Are respondents who omit conflict tactics scale items more violent than those who omit none? A methodological note

In a study of 31,801 active-duty Army military men and women, the authors encountered protocols in which moderate or severe Conflict Tactics Scale (CTS) items had been omitted. Dropping respondents who omit items could introduce biases into the data. We investigated the hypothesis that those respondents who omitted an item differed in aggression from those who omitted none and hence should not be dropped from the analysis without further consideration. We tested this by means of three comparisons of those who omitted one item and those who omitted none: classification of respondents into no aggression, moderate aggression, and severe aggression: differences between mean CTS scores; and comparison of one-time aggressors with those who reported more than one incident in a year. All three methods indicated that the group that omitted one item was more aggressive than the group that omitted none.