This meta-analytic review of 148 studies on child and adolescent direct and indirect aggression examined the magnitude of gender differences, intercorrelations between forms, and associations with maladjustment. Results confirmed prior findings of gender differences (favoring boys) in direct aggression and trivial gender differences in indirect aggression. Results also indicated a substantial intercorrelation (r = .76) between these forms. Despite this high intercorrelation, the 2 forms showed unique associations with maladjustment: Direct aggression is more strongly related to externalizing problems, poor peer relations, and low prosocial behavior, and indirect aggression is related to internalizing problems and higher prosocial behavior. Moderation of these effect sizes by method of assessment, age, gender, and several additional variables were systematically investigated.

Short forms are so-named because they consist of shortened versions of original scales. Creating and using such measures is something of a cottage industry for many practicing scientists. A PsycINFO search conducted in late December 2008 using short form and short forms as alternate keywords returned 5,101 references, over 4,200 of which were articles in refereed journals. Many of these references had an applied measurement or methodological focus, describing the creation and psychometric evaluation of a short form of an extant instrument. For example, Rammstedt and John (2007) recently created a 10-item version of the 44-item Big Five Inventory (John, Donahue, & Kentle, 1991) as a way to measure the five broad dimensions of personality (i.e., Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness) in 1 minute or less. Researchers often create short forms so they can assess a larger number of constructs in a reasonably short testing session. One particular advantage is that use of short forms can ensure that measures of many potentially critical variables are obtained so the researcher has greater latitude in testing alternative hypotheses.Regardless of any compelling reasons for using short forms, the most likely outcome of using them is that the researcher will have a measure of the construct that has poorer psychometric properties than does the original long