Dealing with Possible Baseline Inequalities Between Experimental Groups - The Case of Motor Learning

Area of Science:

• Experimental Design
• Motor Learning Research
• Biostatistics

Background:

• Random assignment is crucial for preventing bias and establishing causality in experiments.
• Simple randomization can lead to unequal baseline group characteristics, especially with small sample sizes.
• Motor learning studies often have small sample sizes, increasing the risk of baseline inequality.

Purpose of the Study:

• To assess the probability of unequal baseline groups using simple randomization in small sample sizes.
• To evaluate the effectiveness of stratified permuted-block randomization in achieving baseline equality.
• To provide guidance for researchers on choosing appropriate randomization methods.

Main Methods:

• Developed code to calculate the probability of unequal groups with simple randomization.
• Examined an example of stratified permuted-block randomization.
• Analyzed variables commonly measured in motor learning experiments.

Main Results:

• Simple randomization shows a high probability of unequal baseline groups in typical motor learning sample sizes.
• Stratified permuted-block randomization significantly improves baseline equality among experimental groups.
• The choice of randomization impacts the reliability of cause-and-effect conclusions.

Conclusions:

• Simple randomization is often inadequate for small motor learning studies due to potential baseline inequality.
• Stratified permuted-block randomization is a superior method for ensuring group comparability in such research.
• A decision flowchart is proposed to aid researchers in selecting randomization techniques.