0

Working on a project, and I have a mediation result where I'm curious if I'm interpreting this right. I'm a little rusty on my mediation interpretations. See image example below.

enter image description here

A pathway is positive

B pathway is negative

C' pathway is positive

The C pathway has a beta = .40, but after running the mediation model the C' pathway has a beta of .67

Mediation: A pathway beta = .56; B pathway beta = -.50

Using state data categorized by county. The directions of the mediation results are throwing my confidence in the interpretation a bit.

X = Pct Poor health status in a county

M = Pct voted for candidate A in a county

Y = Death Rate in a county

My interpretation:

Countywide poor health status positively predicts county death rate, but the result is mediated by the pct who voted for candidate A in the county.

Pct poor health status in a county positively predicted Pct Voted for Candidate A. Counties that were more likely to vote for Candidate A had a lower death rate.

Does that sound accurate? Thanks.

M_S_Pen
  • 49
  • 4
  • It would be helpful to see what happens to c (total) and c' (direct). – POC Mar 08 '21 at 17:47
  • c pathway's beta is .40 but after running the mediation model the c' pathway is .67. I'll put this as an edit in my original post. – M_S_Pen Mar 08 '21 at 18:50

1 Answers1

0

If you are sure that you are not in front of suppressor or confounding effect, and that the model makes sense, then your interpretation seems legit.

POC
  • 346
  • 1
  • 8
  • 23