The dominant issue in personality research over the last decade has been concerned with the fundamental structure of personality and the best measures of that structure. Exploratory factor analysis was used to investigate possible three- and five-factor solutions to the Eysenck Personality Profiler (EPP; Eysenck, Barrett, Wilson, & Jackson, 1992) which consists of 21 primary scales categorized under three super-factors. Little evidence was found to support Costa and McCrae's (1995) unequivocal comment that a five-factor solution fitted the data well. Confirmatory factor analysis was also used, by means of structural equation modelling, to estimate the goodness of fit of three- and five-factor models and little evidence was found to favour one solution over the other. A shorter version of the EPP, which consists of just nine scales, seemed to favour a three-factor solution. Various criticisms of the EPP are also made: some scales have relatively low alpha, there seem to be too many neuroticism scales and the three category response scales seem less than ideal.