Does professional training make a therapist more effective?

Abstract

Jeffrey S. Berman and Nicholas C. NortonUniversity of Texas at AustinIn a recent review, Hattie, Sharpley, and Rogers (1984) concluded that patientstreated by paraprofessional therapists improve more than those treated by profes-sionals. However, this provocative conclusion is based on inappropriate studies andstatistical analyses. The present review omitted problematic studies and organizedthe data to permit valid statistical inference. Unlike Hattie et al., we found thatprofessional and paraprofessional therapists were generally equal in effectiveness.Our analyses also suggested that professionals may be better for brief treatmentsand older patients, although these differences were slight. Current research evidencedoes not indicate that paraprofessionals are more effective, but neither does it revealany substantial superiority for the professionally trained therapist.In his 1952 critique Eysenck noted the lackof evidenc e for the efficacy of psychotherapy,and he concluded that it would be prematureto formulate guidelines for training psycho-therapists. During the intervening years, re-searchers have responded to Eysenck's chal-lenge by demonstrating convincingly that pa-tients who undergo psychotherapy improvemore than those who do not receive treatment(e.g., see Shapiro & Shapiro, 1982; Smith,Glass, & Miller, 1980). However, such evidenceconcerning the efficacy of psychotherapy doesnot necessarily indicate the usefulness ofprofessional training programs. For instance,psychotherapy with highly trained therapistsmight not result in outcomes that are any bet-ter than those obtained by untrained thera-pists. To gauge the benefits of professionaltraining, we also need to know whether trainedtherapists are more effective than those withoutsuch training.One prominent assessment of this issue wasprovided by Durlak (1979), who reviewedstudies comparing professional therapists withparaprofessionals having no formal training.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 218 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.