The relationships between self‐estimates of vocational interests and inventory scores have been the subject of research since 1941, but to date there has been no comprehensive review of that literature. Thirty‐seven articles relating to self‐evaluation of vocational interests were located in manual and computer searches of the psychological literature. The median values of 24 rank correlations (between self‐estimated and measured interests) derived from 11 idiographic studies ranged from .12 to .77. There was considerable variation in individuals' ability to estimate their own interests. A meta‐analytic procedure was used with 14 separate nomothetic studies (N = 4,282) reporting 316 product‐moment correlations. Correlations ranged from ‐.05 to .89 with a mean of .46 and around 99% were positive. Based on the Schmidt‐Hunter formulae, 13% of the observed variance was accountable for by reference to the expected variance and it was concluded that there were true differences in effect sizes across studies. Results indicated significant differences in the correlations depending on which inventory was used, but did not support any proposed influence of vocational interest areas. It is proposed that in the process of interest assessment, our questionnaires and inventories should be evaluated against self‐estimates.