Metainformation
| Tag | Value |
|---|---|
| file | Inferential_Statistics_vufsw-bootstrap-0291-en_vufsw-bootstrap-0291-en |
| name | vufsw-bootstrap-0291-en |
| section | inferential statistics/bootstrap |
| type | schoice |
| solution | TRUE, FALSE, FALSE, FALSE |
| Type | conceptual |
| Program | NA |
| Language | English |
| Level | statistical reasoning |
Question
Look at the mediation model below. Why is the bootstrap method the best method to test the mediation effect?
- TRUE: Because the bootstrap method is based on testing the indirect effect αb and not the difference between the total effect c and the direct effect c’ and because the bootstrap method is reliable even with small samples
- FALSE: Ecause with the bootstrap method the indirect effect is always equal to both αb and c - c’ and because the bootstrap method is reliable is reliable even with small samples
- FALSE: Ecause the bootstrap method is based on testing the indirect effect αb and not the difference between the total effect c and the the direct effect c’ and because with the bootstrap method the indirect effect is always equal to both αb and c - c’
- FALSE: Ecause the bootstrap method is based on testing the indirect effect αb and not the difference between the total effect c and the the direct effect c’ and because the bootstrap method is reliable even with large samples
Solution
The bootstrap method and the Sobel test are better than the Baron and Kenny method because they test a meditation effect by performing a direct test for the indirect effect αb. The Baron and Kenny is a rather outdated method that tests tests a mediation effect by testing the difference between the total effect and the direct effect c-c’ and in which 4 conditions must be satisfied. The first condition (i.e. the existence of a total effect) is not actually necessary. In addition, the Baron and Kenny method puts too much emphasis on statistical significance.
In addition, the bootstrap is better than the Sobel test because the Sobel test uses a point estimate for αb . This requires the sample to be relatively large so that the assumption of multivariate normality is met for αb.
- True
- False
- False
- False