Measure and control of non‐response in a mail survey

Abstract

PurposeThis paper shows several results obtained from the first investigation made in Spain using Dillman's total design method for mail surveys. In the spring of 1998 a survey was carried out on consumer behaviour in one particular Spanish region (Navarra). The response rate was 68.3 percent. One month later, a team of professional interviewers interviewed the people who did not respond to the mail survey. The objective is to compare the results obtained by these interviewers with an estimation of the non‐response after using several different methods and a group of adjustment procedures.Design/methodology/approachThe total design method (TDM) for mail surveys has produced some wonderful results in several countries in Northern and Central Europe, but as far as one can gather it has never been used in any of the countries of Southern Europe surrounding the Mediterranean. The excellent results obtained from mail surveys in the USA, and some of the findings confirmed by disciples of Dillman in several European countries, led to carrying out the first investigation in Spain using TDM through mail surveys. With the aim of resolving some of the above‐mentioned questions, it was decided to make the first‐ever investigation using TDM within Spanish society.FindingsThe results obtained by applying the total design method to a Mediterranean country do not differ greatly from data obtained by other researchers on the efficiency of mailing surveys in other countries. This fact forces one to have serious doubts about the declarations of some experts who place the response rate for mailing surveys at 40 percent in Spain. In this paper, different strategies to reduce the impact of non‐response in a mail survey are outlined. First, a second interview was carried out with a subsample of the interviewees who were non‐responders, with the aim of finding out more about the characteristics of those who do not respond. Later, a strategy of weighting was used using CHAID to identify the best predictors from among the set of variables available.Originality/valueIn the South of Europe there is very sparse use of mail surveys, while there is great use of face‐to‐face surveys: this is the opposite of the tendency in other countries, where telephone surveys are supplanting the face‐to‐face interview. This is the first research about non‐response in mail surveys in the South of Europe.