Personalisation in advance letters does not always increase response rates. Demographic correlates in a large scale experiment

This study was set up to examine whether personalising advance letters by adding names and appropriate salutation, affects the survey cooperation of subgroups in the general population differently, in analogy to findings that subgroups react differently to advance letters. Differential reactions could be an explanation for the mixed findings in the literature on personalisation of advance or cover letters. In a large scale study (N=39,518), information in communal registries was used to study (non)response patterns in subgroups, as a result of personalisation. Advance letters of the Dutch CAPI Labour Force Survey were randomly assigned to addresses. In the non-personalised version (N=30,899), letters were addressed to ‘the inhabitants of…’. In the personalised version (N=8,619), the name or names of the household core were derived from municipal registries and used for addressing the letter. A re-interview addressed the issue whether the advance letter was read by more households when the household was addressed by name. By linking the sample to registries, it was possible to study response behaviour of subgroups. The study focussed on groups the literature indicates as differentially reacting to advance letters, i.e., different age, ethnic, gender, household composition and income groups, and groups with or without a listed telephone number. Hardly any difference in the overall level of cooperation was found if either a personalised or non-personalised letter was used. However, differential reactions were found for listed versus unlisted telephone owners, where only listed households reacted positively to personalisation. In the other subgroups studied, no firm evidence of differential reactions were found. The paper discusses what these results signify for sample composition and the risk of bias.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.