Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Interpretation of mediation analysis results with medeff

    Hi,

    I hope this is the right place to post this question.

    I have performed a mediation analysis with medeff.

    I have an independent variable T, a mediator M, and outcome Y. (All 3 variables are binary, and I use logit)

    I first specify the two equations required by medeff.

    I make reference to the standard mediation triangle, where path a connects T and M, path b connects M and Y, and path c' connects T and Y.

    Equation 1 is simply regressing M on T and some control variables. (this gives and estimate for path a)
    Equation 2 regresses Y on M + X +control variables. (this gives estimates for b and c')

    What I find is the following:

    I get a significant negative association betwee

    Equation 1
    M Coef. Std. Err. z P>z [95% Conf. Interval]
    T -3.458955 .1044635 -33.11 0.000 -3.6637 -3.25421
    Equation 2
    Y Coef. Std. Err. z P>z [95% Conf. Interval]
    T .6081007 .0949606 6.40 0.000 .4219813 .7942201
    M -.1496489 .0989504 -1.51 0.130 -.3435881 .0442904
    So, path a is significant and negative, path b is not significant, and path c' is positive and significant.

    So, based on that, from what I understand, I would conclude that there is no mediation because path b is not significant.

    However, medeff produces the following results:

    Effect Mean [95% Conf. Interval]
    ACME1 .0189176 -.005018 .0438004
    ACME0 .0160403 -.0041128 .0379756
    Direct Effect 1 .1274808 .0903369 .1646811
    Direct Effect 0 .1246035 .0863965 .1633032
    Total Effect .1435211 .1134255 .1722113
    % of Total via ACME1 .1313853 .1098513 .1667854
    % of Total via ACME0 .1114022 .0931433 .141418
    Average Mediation .017479 -.0045654 .0409513
    Average Direct Effect .1260421 .088528 .1640466
    % of Tot Eff mediated .1213937 .1014973 .1541017


    So, medeff says there is 12% mediation (looking at the last row, % of Total Effect mediated.)

    So, my first question is, given the results from the two regression analysis above, should I ignore the medeff results and conclude that there is no mediation?
    Or is there actually mediation?

    Second question, why if medeff says that there is 12% mediation, would the confidence interval for "average mediation" include 0, ie not significant?


    Thank you for any help you can provide.


    T Jones






  • #2
    Hi again, I tried changing the title because this isn't really a question about medeff, it's simply a question of whether I should apply mediation analysis when part of mediation path is not significant? If not, then are the results that a mediation command such as medeff produces simply nonsense? Or are they still interpretable? Also, is it strictly 5% level that we deem it "significant"? What about at 10 or 15% level? Thank you

    Comment


    • #3
      Re: deem it significant: It being the mediation path

      Comment


      • #4
        Cross-posted at http://stats.stackexchange.com/quest...-mediat#123969

        Please see FAQ Advice on our cross-posting policy, which is that you should tell us about it.

        Comment


        • #5
          Hi, sorry, I did not know. Will do so in future. thx

          Comment


          • #6
            Dear Statalisters,

            I have the same question regarding medeff: Does the CI from -0.xx to +0.xx mean for the indirect effect mean, that it is not significant?
            Additionally, I'm not sure whether I can use a continuous variable as IV?

            Thank you and best regards
            Christina

            Comment

            Working...
            X