Sampling variance calculation for effect sizes from pretest-posttest-control group designs

[Kristina Loderer]

I have a question regarding the computation of effect sizes in the form of the correlation coefficient (r) from pretest-posttest-control group designs. I am planning on including such designs in my analysis and have calculated effect size estimates based on the mean pre-post change in the treatment group minus the mean pre-post change in the control group, divided by the pooled pretest standard deviation, as recommended by Morris (2008).

Subsequently, I converted this effect size to ‘r’, or, more accurately, Fisher’s z-transformed coefficients. I am now a bit confused regarding the calculating of the sampling variance – for Fisher’s z-transformed correlation coefficients, the sampling variance is calculated using the formula ” 1 divided by N minus 3 “. Papers dealing with the calculation of effect sizes for these designs only focus on measures of the standardized mean difference (i.e., d or g).

I am wondering whether this sampling variance makes sense given the specific study design that I have obtained the z-transformed ‘r’ from, namely the pre-post-control group design..any thoughts? I feel like this would be too simple, given the complex formulas of sampling variances for d or g for these study designs..

Thanks!

Kristina

[Author]

Posts