Web Survey Bibliography

Title Computer-assisted longitudinal surveys
Year 2002
Access date 21.04.2004
Full text

pdf (30k)

Abstract In longitudinal surveys, respondents are followed across time detecting changes on their measured variables, in order to understand the dynamics of some process. The same respondents are contacted at possibly regular time intervals, and administered by means of a questionnaire that can be dependent from the wave of the survey and from answers collected at the previous times. Longitudinal surveys are likely to have very complex questionnaires, with a very large number of questions and with complicated paths that can be extremely difficult to administer by paper and pencil interviewing, in particular when sophisticated dependent interviewing is involved.
Dependent interviewing makes use of information available on subjects, coming from previous surveys or other data sources, in order to customize the current interview in different ways. Dependent interviewing is used to probe for changes, to avoid asking already known information, to remember previous answers, to give a feedback to respondent before asking new information. Dependent interviewing is one of the main focus in longitudinal surveys, but it is very difficult to implement in practice by paper and pencil interviewing because of the intrinsic logistical constraints of this method. Think only the need for the interviewer to access frequently “historical” information about the current respondent, on related papers, in order to follow the designed question flow: there is an enormous work to prepare materials for PAPI interviewing at each wave of the survey, and for all these efforts, we have an high risk of errors in question administration, producing non pertinent answers and missing responses.
The introduction of computer-assisted interviewing in longitudinal surveys promises to revolution this field of surveys, making really possible to carry out a longitudinal survey implementing advanced interviewing techniques using historical data on subjects, such as complex dependent interviewing, reactive probing techniques, proactive feedback techniques, and consistency checks across times.
Computer-assisted interviewing allows for handling the complexity of a longitudinal questionnaire, virtually without errors, even in self administered interviewing, through the programming logic built in a software system. By means of automated question flow administration and on-line consistency checks, this method guarantees better control on extra-sampling survey errors, providing a significant gain in survey data quality.
These great features are, we have to make it clear, quite theoretical, in the sense that an ideal CAI system could, in principle, implement in some proper way the techniques we are thinking to use in longitudinal surveys. Unfortunately, we have to say the main currently available CAI systems do not implement a satisfactory and effective support for longitudinal surveys, limiting what is really possible to put into effect among all we can think to do.

Access/Direct link Homepage - conference (full text)
Year of publication2002
Bibliographic typeConferences, workshops, tutorials, presentations
Print

Web Survey Bibliography - International Conference on Improving Surveys, 2002 (27)