I am trying to work out the statistics that I need for my Thesis, and I have LOTS of questions.
1. I need to do a meta-analysis on a number of randomised controlled
trials. Since what I'm interested in with these type of trials is the
interaction effect of group x time, ive been told that the relevant
effect size for this is partial eta squared. Are you aware of any
programs that do this? I have not been able to locate one thus far.
2. If no program allows this, then I am happy to calculate by hand, however the formula is: SSeffect/(SSeffect + SSerror), and I know that doing an ANOVA in SPSS will actually give me the associated partial eta squared.
However, what I can't work out is how to do the ANOVA. Usually, I would input raw data and calculate it the normal way - however, in this situation I only have (in most cases) means, standard deviations and number of participants. How can I do an ANOVA - by hand or otherwise, so that I can calculate the SSeffect and SSerror and thus the partial eta squared statistic?
3. Some studies I have raw mean and SD data while for others
I have study statistics (like F and chi-squared values) and others I
only have p-values. Do any programs allow me to do a meta-analysis
with this variable information?
any help will be much appreciated!!!
Advertise on Talk Stats