Hi All,
I really hope that someone may be able to shed some light on an issue that I am having with interpreting some mediation analysis.
I have dichotomous variables (for the manipulations (X), mediators (M) and the outcome(Y)) so the Process macro (and associated syntax) of which is very easy to use and interpret is not an option at this time.
I have used the excel spreadsheet formulated by Nathaniel Herr (adapted from Kenny, 2006; mediation with dichotomous variables) which calculates Baron and Kenny's percentage (proportion) of effect mediated. To get this value I have run descriptive statistics to calculate the SDs for X, Y, and M variables, calculated the covariate (for X,M) and run logistical regressions to find affects of X on M, X on Y, M on Y, and X on Y through M.
Given that I am unfamiliar to this, the spreadsheet doesn't appear to be very clear, and whilst I could report the results obtained for each logistic regression formed within SPSS, this would mean that the spreadsheet makes no sense to undertake. There is obviously some data that needs to be reported from this (likely the Baron and Kenny's percentage (proportion)of effect mediated) but I am a little unsure of how to report this.
Has anyone conducted this analysis before, are able to help with this at all, or have come across papers that have used this method before?
Thanks in advance.
Emma
|