Personality structure among prisoners: how valid is the five-factor model, and can it offer support for Eysenck's...

Research with offenders has tended to use a categorical system of diagnosis or a three-factor model. There is growing evidence among non-offending groups that a five-factor model (FFM), which is more holistic and emphasises strengths as well as limitations, may be more clinically useful.

The purpose of this study was to test the validity of the FFM of personality in a sample of adult male prisoners. It was predicted that they would show higher scores for extraversion (E) and lower scores for emotional stability (ES) [the equivalent of higher neuroticism (N)], that E would be predicted by prison experience and age and that the FFM would be confirmed as a better fit to the data over Eysenck's three-factor [P(psychoticism)EN] model.

Two independent samples of adult male prisoners were asked to complete the International Personality Item Pool while locked in their cells over the lunchtime period. Parallel and confirmatory factor analyses were used to test the hypotheses.

There was a 65 and 75% response rate across the independent samples, resulting in samples of 259 and 183, respectively. Overall, more positive personality traits were evidenced across both samples. E was not predicted by prison experience or age, with increased ES related to increased age in only one sample. The best-fitting model comprised five factors, but represented a variant of the FFM.

The value of applying broad conceptualisations of personality to special samples is discussed, with the importance of accounting for positive, as well as negative, traits acknowledged. The presence of increased positive traits in comparison to negative traits is not in keeping with Eysenck's original personality theory of criminal behaviour, with the study highlighting value in accounting more for environmental factors.