The Wary Student, Part 8: Interaction-Effect Results

Because the negative cooperation p value missed .01 (.033) significance because of the Bonferroni correction, an analysis of covariance (ANCOVA) was attempted in order to increase the power of the statistical tests. The ANCOVA analysis attempts to use a second variable (called a “covariate” or CV) to explain some of the sample variation in the dependent variable (DV), which increases the observed power of the independent variable (IV). In order to run an ANCOVA, two assumptions must be made: first the CV is correlated with the DV but not the DV, and that the DV and and the IV do not interact. This second requirement called the homogeneity of regression, is similar to the homogeneity of variance test used in the ANOVA

Only one of the assumptions for ANCOVA was met. A perason’s correlation showed that the test for neutral cooperation was positively correlated with negative cooperation (F = .298, p < .01) but not correlated with cognitive load (F = .004, p = .955). Then, to test the homogeneity of regression assumption a two-way ANOVA was run with both the IV and the CV as independent variables. The results failed to allow an ANCOVA, but are fascinating in their own right.


Raw visualization of the interaction effect.
Cognitive load decreases the negative cooperation of those who were neutrally cooperative

Specifically, the two-way ANOVA results (F = 22.257, p = .0000048, η2 = .112) show a strong interaction effect. That is, cognitive load strongly effects the negative cooperative behavior of those who are neutrally cooperative. In this two-factor ANOVA, both cognitive load (F = 25.144, p = 0.0000013) and neutral cooperation (F = 15.171, p = 0.00014) were significant. Tukey post-hoc tests were not run because both condition and neutral cooperation had only two levels. Among those who accepted the unfair allocate resources, increased cognitive load barely altered the degree of negative cooperation (decreasing it from an average of .09 to .07 points). However, among those who rejected the unfair division, higher cognitive load reduced negative cooperation (from n = 21, μ = .8571, to n = 20, μ = 0).


The Wary Student, a tdaxp research project
1. Abstract
2. Cognitive Load
3. Cooperative Behavior
4. Method
5. The Experiments
6. Hypotheses
7. Main-Effect Results
8. Interaction-Effect Results
9. Discussion
10. Future Research
11. Bibliography

Leave a Reply

Your email address will not be published. Required fields are marked *