ABSTRACT Construct validation of personality tests begins when items are selected according to substantive and structural postulates; nevertheless, substance has generally been emphasized to the neglect of structure. The present approach shows how a structural model—the interpersonal circumplex—could be used to ensure that test items for measures of interpersonal problems (e.g., loneliness, nonassertiveness) faithfully mirror the theoretical characteristics of their constructs. New item statistics are derived with applicability to item analysis and test construction, and the equivalence of item‐curve fitting and circumplex scoring is illuminated.
Stored only when both papers are in the corpus — 71 citations exist upstream.