GJRM.data (version 0.1-1)

dataDSS: Simulated data with double sample selection

Description

Simulated data with double sample selection and binary outcome.

Usage

data(dataDSS)

Format

dataDSS is a 10000 row data frame with the following columns:

y1

First selection.

y2

Second selection.

y3

Binary outcome.

x1, x2

Covariates.

x3

Covariate influencing only y1.

x4

Covariate influencing only y2.

y3.o

Original outcome, without missingness.

Examples

# Data have been simulated as shown below

n  <- 10000
x1 <- round(runif(n))
x2 <- runif(n)
x3 <- runif(n)
x4 <- rnorm(n)
u  <- rnorm(n)

y1 <- ifelse(-1.55 +     x1 - x2 + x3 + u + rnorm(n) > 0, 1, 0)
y2 <- ifelse(-0.25 - 0.5*x1 + x2 + x4 + u + rnorm(n) > 0, 1, 0)
y3 <- y3.o <- ifelse( -0.75 + x1 + x2 + u + rnorm(n) > 0, 1, 0)

y2 <- y2*y1
y3 <- y3*y2
y3 <- ifelse(y2 == 0, NA, y3)

dataDSS <- data.frame(y1, y2, y3, x1, x2, x3, x4, y3.o)