margins.dat.clogit {catregs}

Computes predicted probabilities for conditional and rank-order/exploded logistic regression models. Inference is based upon simulation techniques (requires the MASS package). Alternatively, bootstrapping is an option for conditional logistic regression models.

Description

Given a model object and a design matrix, this creates a data.frame of the design matrix, with predicted probabilities for each response category. Inferential information about the predict probabilities is supported with simulation. Bootstrapping may be added as an option for conditional logistic regression models.

Usage

margins.dat.clogit(mod,design.matrix,run.boot="no",num.sample=1000,
prop.sample=.9,alpha=.05,seed=1234,rounded=3)

Arguments

mod

A conditional logistic regression model as estimated in the Epi package or an Exploded logistic regression model as estimated in the mlogit package.

design.matrix

Design matrix of values for the independent variables in the regression model. Unlike the design matrices in the margins.des function, the design matrix for a conditional logistic regression entails multiple rows, corresponding to the number of response options.

run.boot

Whether to compute confidence intervals around the predicted probabilities using bootstrapping. Defaul is "no."

num.sample

num.sample is the number samples drawn to compute the sampling distibution.

prop.sample

prop.sample is the proportion of the original sample to include in the sampling distibution samples. Default is .9

alpha

The alpha value for confidence intervals. Default is .05.

seed

Sets a seed so that random results are reprodicible.

rounded

How many decimal places to show in the output.

Value

des

Returns a data.frame containing the design matrix and additional columns for the predicted probabilities.

boot.dist

The full bootstrapped distribution for the probabilities.

Author(s)

David Melamed

Examples

data("LF06travel")
m1 <- Epi::clogistic(choice ~ train + bus +
time  + invc, strata=id, data=LF06travel)
design <- data.frame(train=c(0,0,1),bus=c(0,1,0),time=200,invc=20)
design
margins.dat.clogit(m1,design)