Standard drink definitions (10 g ethanol) were provided with pictures (e.g., a glass of beer) and the number of drinks in typical containers. Respondents selected a descriptor for their cigarettes use: “Never smoked or never smoked regularly”, “Do not smoke now but used to smoke”, “Occasionally smoke (on average, < 1/day)”, “Currently smoke cigarettes regularly (≥ 1/day)”. Respondents indicated how many servings of fruit (fresh, frozen, canned or stewed) and how many servings of vegetables (fresh frozen, canned) they ate per day. Examples were given to illustrate serving sizes. Respondents indicated separately for weekdays Crenolanib ic50 and weekends how much time

they were physically active, including walking to campus or shops, housework, shopping, sport, and exercise. Respondents indicated their height in

metres or feet and inches and their weight in kilograms or pounds. There were a total of 78 questions in the questionnaire though it should be noted that with branching and skip patterns most participants (e.g., non-drinkers) will not have been presented with all of the questions. Of 7130 students invited, 3283 (46%) participated. University response rates ranged from 53% to 72% (63% overall) while polytechnic response rates ranged from 15% to 36% (24% overall). Response did not vary by age and gender, but Māori were less Obeticholic Acid molecular weight likely to participate (42%) than non-Māori (48%; p < 0.001). Table 1 summarises risk behaviour and overweight/obesity prevalence, by gender, as a function of latency to response. Late respondents were significantly more likely to be also binge drinkers

in high school and to be physical inactive. The differences for being overweight/obese, smoking, and diet were in the expected direction but non-significant. We conducted the analyses separately for the polytechnic colleges versus universities finding results that were consistent for all five parameters so we have reported only the combined results. Table 2 shows prevalence estimates adjusted under the assumption that non-respondents have the same prevalence of these behaviours as late respondents, and the extent of non-response bias in absolute and relative terms. Late respondents had a higher prevalence of binge drinking and non-compliance with physical activity guidelines. Differences in the prevalence of non-compliance with dietary guidelines, smoking and overweight/obesity were non-significant but in the expected direction. The apparent non-response bias for binge drinking was mainly driven by differences among men. For physical activity, the effects were mainly driven by differences among women. Notably, smokers were significantly over-represented among female late respondents even though the overall result was non-significant.