Related Experiment Videos
Assessing spurious "moderator effects": Illustrated substantively with the hypothesized ("synergistic") relation
1University of Illinois, Urbana-Champaign.
Psychological Bulletin
|May 1, 1990
Summary
Traditional regression analysis confirmed a spurious link between math ability and spatial skills in advanced math. A more rigorous approach revealed this finding was an error, highlighting the need for improved moderator analysis methods.
Area of Science:
- Psychology
- Educational Psychology
- Quantitative Psychology
Background:
- Hierarchical multiple regression analysis is a common method for assessing moderator variables.
- This technique may have drawbacks for corroborating psychological theories.
- Previous research has explored the relationship between mathematical ability and spatial visualization.
Purpose of the Study:
- To examine the traditional methodology for assessing moderator variables.
- To illustrate the possible drawbacks of hierarchical multiple regression analysis.
- To test the hypothesis of a synergistic relation between mathematical ability and spatial visualization in advanced mathematics.
Main Methods:
- Empirical analysis of a large dataset (400,000 subjects).
- Application of traditional hierarchical multiple regression analysis.
- Utilized a more systematic approach for re-analysis.
Main Results:
- The traditional methodology initially confirmed a synergistic relation between mathematical ability and spatial visualization.
- Further analysis using a more systematic approach demonstrated this finding to be spurious.
- The study identified potential Type I and Type II errors in moderator effect assessment.
Conclusions:
- The traditional methodology for assessing moderator variables can lead to spurious findings.
- A more systematic approach is necessary for accurate moderator effect assessment.
- Suggestions are provided for modifying traditional methods to minimize errors in applied and theoretical research.