it depends  on the type of variables whether continuous or desecrate variables.

you have t-test, one way anova , correlation and chi square, for bi-variate analysis

If women empowerment is treated like a dichotomous variable, then why not try for logistic regression.

Yes, nicely explained by Michael R. Frone.

you can also make the modgraphs of interaction effect of microcredit*age and women empowerment

Hi,

For you question: How to analyze the moderating effect of demographic variables on dependent variable using spss? 

Kindly see the following links and its very useful to you.

http://www.theanalysisfactor.com/spss-glm-choosing-fixed-factors-and-covariates/

http://www.personal.psu.edu/jxb14/M554/specreg/process2012.pdf

http://www.aessweb.com/pdf-files/aefr-2014-4(7)-893-907.pdf

http://en.wikipedia.org/wiki/Moderation_(statistics)

Thank you all of you for giving valuable advice to me.  After referring those things if have doubt I will contact with you.

Hi