| Literature DB >> 11497023 |
E W Lambert1, A Doucette, L Bickman.
Abstract
The pre-post design has been the workhorse of outcome evaluations for many years. Using data from a study of 984 treated children (ages 5 to 17 years), this article argues that there are two structural problems with the pre-post evaluation of outcome: (1) excessively large intervals of uncertainty for individual outcomes and (2) paradoxical inconsistencies in the evaluation of groups. These problems can be solved by designs with three or more repeated measures analyzed with longitudinal multilevel analytic models.Entities:
Mesh:
Year: 2001 PMID: 11497023 DOI: 10.1007/bf02287244
Source DB: PubMed Journal: J Behav Health Serv Res ISSN: 1094-3412 Impact factor: 1.505