Effects of the Content of Personality Questionnaires on the Transitivity of Ratings (the Case of the Grit Scale Items)

Likert scale is widespread in psychological research when there is a need to measure different psychological traits. Its use implies that a respondent can order psychological traits in “more or less” terms. However, researchers usually do not test this assumption assumption. In this study we checked one of the prerequisites of order stucture of psyhological traits, that is assumptions of transitivity (i.e. if A > B и B > C then A > C). We also tested the hypothesis that transitivity errors occur because of the content of statements. We conducted a study where participants (N=250) made paired comparisons of manifestation of six psychological characteristics in different activity areas in which participants were engaged (i.e. work, study, etc.). Three statements were taken from consistency of interest subscale from the grit scale (Duckworth, 2007), the other three we elaborated based on the description of the construct of individual interest (Renninger, 1992). Results showed that most respondents (68.8%) did not have transitivity violaitions in their responses. However, those who made errors in transitivity made more errors in grit statements (24%), rather than in individual interest statements (16%). We suppose these differences are connected with the abstractness of the contents of grit statements and more concrete and formulated in behavioural terms content of the statements from individual interest scale.