Re: significance of correlation affected by sample size?



Conceptually:

Statistical significance = effect size x sample size

e.g., in concrete form

t = d x (df^0.5)/2 ... (for independent t using Cohen's d)

.... as d or df increases t (and statistical significance) increases ...

This is taken straight from Robert Rosenthal's book chapter in the
Handbook of Research Synthesis.



Thom

.