Accounting for baseline trends in intervention studies: Methods, effect sizes, and software

Abstract

Single-case experimental studies are relevant and important to substantiate the effectiveness of behavioral, clinical, and educational interventions. When a baseline trend is present in intervention studies, it is challenging for researchers to determine, visually or statistically, if an effect is due to the intervention or to the naturally occurring trend. In this paper, we demonstrated and appraised four methods that can assess an intervention effect in the presence of baseline trends. These four methods quantified an intervention effect as a phase change effect size (i.e., mean phase difference and the slope and level change) or a nonoverlap effect size (i.e., Tau-UAB-A and Tauc). Empirical data from an intervention study were used in the demonstration. All methods were evaluated in terms of types of intervention effects assessed, assumptions, specialized computing tools, and application issues (e.g., missing scores). To empower researchers to use these four methods, we provide a summary of each method’s appropriate inferences about an intervention effect and its computing tools’ strengths and limitations. Based on the results, we recommend using multiple baseline trend-controlled methods to draw a conclusion about an intervention effect.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 21 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.