Zastosowanie testów psychologicznych do oceny osób odpowiadających na pytania w wywiadzie kwestionariuszowym

Abstract

In order to assess the degree of confidence with regard to the results obtained in questionnaire interviews, a researcher may measure the appropriate traits of respondents which influence the validity of their answers. On the basis of saturation of a given group of respondents with such traits, a researcher may estimate the quality of their answers obtained in interviews. It was assumed that the basic traits which are likely to disturb the validity of answers, are: 1° poor vocabulary, 2° a predisposition to show off, 3° a predisposition to answer the questionnaire questions without thinking over their contents. Three tests were prepared to measure these traits: 1° a test of the vocabulary (V), 2° a test of lying (L), 3° a test of dummy, or “sleeper” questions (D). The basic hypothesis was formulated, as follows: respondents showing a high level on the score of the V-test, and low scores on the L-test and D-test are giving correct answers more frequently than those having the opposite scores. The hypothesis was corroborated in methodological investigations conducted in Włocławek in 1973. Coefficients of correlation between particular scores and the numbers of correct answers (A) are surprisingly high: rVA=0,65, rLA=0,40, rDA=0,53.