Research attrition and sampling bias in psychotherapy outcome studies

Bridgette Michelle Sacco-Laurens (1999)

Participant attrition in outcome studies threatens both conclusions about treatment effectiveness and generalizability. Many researchers do not report the number or type of participants who drop out of research. A review of the literature that did report this information revealed applied settings found demographic differences among attrition groups more often than controlled studies. The most common differences were related to social class variables, such as education, income, or socioeconomic status. This study examined the impact of research attrition on sample representativeness in outpatient private practice settings. Intake information on participants who returned all requested follow-up materials were compared to those with fewer or no response to research follow-up requests. Major findings indicated that demographic variables related to social class, for example: education, occupation, and socioeconomic status, were significantly lower for the research attrition group, A therapist rating of symptom severity at intake was the only diagnostic variable related to research attrition. Previous treatment experience and attendance patterns were also significantly different among the response groups. Research compliance was moderately predicted by previous treatment experience, years of client education, occupation status and socioeconomic status.