SPSS is good for conditional logistic regression analysis. Look at youtube for this case:

https://www.youtube.com/watch?v=be-M0tMMKyU

Hello Dawit Tiruneh. SPSS has no dedicated conditional logistic regression command. As the video Béatrice gave you shows, SPSS users have traditionally tricked the COXREG command into estimating the desired conditional logit mode. Here is some more (text-based) info about how to do that:

• https://www.ibm.com/support/pages/conditional-logistic-regression-using-coxreg

If you do have access to Stata, I suggest that you use it instead, as it does have a command (clogit) for estimating conditional logit models (with no need for trickery).

• https://www.stata.com/manuals/rclogit.pdf

If you only have access to SPSS, another option is to use generalized estimating equations (GEE) to account for the paired (or more generally, matched) data. Several years ago, I used Stata to estimate some conditional logit models for a physician-researcher. A few years after that, when SPSS added the GENLIN command with GEE as a feature, I used it to re-estimate the same models. The results did not match exactly those from clogit in Stata (and I did not expect them to match exactly), but they were very similar--see this old thread in the usenet newsgroup for SPSS:

• https://groups.google.com/forum/#!topic/comp.soft-sys.stat.spss/zkxR016mZxM

You may have to copy the table of results shown there and paste it somewhere that you can change the font to fixed to make things line up.

HTH.

kindly, thanks so much for your contributions and clear response to this question

It is not possible to do conditional logistic regression directly on SPSS. Instead, we need to format the data accordingly and use either multinomial logistic regression (matched one to one) or cox regression (one to one or one to many).