In a survey conducted in Ontario, front-line staff of specialized addiction treatment services were asked to indicate the extent to which they believed 53 different treatment processes to be necessary for the effective treatment of people with alcohol and drug problems. Cognitive-behavioural processes were generally rated as almost essential for treatment to be effective. Other processes received mixed ratings, while confrontation and pharmacological treatment were, on average, rated as detrimental. Factor analysis identified three interpretable dimensions of beliefs (a) cognitive-behavioural, (b) disease, and (c) medication. A fourth dimension involved both psychodynamic and conditioning processes and did not clearly correspond with any known therapeutic approach. Scores on scales developed using items from these four dimensions were variously influenced by respondents age, education, place of work, and certification status. Two groups identified using cluster analysis differed primarily with respect to scores on the disease scale. Logistic regression analysis showed that age, certification status, and place of work were associated with membership in these groups. Implications for treatment system development and research are discussed.