This study examined the relationship between subjects’ actual test derived scores and their estimates of what those scores would be. Sixty subjects completed the 16 PF (form D) and then estimated the scores on each dimension for themselves and another person they knew well. The results showed significant positive correlations on 9 of the 16 dimensions for themselves. The dimensions they were best at estimating were Desurgency-Surgency, Untroubled adequacy-guilt proneness and Threctia-Parmia. Only two correlations (both negative) reached significance concerning their ability to predict another known person’s scores. Whereas subjects believed they were like the other person they nominated (13 of the 16 correlations were significantly positive), in actual fact their test derived scores showed only two significant findings, one positive and the other negative. Results are discussed in terms of lay theories of personality and their relationship to personality assessment.