Estimating slope and level change in N = 1 designs

Area of Science:

• Behavioral Science
• Psychology
• Research Methodology

Background:

• Single-case designs are crucial for evaluating interventions.
• Accurate estimation of treatment effects requires separating level and slope changes.
• Existing methods may not adequately control for baseline trends or serial dependence.

Purpose of the Study:

• To propose and validate a novel procedure for separately estimating level and slope changes in single-case designs.
• To eliminate baseline trend from data before assessing treatment effectiveness.
• To evaluate the bias and precision of the proposed estimators through simulation.

Main Methods:

• A new procedure for estimating level and slope change is detailed.
• Baseline trend is removed from the data series prior to analysis.
• A simulation study assessed estimator performance under various conditions (data generation models, serial dependence, trend, level/slope change).

Main Results:

• The proposed procedure provides unbiased estimates for level and slope changes.
• Baseline trend is effectively controlled.
• Slope change estimation shows acceptable efficiency.
• Level change estimator variance may be a concern with highly negatively autocorrelated data.

Conclusions:

• The new procedure offers a robust method for analyzing single-case designs by accurately estimating level and slope changes.
• The method effectively controls for baseline trends, enhancing treatment effect assessment.
• Further consideration of data characteristics, like negative autocorrelation, is advised for level change estimation.