Outcome effect expectancies are considered a key determinant of behavior in social learning models of substance use. The present study reports the development and psychometric properties of the Cocaine Effects Questionnaire (CEQ-P), designed specifically for use with patients in treatment for substance abuse. Items generated from patients and treatment staff were Likert-rated for frequency by 178 patients in substance abuse intensive day treatment. Seven components were derived and reduced into a 33-item measure with good reliability. The components loaded on two higher order components: positive effects (five scales) and negative effects (two scales). The CEQ-P shows good construct and concurrent validity in comparison with measures of similar constructs and with cocaine use variables. Few gender or racial differences were significant. Greater pretreatment cocaine use was associated with expecting less frequent positive effects and more frequent social withdrawal, consistent with clinical reports of later stage cocaine use. Alternatively, greater urges to use correlated positively with expected positive effects, specifically with expecting enhanced well-being, pain reduction, and sexual enhancement from cocaine. Less cocaine use during the 3-months after intensive treatment was predicted by expecting more negative effects from cocaine pretreatment. The CEQ-P is likely to be useful for both clinical and research purposes. Treatment implications were discussed.