Token monetary incentives improve mail survey response rates and participant retention: results from a large randomised prospective study of mid-age New Zealand women.

AIM: To examine if a small token monetary incentive (NZ$5) increases mail survey response rates and participant retention of 40-50 year old New Zealand women. METHOD: In 2009, 2500 women were randomly selected from the New Zealand electoral rolls for a prospective study investigating factors related to the prevention of weight gain. At baseline, 400 women were randomly assigned to receive NZ$5 with the initial survey mail-out in addition to nonmonetary gifts to encourage participation (pen, tea bag, entry in lottery draw) received by all women. At 2 years, 400 women (200 received NZ$5 at baseline and 200 had not) were randomly assigned to receive the same token monetary incentive. At 3 years, all women identifying as an ethnic minority (n=234) and 300 randomly selected women of other ethnicities received the token monetary incentive with the initial mail out. RESULTS: The baseline response rate for women who received NZ$5 was significantly higher than for women who did not (76% vs 64%, p<0.001). At 2 years, retention rate for all women who received NZ$5 was significantly higher than for women who did not (88% vs 80%, p<0.001). At 3 years, among those women not identifying as an ethnic minority, the retention rate for those who received NZ$5 was significantly higher than for those who did not (84% vs 77%, p=0.014). CONCLUSION: Inclusion of a small token monetary incentive significantly increases mail survey response rates and participant retention in mid-age New Zealand women.

Agreement between measured and self-reported height, weight and BMI in predominantly European middle-aged New Zealanders: findings from a nationwide 1989 survey.

AIM: To examine the agreement between self-reported and measured height, weight and BMI in 40-50 year old New Zealand men and women. METHOD: Self-reported and measured height, weight and derived BMI were examined using data from 345 40-50 year old participants in the 1989/1990 Life In New Zealand Survey. Factors associated with biased reporting were assessed using regression models. RESULTS: Height was overestimated by men (1.08 cm, 95%CI 0.58, 1.59 p<0.001) and women (0.61 cm, 95%CI 0.35, 0.87, p<0.001), contributing to a small but statistically significant underestimation of BMI (men 0.31 kg/m2, 95%CI 0.14, 0.48, p<0.001; women 0.26 kg/m2 95%CI 0.11, 0.41, p<0.001). Weight was not statistically significantly misestimated by either sex. Bias in self-reported BMI did not differ by age, sex, ethnicity, SES or measured BMI category (all p-values <0.231). Agreement between BMI categories based on self-reported and measured data was very good (kappa: men 0.80, women 0.93). CONCLUSION: Self-reported height and weight from New Zealand 40-50 year olds in 1989 produced BMI estimates valid for use in epidemiological studies, especially when used as a continuous variable. Our analyses need to be replicated using data from a current and representative New Zealand sample.