An Improved Rank Correlation Effect Size Statistic for Single-Case Designs: Baseline Corrected Tau

Area of Science:

• Psychology
• Behavior Analysis
• Research Methodology

Background:

• Single-case experimental designs (SCEDs) face challenges in measuring treatment effects when pretreatment performance shows an improving trend.
• Existing effect size statistics, like Tau-U, may inaccurately attribute baseline improvements to interventions due to inadequate trend control.

Observation:

• The Tau-U statistic, while popular, has limitations including inflated values, lack of bounds, inability to be visually graphed, and a high Type I error rate.
• These limitations can lead to the incorrect conclusion that ineffective treatments are effective.

Findings:

• A novel effect size statistic, Baseline Corrected Tau, is proposed, utilizing rank correlation and robust regression for improved baseline trend control.
• Field-testing with published and simulated single-case time series data demonstrated the efficacy of Baseline Corrected Tau.
• A web-based calculator for Baseline Corrected Tau is introduced to aid single-case investigators.

Implications:

• Baseline Corrected Tau offers a more accurate and reliable method for quantifying treatment effects in SCEDs, particularly when baseline trends are present.
• This improved statistic can reduce Type I errors, ensuring more valid conclusions about intervention effectiveness.
• The availability of a web-based calculator promotes the adoption and practical application of this advanced statistical tool in clinical and research settings.