Hello Ylenia,

The answer depends on how you intend to use the means & SDs for your analysis.

The pretest and posttest scores for a given group are likely to have different SDs, so I'm unclear as to why you'd want to combine these into a single value that will likely represent neither occasion accurately. If, on the other hand, your metric of interest was pre-to-post change (Mpost - Mpre), you could compute the SD of the difference scores if the pre-post correlations were given for each group: SD(diff) = sqrt (SDpre^2 + SDpost^2 -2*corr(pre,post)*SDpre*SDpost). If pre & post scores are uncorrelated (not very likely), then SD(diff) = sqrt(SDpre^2 + SDpost^2).

If your analysis had to do with standardized differences, it would be easiest to use Glass's delta: (Mpost - Mpre) / SDpre for each group. These then represent change relative to the baseline values.

Good luck with your work.

Hello Ylenia,

The answer depends on how you intend to use the means & SDs for your analysis.

The pretest and posttest scores for a given group are likely to have different SDs, so I'm unclear as to why you'd want to combine these into a single value that will likely represent neither occasion accurately. If, on the other hand, your metric of interest was pre-to-post change (Mpost - Mpre), you could compute the SD of the difference scores if the pre-post correlations were given for each group: SD(diff) = sqrt (SDpre^2 + SDpost^2 -2*corr(pre,post)*SDpre*SDpost). If pre & post scores are uncorrelated (not very likely), then SD(diff) = sqrt(SDpre^2 + SDpost^2).

If your analysis had to do with standardized differences, it would be easiest to use Glass's delta: (Mpost - Mpre) / SDpre for each group. These then represent change relative to the baseline values.

Good luck with your work.

Thank you David Morse for your answer.

I need to calculate the standardized mean difference so i guess that Glass's delta is the ricorrect approach, isn't it?

Thank you.

David Morse gives an excellent suggestion. Ingram Olkin and Gene Glass were two of the founders of the subject and I found their work to be very helpful when I was in grad school. I have given links to their work below in case you would find it helpful as well. Best wishes, D. Booth

https://www.google.com/search?q=ingram+olkin+metaanalysis&rlz=1C1CHBF_enUS874US874&oq=ingram+olkin+metaanalysis&aqs=chrome..69i57.21042j0j1&sourceid=chrome&ie=UTF-8

https://www.google.com/search?rlz=1C1CHBF_enUS874US874&ei=jkT0XYutA4nKtQbmt7AI&q=Gene+Glass+meta+analysis&oq=Gene+Glass+meta+analysis&gs_l=psy-ab.3..0.11963.11963..13165...0.0..0.364.720.1j0j1j1......0....1..gws-wiz.......0i13.9VaZZ-eEHV0&ved=0ahUKEwiLrMLYh7TmAhUJZc0KHeYbDAEQ4dUDCAs&uact=5