Hi,
I am doing a three way anova on my data. The correct design is two factor repeated (ie, within subject), and one of the factors between subject. For this, I need to use SPSS (as opposed to sigmaplot or prism). When i do three way anova (in sigmaplot), with no factors repeated, I get very high significance two of the factors, and all three are significant. I understand this is not the correct design for my data, but I was under the assumption, that once I figure out spss and do the stats using two of the factors repeated, I would only get better significance.
Thing is, I have now done the analysis, using one between subject, and two within subjects. (i.e. two factor repeated, one factor not repeated). This is the correct way given my experimental design. With this, there is MUCH LESS significance, and only one factor is significant.
My question, thus is, is it even POSSIBLE for a repeated measure analysis of the data to give far LESS significance then a non-repeated measure analysis of the data?
Thank you,
Sonya