| Abstract: | School psychologists commonly conduct interventions for individual students and evaluate their success through single-case research designs, most of which include a baseline observation period. However, concluding intervention effectiveness from a single-case design is problematic when pronounced baseline trend is evident. With positive baseline trend, interpretation of the measured mean difference between phases is ambiguous, and an invalid interpretation is probable. Although an adequate method for controlling baseline trend exists, it is rarely applied in published studies. This article seeks wider application of baseline trend control by describing a modified technique that blends more easily with visual analysis. A second goal of this article is to encourage better informed use of baseline trend control, by empirically demonstrating its need in published studies, and its impact on R2 effect sizes and on autocorrelation. Finally, baseline trend control is combined with elimination of autocorrelation, for results that can withstand the scrutiny of the broader research community. (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
