Means of increasing response rates in a Norwegian dietary survey among infants - results from a pseudo-randomized pilot study

Postal surveys are widely used in scientific studies, including dietary surveys, but few studies about methods to increase participation in national dietary surveys are published. In the present study we compared response rates in a pilot study to a national dietary survey among infants using two different incentives (gift certificate or lottery), personalization in the form of handwritten name and address vs. a printed label and mode of sending out invitations (e-mail or postal invitation). In this parallel-design pseudo-randomized pilot trial, a nationally representative sample of 698 mothers of infants aged 6 and 12 months was drawn from the Norwegian National Registry and invited to complete a food frequency questionnaire about their infant's diet. One half of the mothers of 6 month olds were randomized by alternation to the lottery group (n = 198) and offered to participate in a lottery of two prizes (500 EUR and 1000 EUR). The other half (n = 200) was offered a gift certificate (50 EUR) upon completion of the questionnaire. Each incentive group was randomized by alternation to receiving an invitation with handwritten name and address or a printed label. For the mothers of infants aged 12 months (n = 300), 150 mothers received an e-mail invitation and 150 mothers received a postal invitation. Logistic regression was used for testing differences between the groups. The response rate was significantly higher (p = 0.028) in the gift certificate group (72%) than in the lottery group (62%). No difference was seen between those receiving an invitation with a handwritten name and address (68%) compared to a printed label (66%, p = 0.72). A somewhat higher response rate was seen when using the postal (50%) compared to the e-mail invitation (43%, p = 0.25). In this pseudo-randomized parallel-design trial of women participating in a national dietary survey among infants, the response rate was higher when offered a gift certificate than when participating in a lottery. Handwritten name and address did not affect participation compared to a printed label. Only a moderate difference was seen between the postal and e-mail invitation. Others conducting similar methodological studies are encouraged to publish their results to expand the knowledge basis in this area.