Multinomial Regression with Survey Data

Source: R/svymultinom.R

svymultinom uses the withReplicates function to compute the replicate-based estimate of the variance-covariance matrix of coefficients for a multinomial regression fitted by multinom.

svymultinom(formula = NULL, weights = NULL, data = NULL)

# S3 method for svymultinom
coef(object, ...)

# S3 method for svymultinom
vcov(object, ...)

# S3 method for svymultinom
formula(x, ...)

# S3 method for svymultinom
predict(object, ...)

# S3 method for svymultinom
model.frame(formula, ...)

# S3 method for svymultinom
print(x, ...)

# S3 method for svymultinom
summary(object, ...)

# S3 method for summary.svymultinom
print(x, digits = 4, ...)

# S3 method for svymultinom
update(object, ..., evaluate = TRUE)

Arguments

formula

regression formula
weights

weights for regression
data

dataset for regression
object

an object of class 'svymultinom'
...

additional arguments
x

an object of class 'svymultinom'
digits

minimal number of significant digits. See print.default.
evaluate

a logical value. If TRUE, the updated call is evaluated. Default is TRUE.

Value

An object of class svymultinom is returned:

call

the function call,

NAIVEreg

the naive multinomial regression object,

vcov

the replicate-based estimate of the variance-covariance matrix of coefficients,

...

Methods (by generic)

  • coef: Extract coefficients

  • vcov: Extract the var-cov matrix of coefficients

  • formula: Extract the regression formula

  • predict: Predict with new data

  • model.frame: Extract the model frame

  • print: Print results of svymultinom nicely

  • summary: Summarize results of svymultinom nicely

  • print: Print summary of svymultinom nicely

  • update: Update svymultinom

See also

multinom, withReplicates

Examples


if (FALSE) {
rm(list=ls())
library(CMAverse)

# multinom
n <- 1000
x1 <- rnorm(n, mean = 0, sd = 1)
x2 <- rnorm(n, mean = 1, sd = 1)
x3 <- rbinom(n, size = 1, prob = 0.4)
linearpred1 <- 1 + 0.3 * x1 - 0.5 * x2 - 0.2 * x3
linearpred2 <- 2 + 1 * x1 - 2 * x2 - 1 * x3
py2 <- exp(linearpred1) / (1 + exp(linearpred1) + exp(linearpred2))
py3 <- exp(linearpred2) / (1 + exp(linearpred1) + exp(linearpred2))
py1 <- 1 - py2 - py3
y <- sapply(1:n, function(x) sample(size = 1, c(1:3), prob = c(py1[x], py2[x], py3[x])))
w <- ifelse(x3 == 0, 0.4, 0.6)
data <- data.frame(x1 = x1, x2 = x2, x3 = x3, y = y)

reg <- svymultinom(y ~ x1 + x2 + x3, weights = w, data = data)
coef(reg)
vcov(reg)
formula(reg)
predict(reg, newdata = data[1, ], type = "probs")
model.frame(reg)
summary(reg)
update(reg, weights = w[1:500], data = data[1:500, ])
}