McGuire, Carlisle, and Young (1965) proposed that any deviant sexual behavior is the direct product of a deviant sexual preference. This theory evolved into the sexual preference hypothesis (Lalumiere and Quinsey, 1994), which holds that men who engage in sexually deviant behaviors do so because they prefer them to socially acceptable sexual behaviors. Classical conditioning theory would argue that this type of sexual interest is acquired through a concurrent experience of deviant stimuli and a sexually aroused physiological state that leads to a psychological association. This association would then be reinforced through masturbation to deviant fantasies (operant conditioning), while arousal experienced with nondeviant stimuli would gradually be extinguished due to a lack of reinforcement. This article focuses on various mechanisms for determining the nature of a person's erotic responses to various stimuli. The measurement of physical responses related to erotic excitement include penile plethysmography and pupillometry. Self-report measures of sexual response to various presented stimuli can include interviews, questionnaires, and card sorts. Procedures that use attention levels and duration for various stimuli and information processing are also reviewed. The problems inherent in each assessment procedure are discussed. The authors note that although plethysmography is generally considered the measure of choice in determining sexual preference, it is, like most other assessments of sexual preference, vulnerable to faking. The authors favor a combination of a card sort and the use of the Multiphasic Sex Inventory as the most cost-effective assessment of sexual preference and the measurement of the likelihood of faked responses. The discussion of the future development of assessment techniques focuses on attentional methodologies. 1 table and 133 references