Abstract
Three studies were conducted exploring the effects of two questionnaire formats (grouping versus randomized items) on internal consistency and test-retest reliability, convergent and discriminant validity, and accuracy in respondent-provided descriptions. The results of these studies (two correlational, one experimental) were reasonably consistent in showing that neither format is clearly superior to the other on any of the examined psychometric properties, although a weak trend showed that the grouped item format produced slightly superior measurement properties. Over-all, the results did not support the usefulness of grouping question-naire items. Implications and future research directions are briefly discussed.
