The Aphasiology Archive
University Library System, University of Pittsburgh
    

Is Busk and Serlin’s measure of therapy effect size d a suitable measure for use in therapy studies? Evidence from simulations

Howard, David and Rose, Miranda
Is Busk and Serlin’s measure of therapy effect size d a suitable measure for use in therapy studies? Evidence from simulations. In Clinical Aphasiology Conference: Clinical Aphasiology Conference (2009 : 39th : Keystone, CO : May 26-30, 2009) / : (2009).

Full text available as:
PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract

Beeson & Robey (2006) advocate Busk & Serlin’s d as a good measure of effect size, suitable for meta-analysis of single case experimental designs of therapy. Unfortunately there are no available data supporting the use of this measure.

We propose a set of criteria for a suitable measure of effect size and investigate whether d meets them in 1360 simulations each involving 10000 measures of d.
We show that d always overestimates the true effect size and this is primarily related to the number of items in the baseline series and the degree of lag 1 autocorrelation in the series.

We show that Busk & Serlin’s d does not meet any of our suggested criteria. It is difficult to see how its use can be justified.

EPrint Type:Clinical Aphasiology Paper
Subjects:UNSPECIFIED
ID Code:1998
Conference:Clinical Aphasiology Conference: Clinical Aphasiology Conference (2009 : 39th : Keystone, CO : May 26-30, 2009)
Additional Information:USED WITH PERMISSION.