My dependent variable is dichotomous and I have five mediators. To my understanding the Sobel test is limited to one mediation. However, I did find an article by Preacher and Hayes 2008 where they had four mediators. On Hayes website you can download the Macro called indirect for SPSS. Has anyone ever tried using this before? I recognize the obvious answer would be to use Lisrel; however, I prefer SPSS.
If you are mediating variables in the model indirectly impact the LISREL it is better to use
I've used the INDIRECT macro on spss several times. It's very easy to use - the instructions are all on the downloadable file.
Basically, there is one line of syntax where you have to insert your variable names (this all clearly labelled so you know what goes where). The output is very easy to interpret too (if you are familiar with mediation analysis).
I have that macro and I will be using it. Do you just report the confidence intervals? Lines of syntax? I just click on regression under analyze and the INDIRECT macro is there. Is there a reason for syntax? Also, have you published with these results?
When I have used Preacher & Hayes, I have used syntax downloaded from Andrew Hayes' website (I can't find the link in my version of SPSS).
The downloads can be found below.
http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html
I have it already. Thank you though. Do you know if the mediators can be dichotomous? I do not believe so. For example, "Did you pledge to remain a virgin until marriage?" Yes, No.
i don't think SOBEL can but PROCESS may be able to. I know the outcome can be dichotomous.
I have discussed in detail the philosophy of Sobel's test with example, kindly find the attached file,
Article Making Sense of Mediating Analysis: A Marketing Perspective
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS