In reply to lithos:
Hi again,
The article from the document (Cohen 1992)gives a table of estimated sample sizes for 8 types of design including ANOVA.
Is running it through gpower3 better because it factors in variation of measurement apparatus?
I downloaded gpower3 and followed the instructions but the input parameters were different, will that matter? Eg there was no f squared treatment effect parameter but there was an error parameter
The gpower3 doesn't function fully on my computer, the program window doesn't fully expanded and I have the opaque pane at the bottom of the small window (like when you can't print an article directly from a PDF link but have to save it to enable printing controls) i removed and re-installed, repaired, changes resolutions but nothing worked. Any suggestions or is it just a matter of trying on a different computer?