# How to conduct mediation analysis with ordinal (probit) regressions

#### rezarta

##### New Member
Hello everyone,

I need to conduct a mediation analysis in which all the variables (i.e., the predictor, mediator, and outcome) are single items measured on 4-point Likert type scales (i.e., strongly disagree, disagree, agree, strongly agree). I have conducted ordinal (probit) regression analyses to assess the impact of predictors on the outcome variable, however I was asked by reviewers to assess a mediation model. I did some research and I understand that the usual linear regression methods for assessing mediation do not work with ordinal regressions (I often use Hayes SPSS macros to conduct mediation analyses). I could find some guidance for mediation using logistic regressions, but I couldn't find much on ordinal regressions (except for some descriptions for doing such analyses on MPlus, which I never used before). Are there ways to conduct such analyses using SPSS? Not sure if it's relevant, but the sample size is also large (over 900 participants).

I'd really appreciate your input.

Thank you so much in advance.
Rezarta