von PonderStibbons » Mi 13. Jul 2011, 08:52
Formeln kenne ich leider keine, womöglich kannst Du Dir den Referenzartikel aus dem unten zitierten Artikel besorgen.
(Aus: REVIEW OF EDUCATIONAL RESEARCH 2010, Vol. 80, pp 4-33 )
"Peterson and Brown (2005) compared the accuracy of several equations that translate standardized beta coefficients (β) to correlation coefficients (r). When betas are reasonably close to zero (| r | ≤ .12), their analyses showed that the best equation for estimating correlation coefficients is simply r = β. Moreover, in a sample of more than 1,500 studies in the social sciences, they found that the number of predictors in a regression equation was not significantly related to the difference between r and β. In other words, β corresponds to r equally well, regardless of the number of covariates in the regression equation. As a result, the current study directly substituted beta coefficients for correlation coefficients.
Referenz:
Peterson RA, Brown SP (2005). On the use of beta coefficients in meta-analysis. Journal of Applied Psychology 90:175–181.
HTH
P.