This may help
http://www.nt.ntnu.no/users/skoge/prost/proceedings/ifac2008/data/papers/4226.pdf
Apparently R does that with the package ggm. Look at this link for an example with explanation: http://www.statmethods.net/stats/correlations.html
There several so-called "pseudo" correlations for statistics like logistic regression. Can you be more specific about what statistic you want?
Dear Daniel,
You are right for logistic but i need to find the pseudo correlation between diversity and biomass. Because biomass and diversity were calculated both from tree diameter, thats why I need to clarify their psuedo correlations. If I do the regression between them and i get high R2, reviewer may have question. If you have further question, please ask. Please suggest me the best way.
Dear Arshad,
The pseudo correlations, like Cox and Snell's, are methods to use the deviance statistic in a way that is conceptually similar with Pearson's correlation. There are a few of these because there is no simple agreed upon way of doing this with the deviance. Similar problems occur with mixed models. Here it sounds like either you will just have a large correlation, or is it that you want the correlation between diversity and biomass conditional on tree diameter. Common statistics for that are the part and partial correlations, which I believe are both in SPSS (I don't have SPSS, so can't check).
Does that make sense?
Dan
You may refer Rao and Boiroju (2013) Generation of correlated samples, InterStat Journal.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis