I want to know which one is the best for interpretation of logit regression.
Hi Vivian,
I strongly recommend reading: Regression Models for Categorical Dependent Variables Using Stata, J. Scott Long and Jeremy Freese, Stata Press, 2014, ISBN-13:
978-1-59718-111-2, even if you do not use Stata, it provides excellent examples.
http://www.stata.com/bookstore/regression-models-categorical-dependent-variables/
You should have a look at the book:
Hosmer Jr, David W., Stanley Lemeshow, and Rodney X. Sturdivant. Applied logistic regression. John Wiley & Sons, 2013.
You can find detailed information regarding binary and multinomial logistic regression.
Which you should use depends on the research question you want to answer. Typically, relative risk will be more informative and succinct since it is usually invariant and closely tied to the model parameters. The marginal effect differs, I believe, according to the values of one or more independend variables corresponding to the point at which it is evaluated. So, if the question is a general question about which IVs are influential then go with relative risk. For a question about the amount of effect at specific values of IVs then the marginal effect provides the most direct answer.
Relative risk is ratio of probabilities while marginal effects is change in odds by changing one unit.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis