Handwritten 'post-it' notes, questionnaire formats and response to a postal questionnaire survey

We were interested to read the recent letter by Osler et al.1 comparing the effect of using stamped return envelopes instead of the usual pre-printed business reply envelopes on postal survey response rates. Our research is based in a deprived inner-city area that has, for many years, given low response rates of 40–60% in postal surveys of samples drawn from family practice registers.2–4 A very high family practice ‘list inflation’,5 high population mobility, low levels of education and a high proportion of people whose first language is not English may contribute to this non-response. In anecdotal reports, using a handwritten ‘post-it’ note to encourage questionnaire response has been reported to increase response rates substantially.6 The size and format of the questionnaire might also be expected to have an effect on response rates. We report a randomized comparison of the effect these two interventions on response to a questionnaire survey. The purpose of the questionnaire was to evaluate experiences of healthcare utilization in participants aged ≥65 years. The trial had a factorial design, and randomization was in blocks, stratified by family practice. The study was approved by the research ethics committee of St Thomas’ Hospital, London. We evaluated two interventions. A small yellow ‘Post-it’ note bearing the handwritten phrase ‘Many thanks for all your help’ was attached to the front of half of the questionnaires. In addition, half of the questionnaires were produced in A4 format and half in A5 booklet format (50% smaller). The questions and layout were otherwise identical. The questionnaires were printed on yellow paper. Following an initial mailing, non-responders received a second mailing of the same questionnaire after a period of ∼3 weeks. An initial sample of 3000 subjects was taken by sampling 200 subjects from each of the 15 family practices. The study had 90% power to detect a 6% difference in response rates. The results are presented in Table 1. There were 1,130 questionnaires returned over two mailings giving a response rate of 38%. Questionnaires were returned as undeliverable only for 16 subjects. There was little evidence that either intervention increased the response rate. The response rate for questionnaires sent with or without a handwritten ‘Post-It’ note attached differed by only ∼3% over the two mailings. Similar response rates were obtained for the two questionnaire formats. This suggests that there may be positive and negative attributes for each. The A4 format is designed to be more legible due to its larger font size, but it also uses twice as much paper, and may therefore give the impression of a being a lengthier questionnaire, whereas for the A5 format the reverse is true. Systematic reviews have shown varying results for methods that seek to increase the personalization of the survey.7,8 For example, the effect of using handwritten notes has been found to range between an 11% decrease in response to a 6% increase.7 Response rate in relation to questionnaire size and presence/absence of handwritten ‘Post-it’ message in a postal survey in patients aged ≥65 years in South London Response rate in relation to questionnaire size and presence/absence of handwritten ‘Post-it’ message in a postal survey in patients aged ≥65 years in South London We share the concern of Osler et al. regarding the need for further examination of response effects in population surveys. However, our results suggest that in contexts in which the propensity to return a questionnaire is low, subtle modifications to postal survey strategies may be expected to have limited effectiveness, and alternative survey methods should also be considered.