So, I calculated a negative binomial regression model and I am trying to estimate the mean marginal effects in R.

To do this, I used the mfx package and wrote the following code:

negbinmfx(formula = attack ~ cinc + alliance_memberships + maj.pow.fac + 
            I(log(int.users2)) + myedu + I(log(gdp_pc)) + polity2 + years_since_dispute,
          data = data,
          atmean = TRUE)

My dependent variable measures the number of cyberattacks a country initiates per year.

The output from this command is the following:

Marginal Effects:
                           dF/dx   Std. Err. z P>|z|
cinc                  3.2182e-33  1.8067e-27 0     1
alliance_memberships  5.6471e-35  3.1703e-29 0     1
maj.pow.facyes       -7.5705e-35  4.2501e-29 0     1
I(log(int.users2))    2.2773e-34  1.2785e-28 0     1
myedu                -4.2954e-35  2.4115e-29 0     1
I(log(gdp_pc))       -1.5224e-34  8.5471e-29 0     1
polity2              -2.6135e-35  1.4672e-29 0     1
years_since_dispute  -1.4429e-32  7.9944e-27 0     1

dF/dx is for discrete change for the following variables:

[1] "maj.pow.facyes"

Now, what I am unclear about is the last part. maj.pow.fac is a simple dummy variable with the levels "yes" and "no".


dF/dx is for discrete change for the following variables: [1] "maj.pow.facyes"

...mean that these marginal effects are calculated by holding the dummy maj.pow.fac constant at the level "yes", or what do the effects mean in relation to the dummy?

I am sorry if this question is simple to some. I am a complete beginner and have not been able to figure this out using the mfx package's documentation.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.