Four experimental interview styles, designed to differ in the extent of their use of active fact-oriented and active feeling-oriented techniques, were compared in relation to their use in the initial diagnostic interviews with the mothers of children referred to a psychiatric clinic. All four styles proved to be effective in eliciting emotions and feelings, but the findings suggested that each was effective for different reasons. It appeared that emotional expression could be encouraged by the interviewer's response to emotional cues, by a reflective style with little factual cross-questioning, by the use of direct requests for self-disclosures, by the optimal (but not necessarily maximal) use of interpretations. and expressions of sympathy, and by direct requests for feelings.