README.md

In staudtlex/ame: Average Marginal Effects

ame

ame computes simulation based (Krinsky-Robb) average marginal effects and first differences for various estimators in R.

Main goals: compute marginal effects and standard errors for: + continuous variables (dy/dx) + dichotomous variables (first differences) + interactions of continuous variables (dy/dx) + interactions of categorial variables (first differences) + spline-functions (for generalized additive models)

Installation

To install ame, simply run

devtools::install_github("staudtlex/ame")

Examples

1. Logit model, graduate school admission data

library(ame)

# graduate school data
gsa <- read.csv("http://www.ats.ucla.edu/stat/data/binary.csv")
gsa$rank <- as.factor(gsa$rank)

# run glm logit model
glm_fit <- glm(admit ~ gre + gpa + rank, data = gsa, family = binomial(link = "logit"))
summary(glm_fit)

Coefficients:
             Estimate Std. Error z value Pr(>|z|)    
(Intercept) -3.989979   1.139951  -3.500 0.000465 ***
gre          0.002264   0.001094   2.070 0.038465 *  
gpa          0.804038   0.331819   2.423 0.015388 *  
rank2       -0.675443   0.316490  -2.134 0.032829 *  
rank3       -1.340204   0.345306  -3.881 0.000104 ***
rank4       -1.551464   0.417832  -3.713 0.000205 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

# compute average marginal effects
glm_ame <- ame(glm_fit, cont_vars = c("gre", "gpa"), fac_vars = c("rank"), nsim = 1000)

Use the summary() command to display the average marginal effects and first differences of the model variables:

Continuous variables: gre, gpa 
Factor variables:     rank 

             dydx  Std. Error z value  Pr(>|z|)    
gre    0.00043102  0.00020583  2.0941 0.0355246 *  
gpa    0.15578509  0.06248429  2.4932 0.0124070 *  
rank2 -0.15452167  0.07524027 -2.0537 0.0392037 *  
rank3 -0.28434508  0.07528385 -3.7770 0.0001556 ***
rank4 -0.31680193  0.08239552 -3.8449 0.0001182 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Note: For a factor variable f, dydx corresponds to the first difference E(Y|f_i) - E(Y|f_0)

2. Linear model, mtcars-data

# linear model using glm
lm1 <- glm(mpg ~ cyl * hp + wt, data = mtcars, family = gaussian)
summary(lm1)

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) 52.017520   4.916935  10.579 4.18e-11 ***
cyl         -2.742125   0.800228  -3.427  0.00197 ** 
hp          -0.163594   0.052122  -3.139  0.00408 ** 
wt          -3.119815   0.661322  -4.718 6.51e-05 ***
cyl:hp       0.018954   0.006645   2.852  0.00823 ** 
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Compute average marginal effects and first differences.

lm1_ame <- ame(lm2, cont_vars = c("cyl", "hp", "wt"), nsim = 1000, seed = 99)
summary(lm1_ame)

Display results:

Continuous variables: cyl, hp, wt 
Factor variables:      

         dydx Std. Error z value  Pr(>|z|)    
cyl  0.031662   0.602009  0.0526  0.938894    
hp  -0.046123   0.014881 -3.0993  0.001901 ** 
wt  -3.133662   0.650566 -4.8168 1.429e-06 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Note: For a factor variable f, dydx corresponds to the first difference E(Y|f_i) - E(Y|f_0)

Note

This version of ame supports glm objects with + gaussian model family with "identity" link + binomial model family with "logit"/"probit" link

staudtlex/ame documentation built on Feb. 12, 2020, 10:26 p.m.