I calculated average marginal effects for negative binomial models by using the following code.
library(margins)
summary(m3 <- glm.nb(V7 ~ V8 + V3 + V10 + V12 + V2 + V5, data = data)) cplot(m3, what = "effect")
But I got the following error: What do I have to do fix this? What is the code to calculate elasticities for random parameter negative binomial models?
> summary(m3 <- glm.nb(V7 ~ V8 + V3 + V10 + V12 + V2 + V5, data = data))
Call:
glm.nb(formula = V7 ~ V8 + V3 + V10 + V12 + V2 + V5, data = data,
init.theta = 0.4868761319, link = log)
Deviance Residuals:
Min 1Q Median 3Q Max
-1.3170 -1.0447 -0.7771 0.1641 2.2438
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -8.979e-01 7.002e-01 -1.282 0.199738
V8 -1.444e-06 4.385e-06 -0.329 0.741903
V3 4.422e-02 1.141e-02 3.874 0.000107 ***
V10 1.933e-05 3.184e-03 0.006 0.995156
V12 -2.394e-03 6.119e-03 -0.391 0.695662
V2 8.191e-03 1.841e-01 0.045 0.964502
V5 5.945e-03 5.945e-03 1.000 0.317260
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for Negative Binomial(0.4869) family taken to be 1)
Null deviance: 269.15 on 285 degrees of freedom
Residual deviance: 232.13 on 279 degrees of freedom
AIC: 726.87
Number of Fisher Scoring iterations: 1
Theta: 0.4869
Std. Err.: 0.0901
2 x log-likelihood: -710.8690
> cplot(m3, what = "effect")
Error in eval(model[["call"]][["data"]], env) :
promise already under evaluation: recursive default argument reference or earlier problems?