Hi everyone,
I'm working on a meta-analysis and i have a study that shows its results in a table. Its results were analized with a General Lineal Model, and its show the R^2 that results from the relation between the dependent and the independent variable.
It is possible calculate the z Fisher with the R^2 that is given as result from a General Lineal Model????
Thanks in advance
Fernando
I presume the study has a single independent variable and you mean the Fisher z transformation.
z = 0.5 loge [(1+R)/(1-R)] (z becomes standard normal variate))
If you are only using a general linear model (opposed to a generalized linear model), then you can simply take the square root of your R2 and input it into the formula given by James above. If you are using a generalized linear model (say, with a binary outcome), the many R2 metrics one can use are not equivalent to the R2 above. Instead it is possible to convert the odds ratio for your independent variable into the functionally equivalent phi coefficient (~r) and then plug into the above z equation as well.
http://psycnet.apa.org/index.cfm?fa=buy.optionToBuy&id=2007-04834-017
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis