Inconsistency in responding: Just inconsistency or something more?

Whereas Thompson (Social Behaviour and Personality, 3, 111–115, 1975) reports that response inconsistency is not relevant for the validity of Eysenckʹs P scale, data presented by Burgess and Kilfedder (Fourth European Conference on Personality, 1988), Burgess and Higgins (Fifth European Conference on Personality, 1990a) and Burgess and Perera (Fifth European Conference on Personality, 1990b) demonstrate that inconsistency contributes to the total variance of the P scale. The scope of this study was to examine the relationship between inconsistency and psychoticism and a number of other well-established personality dimensions, in addition to some self-rating scales. Subjects were twice asked to rate their own shyness on a seven-point scale during a 90-min period in which they responded to various personality questionnaires. Inconsistency was defined as an absolute difference in self-ratings. A somewhat higher inconsistency in response (t = 1.02; P> 0.05) was found in male students (M = 0.80; SD = 1.08; N = 150) compared to female students (M = 0.69; SD = 0.90; N = 211). Inconsistency was not significantly correlated with either Eysenckʹs personality scales, sensation seeking scales, machiavellianism, or with self-ratings on the following attributes: moody, conscientious, unpredictable, reliable and unsystematic. Marginally significant multiple determination coefficients show that inconsistency is not predictable from the described set of predictors. The findings indicate that in samples drawn from the normal population inconsistency can be treated simply as a measurement error and is not an indicator of psychoticism or some other psychological construct.