Face-to-face surveys using paper or a computer: Difference in attitudes

This article examines the implications for data quality of using computer-assisted personal interviewing (CAPI) compared to paper-and-pencil interviewing (PAPI) in an electoral survey. Quality is measured by considering the average of unanswered questions in each modality and the number of non-responses in single-answer multiple choice questions. The analysis of the questionnaire as a whole reveals that the number of “don’t know” responses is higher than the “no answer” option, although the second is more frequent in PAPI. In forced choice questions requiring a single answer, the number of non-responses is greater than the “don’t know” responses; a figure that increases in PAPI. Of the six multiple choice questions, only three show significant differences, although the results are inconclusive.