ERIC Number: ED430182
Record Type: RIE
Publication Date: 1999-Apr-23
Limits of Meta-Analysis as a Basis for Justifying Individual Counseling Interventions.
Hummel, Thomas J.; Lichtenberg, James W.
The first part of this paper examines using meta-analysis as a basis for making probabilistic statements about client outcome by revisiting Smith and Glass's (1977) classic paper concerning meta-analysis of psychotherapy outcome studies. While it may be argued that making probabilistic statements was not the primary purpose of meta-analysis, Smith and Glass took it in that direction when they presented a figure depicting two overlapping normal distributions, one representing the treated population and the other representing the control population. They pointed out that a person at the mean of the treated population fell at the 75th percentile of the control group. From this it can be deduced that the probability is 0.75 of an individual randomly drawn from the treated population being above the mean of the control population. This paper assesses the justification for this statement using numerical analysis and model fitting techniques. Results show that normal distribution is probably not an appropriate model for treated subjects and working backwards from effect size distributions to client distributions seems doomed to failure. The second part of the paper considers the kinds of probabilistic statements that might be offered to clients and how three models of client outcomes might relate to the statements. Six figures depict the different model analyses. (MKA)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Conference of the American Educational Research Association (Montreal, Quebec, Canada, April 19-23, 1999).