Description Usage Arguments Value Author(s) References See Also Examples
Compute average marginal effects for generalized linear and generalized additive models using Krinsky-Robb Monte-Carlo simulation.
1 | ame(model, cont_vars, fac_vars, nsim = 100, seed = 1)
|
model |
an object of class |
cont_vars |
a character vector containing the names of continuous model variables. |
fac_vars |
a character vector with factor model variables. |
nsim |
number of simulations. Default is |
seed |
Seed value for the random number generator. Default value is |
ame
returns an object of class "ame"
. The functions summary
(summary.ame
) and print
(print.summary.ame
) are used to display a summary of the computed average marginal effects.
An object of class "ame"
is a list that contains at least the following components:
|
a named list containing average marginal effects for each continuous variable. If no variable is specified, defaults to |
factor |
a named list containing first differences for each factor variable. If no factor is specified, defaults to |
Furthermore, objects of class "ame"
store the model family and link in the attributes model_family
, model_link
, and the random number generator seed and number of draws from the multivariate normal in the attributes seed
and n_sim
.
Julia Partheymueller, Alexander Staudt (astaudt@mail.uni-mannheim.de)
Krinsky I, Robb AL (1991). "Three Methods for Calculating the Statistical Properties of Elasticities: A Comparison." Empirical Economics 16(2), 199–209.
Brambor T, Clark WR, Golder M (2006). "Understanding Interaction Models: Improving Empirical Analyses." Political Analysis 14(1), 63–82. url: http://localgov.fsu.edu/readings_papers/Research Methods/Brambor_et_al_Multipolicative_Interactions.pdf
King G, Tomz M, Wittenberg J (2000). "Making the Most of Statistical Analyses: Improving Interpretation and Presentation." American Journal of Political Science 44(2), 347–361. url: http://www.polmeth.wustl.edu/files/polmeth/king98f.pdf.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 | # generate some data (factor, dummy, continuous)
set.seed(100)
y <- rbinom(100, size = 1, prob = 0.5)
mcat <- as.factor(rpois(100, 3))
dummy <- as.factor(rbinom(100, size = 1, prob = 0.5))
cont <- runif(100, -10, 10)
data <- data.frame(y, mcat, dummy, cont)
# run glm model
glm_fit <- glm(y ~ mcat + dummy * cont, data = data, family = binomial(link = "logit"))
summary(glm_fit)
# compute average marginal effects
glm_ame <- ame(glm_fit, cont_vars = c("cont"), fac_vars = c("mcat", "dummy"), nsim = 1000)
summary(glm_ame)
# linear model using glm
lm1 <- glm(mpg ~ cyl + hp + wt, data = mtcars, family = gaussian)
lm2 <- glm(mpg ~ cyl * hp + wt, data = mtcars, family = gaussian)
summary(lm1)
summary(lm2)
# compute average marginal effects
lm1_ame <- ame(lm1, cont_vars = c("cyl", "hp", "wt"), nsim = 1000, seed = 99)
lm2_ame <- ame(lm2, cont_vars = c("cyl", "hp", "wt"), nsim = 1000, seed = 99)
# display results
summary(lm1_ame)
summary(lm2_ame)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.