SPSS and SAS programs for determining the number of components using parallel analysis and velicer's MAP test

Area of Science:

• * Quantitative Psychology
• * Statistical Software Development

Background:

• * Standard statistical software often lacks validated methods for determining the number of components in factor and principal component analyses.
• * Researchers frequently rely on simpler, yet statistically flawed, methods like the eigenvalues-greater-than-one rule.
• * Validated techniques such as parallel analysis and Velicer's minimum average partial (MAP) test are widely recommended but underutilized.

Purpose of the Study:

• * To introduce brief and efficient programs for conducting parallel analysis and the MAP test.
• * To facilitate the integration of validated component determination procedures within SPSS and SAS environments.
• * To encourage broader adoption of statistically sound methods in factor and principal component analyses.

Main Methods:

• * Development of programming scripts for SPSS and SAS.
• * Implementation of parallel analysis algorithms.
• * Implementation of Velicer's minimum average partial (MAP) test algorithms.

Main Results:

• * Successfully created functional and efficient programs for parallel analysis and MAP test within SPSS and SAS.
• * Demonstrated the feasibility of conducting advanced component selection techniques in widely used statistical software.
• * Provided researchers with accessible tools to replace flawed component determination rules.

Conclusions:

• * The developed SPSS and SAS programs enable researchers to readily implement validated methods for determining the number of components.
• * Increased accessibility of parallel analysis and MAP test in familiar software environments is expected to improve the quality of factor and principal component analyses.
• * This work addresses a critical gap in statistical software, promoting more rigorous data analysis practices.