Reinterview

A reinterview occurs when an original respondent is recontacted by someone from a survey organization—usually not the original interviewer—and some or all of the original questions are asked again. Reinterviewing can serve more than one purpose in survey research, including (a) verifying that interviews were actually completed as the researchers intended with sampled respondents, (b) checking on the reliability of the data that respondents provided when they were originally interviewed, and (c) further studying the variance of survey responses.

As part of quality assurance efforts to monitor the quality of survey data collection when the questionnaire is interviewer-administered, some part of the recontacts made to selected respondents may include asking some of the original questions again, especially the questions that the researchers deem as key ones. Although there are many reasons why a respondent may not provide the exact same answer during the reinterview, including some legitimate reasons, the purpose of re-asking some of the questions is not to match answers exactly but rather to make sure there is no consistent pattern of deviation from the original data that could signal that the questionnaire was not administered properly by the interviewer, including the possibility that the interviewer falsified some or all of the data.

A reinterview also provides data that can be used by researchers to test the reliability of the original data. For example, demographic characteristics are unlikely to change if the reinterview is conducted within a few weeks or even a few months of the original data collection, although some might, such as someone turning one year older in age, or someone becoming a college graduate because of a recent graduation, or now becoming employed in a new occupation in the interim since first being interviewed. Other types of questions, such those concerning behaviors, experiences, perceptions, knowledge, attitudes, and opinions, also are unlikely to change much within a few weeks or even a few months, but they can be expected to be more likely to change than will demographics. Small changes in these types of variables do not necessarily mean the original data are unreliable, but large changes often signal problems with the quality of (1) the original interviewing, (2) the questionnaire wording, (3) the data collection performed during the reinterview, or all three of these factors. Data gathered from re-asking the same questions in the reinterview also provide researchers with additional ways to understand the variance that is associated with their questionnaire items.

Granted, reinterviewing is costly, but when survey budgets allow for it, there are many benefits that can be gained.