A simple recommendation for the analysis of matching data.

Area of Science:

• Psychology
• Statistics

Background:

• The matching paradigm is widely used in psychology.
• Accurate statistical analysis is crucial for interpreting matching data.
• Failure to account for chance matches can lead to erroneous conclusions.

Purpose of the Study:

• To demonstrate the utility of the z test in analyzing matching data.
• To establish the expected number of chance matches and variance in matching paradigms.
• To provide guidelines on sample size for achieving desired statistical power.

Main Methods:

• Utilized the z test for analyzing data from matching paradigms.
• Derived the expected number of chance matches and associated variance.
• Evaluated Type I error rates and power at different sample sizes.

Main Results:

• The expected number of chance matches in a matching paradigm is 1.0, with a variance of 1.0.
• The z test maintains Type I error rates near the nominal significance level with sample sizes of 80 or 110.
• Larger sample sizes may be necessary to achieve 0.80 power, depending on effect size.

Conclusions:

• The z test is a suitable and accessible statistical tool for analyzing matching data.
• Researchers must account for chance expectations in matching studies.
• Sample size considerations are vital for ensuring adequate statistical power in matching research.