Hello Reseachgate community.
I have perused several recent sources to either find data or power tables missing and there I cannot seem to find the best source for an appropriate minimum sample size for a conditional process (moderated mediation) analysis.
With 4 variables (3 predictors, 1 outcome) and assuming power .80 with alpha .05 and small to moderate correlations between all (?) variables.
Could anyone point me in the right direction, please?
so far... "According to Preacher et al., (2007), assuming moderate strength correlations (r = .39), an a priori power analysis with bootstrapped bias-correction for testing the indirect effect (a = .05; 1 - b = .86) a sample size of at least 50 is required to test the indirect effect in a first-stage moderated mediation analysis." This seems too low, although if I bring it down to weaker correlations I need at least 200?
Please see the table from Preacher et al. (2007) attached, fair use considered...
- Badges
- Science topic
- Similar topics
- Statistics
- Survey