Representatives of 103 agencies were interviewed across the three research sites to determine agency auspices, staffing practices, services rendered, revenue sources and expenditures, clientele, and relations with police agencies. Both public and private agencies -- some sponsored directly by police departments -- were studied. The research could only partially substantiate police complaints that referral agencies were often unavailable when needed most. Agencies which were open 24 hours a day were no more likely to receive referrals than were agencies open only during the daytime. Officers' referral behavior in initial response to calls for service, their attitudes toward referral and social service provision, and their relations with referral agencies were also studied. Data were drawn from observation of officer activities while on patrol and from interviews with observed officers and their supervisors. Although the number of encounters suited to referral is undetermined, referrals occurred in only 5 percent of the more than 5,700 observed police-citizen encounters. Suggestions for referral also occurred in an additional 7 percent of the observed encounters. Referrals were usually determined by the nature of the problem regardless of citizen behavior characteristics or officer attitudes. When citizens were referred, they were generally satisfied with police handling of the problem and the referral agency response. Several recommendations for further research are made, including the need to assess referral's effectiveness in meeting the goals for which it was designed. Tables, diagrams, footnotes, and 17 references are provided. An interview form and a list of types of referral agencies are appended. For companion documents see NCJ 78362 and 77317.