As per literature, there are 4 types of logistic regression
1. Wald
2.Forward stepwise
3. Backward stepwise
What is the need of Goodness of fit model for computing logistic regression?
We had chosen calculation of unadjusted odd ratio in SPSS, in order to adjust the odd ratio what should be do?
For a presentation, I have computed only unadjusted odd ratio and not adjusted odd ratio, if the audience queries the reason for not adjusting odd ratio what should be my answer?
How do we any ways compute adjusted odd ratio in SPSS?
Can we graphically depict the computed odd ratio along with their Confidence Interval? If yes, then do we use excel or SPSS please tell how?
Please do respond as soon as possible as presentation on 15th October 2016
Please if possible do provide references
Thank you
Dear Ashish,
Refer the following links:
1. http://www.feinberg.northwestern.edu/sites/bcc/docs/2011-lectures/Lecture3LinearLogisticRegression2011.pdf
2. http://www.isqr.uni-freiburg.de/logistic.pdf
First of all, forward and backward stepwise model selection are not "different types of regression", they're just different strategies for model criticism. They still use the same regression technique.
It looks like you have a lot of unrelated questions, so rather than asking a bunch of questions it would be better for you to read an introductory text on logistic regressions and then ask more specific questions if you have any left over. Senthilvel gave some suggestions above. I would also recommend the logistic regression chapter in this book https://www.amazon.com/Using-Multivariate-Statistics-Barbara-Tabachnick/dp/0205849571
I don't think it's fair of you to expect people on ResearchGate to teach you logistic regression on very short notice. It's too bad that you have a presentation coming up soon, but it was your responsibility to learn the relevant statistics before you needed to present them; it's not very appropriate to ask a very broad/vague question like this and ask for urgent answers. If you don't have time to learn logistic regression before your talk, you can find other ways around it (e.g. present the data without statistics, present some preliminary statistics but acknowledge that you have more work to do, or do a simpler--even if not quite correct--analysis for temporary purposes).
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis