Testing for a General Changepoint in Medical and Psychometric Studies: Changes Detection and Sample Size Planning

Area of Science:

• Statistics
• Biostatistics
• Psychometrics

Background:

• Change detection is crucial in medical and psychometric research.
• Existing methods for change point detection often require complex computations like resampling or simulations.
• Accurate power analysis and sample size planning are essential for study design.

Purpose of the Study:

• Introduce a new, computationally simple method for change detection using the pseudo Score statistic.
• Determine the sampling distribution of the pseudo Score statistic under the alternative hypothesis.
• Facilitate power analysis and sample size planning in segmented regression with unknown breakpoints.

Main Methods:

• Utilized the recently introduced pseudo Score statistic for change detection.
• Determined the sampling distribution under the alternative hypothesis.
• Developed a method that bypasses the need for resampling or simulations to find critical values.

Main Results:

• The proposed method offers known null and alternative distributions, simplifying power and sample size calculations.
• Simulation studies demonstrated superior performance compared to other Tests for a Change Point (TFCP) with both normal and binary data.
• Real data analyses were successfully conducted on genomic and SAT critical reading datasets.

Conclusions:

• The new pseudo Score statistic-based method provides a computationally efficient and accurate approach to change detection.
• This method enhances power analysis and sample size planning in segmented regression.
• The tool is accessible via CRAN and a user-friendly Shiny App, supporting medical and psychometric research.