Send to:

Choose Destination
See comment in PubMed Commons below
J Clin Epidemiol. 1997 Aug;50(8):869-79.

Methodological problems in the retrospective computation of responsiveness to change: the lesson of Cronbach.

Author information

  • 1Department of Clinical Epidemiology and Biostatistics, McMaster University of Hamilton, Ontario, Canada.



To examine the relation between responsiveness coefficients derived directly from a calculation of average change resulting from a treatment intervention (Responsiveness-Treatment or RT) and those derived from retrospective analysis of changed and unchanged groups (Responsiveness Retrospective or RR) based on a global measure of change.


Two approaches were used. First, we used simulation methods to examine the analytical relationship between the RT and RR coefficients. We then located eight studies where it was possible to compute both RT and RR coefficients. As anticipated from theoretical arguments, the RR coefficients were larger than the RT coefficients (1.50 versus 0.41, p < .0001). Within study there was no predictable relationship between the two indices. Across studies, the magnitude of the RR coefficient was strongly related to the correlation with the retrospective global scale, and unrelated to the magnitude of the RT coefficient. The simulated curves fit well with the observed data, and substantiated the observation that the relation between RT and RR coefficients is complex and only weakly related to the size of the treatment effect.


Retrospective methods of computing responsiveness yield little information about the ability of an instrument to detect treatment effects, and should not be used as a basis for choice of an instrument for applications to clinical trials.

[PubMed - indexed for MEDLINE]

LinkOut - more resources

Full Text Sources

Other Literature Sources

PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Elsevier Science
    Loading ...
    Write to the Help Desk