivation level, the range was 4 percent under-representation to 43 percent over-representation; at the lowest level, it was 6 percent under-representation to 21 percent over representation. Of nine studies reporting educational level, four under-represented women without school qualifications, and three under-represented women with degrees. One of five studies over-represented non-white groups and under-represented white women (by 9 percent). Response patterns varied by topic and recruitment and data collection methods. Future research should focus upon the methods used to identify, reach, and engage women to improve representativeness in studies addressing health behaviors.