Variety and frequency scales of antisocial involvement: Which one is better?

Purpose. The aim of this study was to examine whether a scale including frequency scores of antisocial behaviour is a more sensitive and better measure of antisocial involvement than a variety scale. Methods. Data from a representative sample of 1,292 Norwegian students aged 13 and 14 years was used to compare a 17-item variety scale with two versions of a frequency scale covering the same 17 items. Internal consistency of the scales, stability coefficients (for 1-year and 2-year intervals), and associations with conceptually related variables (e.g. aggression, opposition, alcohol consumption) were examined. Results. Results indicated that using a scale including the (raw) frequencies of antisocial acts committed instead of a variety scale would result in reduced internal consistency, lower stability over time, smaller group differences and weaker associations with conceptually related variables. Similarly, in regression analyses the (raw) frequency scale contributed little to the explained variance in conceptually related variables over and beyond that contributed by the variety scale. Conclusions. Although the results for a log-transformed frequency scale were about as good as for the variety scale, both practical, methodological, and to some extent, conceptual considerations argued against replacing the variety scale with the transformed frequency scale. However, there may be some minor benefits from using an appropriately transformed frequency scale as a supplement to a variety scale, and there are also situations where inclusion of frequency scores in antisocial scales may be essential.