The purpose of the present study was twofold: First, to test Rachman's (1976) theory (Behaviour Research and Therapy, 14, 125-131) which predicts that "directly" conditioned fears will differ from "indirectly" conditioned fears in magnitude and anxiety response patterns. Secondly, to explore validity issues related to the questionnaire methodology typically used in fear acquisition research. The questionnaire comprised 3 anxiety scales and 3 fear-onset questions used in prior research, a specifically developed 36-item fear list and 2 validity-check items. One hundred and ninety-one Ss completed the questionnaire. After selecting and rank-ordering 10 fears from the fear list, Ss answered questionnaire items for their first- and then their tenth-ranked fear. Results failed to confirm Rachman's predictions: A significantly greater proportion of Ss ascribed highly-feared and moderately-feared situations or things to direct conditioning. In addition, differential anxiety response patterns were not present for different levels of fear. However, results supported the prediction that direct-conditioning ascriptions would be endorsed with greater certainty. The findings suggest that direct-conditioning experiences may be more memorable than indirect-conditioning events. The theoretical and methodological implications of the findings are discussed. It is suggested that future research either employ methodologies more suited to investigating causal relationships or that Rachman's (1976) theory be blended with an attributional account of fear acquisition. © 1995.