Forty-six patients were randomly allocated to receive aversion therapy for homosexual impulses according to a classical, avoidance, or backward conditioning paradigm.
Before and three weeks after five days of treatment they were shown a film containing pictures of nude women preceded by pictures of a red circle and of nude men preceded by a green triangle. Prior to treatment they were also conditioned to tones followed by painful electric shocks. During all procedures their galvanic skin and penile volume responses were measured.
Three weeks after treatment the patients showed significantly less penile volume increase to the pictures of men and less penile volume decrease to the pictures of women; but no penile volume increase to the pictures of women. Subsequently each patient was to receive six booster treatments at monthly intervals. Twenty attended for all six. These patients responded better than those who received fewer booster treatments.
At one year following treatment approximately half the patients reported a decrease in homosexual feeling and half an increase in heterosexual feeling. Approximately a quarter reported an increase in heterosexual intercourse and a quarter a cessation of homosexual relations. There was no significant difference in efficacy between the three forms of treatment. The fact that backward conditioning was not less effective than the other two forms was considered to support the contention that these aversion therapies do not act by setting up conditioned reflexes. An alternative mode of action is suggested, compatible with the finding that measures of conditioning in the aversion procedure administered prior to treatment correlate with treatment response.