| pubmed-article:7774980 | pubmed:abstractText | We describe, illustrate, and make available a menu-driven PC program which can be used to assess the effect of a treatment on growth when random allocation of subjects to the treatment and control groups is not feasible. Three different estimators of, and confidence intervals for, this effect are computed, namely, the simple gains, standardized gains, and covariance adjusted estimators. It is shown by means of several examples that these estimators can differ substantially, and some guidelines for choosing between them in specific circumstances are provided. | lld:pubmed |
