mice.impute.hecknorm2step: Imputation by Heckman's model for continuous Missing Not At...

In miceMNAR: Missing not at Random Imputation Models for Multiple Imputation by Chained Equation

Description

Imputes univariate continuous Missing Not At Random (MNAR) outcome using Heckman's model with a two-step estimator.

Usage

mice.impute.hecknorm2step(y, ry, x, JointModelEq, control, ...)

Arguments

y	Incomplete data vector of length n.
ry	Vector of missing data pattern of length n (FALSE=missing, TRUE=observed).
x	Matrix (n x p) of complete covariates.
JointModelEq	Output JointModelEq provided by MNARargument() function.
control	Output control provided by MNARargument() function.
...	Other named arguments.

Details

Imputation of continuous MNAR outcome using two-step Heckman's model (Galimard, 2016). This imputation model requires to define a selection and an outcome equation. The imputation algorithm consists of the following steps:

1. Compute Inverse Mill's Ratio (IMR) from the selection equation probit model

2. Include IMR as predictor in the imputation model

3. Draw imputation parameters using approximate proper imputation for the linear model and adding the Heckman variance correction as detailed in Galimard et al (2016)

4. Draw imputed values from their predictive distribution

Value

A vector of length nmis with imputations.

Note

mice.impute.hecknorm() should be preferred as mice.impute.hecknorm2step(). Accordingly, for continuous outcome, mice.impute.hecknorm() is a default. However, it is possible to modify the method argument by hand.

Author(s)

Jacques-Emmanuel Galimard and Matthieu Resche-Rigon

References

Galimard, J.-E., Chevret, S., Protopopescu, C., and Resche-Rigon, M. (2016) A multiple imputation approach for MNAR mechanisms compatible with Heckman's model. Statistics In Medicine, 35: 2907-2920. doi:10.1002/sim.6902.

See Also

miceMNAR

Examples

require("GJRM")
require(mvtnorm)
require(sampleSelection)

# Generation of a simulated dataset with MNAR mechanism on a continuous outcome

X1 <- rnorm(500,0,1)
X2 <- rbinom(500,1,0.5)
X3 <- rnorm(500,1,0.5)
  
errors <- rmvnorm(500,mean=c(0,0),sigma=matrix(c(1,0.3,0.3,1),nrow=2,byrow=TRUE))

Y <- X1+X2+errors[,1]
Ry <- ifelse(0.66+1*X1-0.5*X2+X3+errors[,2]>0,1,0)

Y[Ry==0] <- NA
  
simul_data <- data.frame(Y,X1,X2,X3)

JointModelEq <- generate_JointModelEq(data=simul_data,varMNAR = "Y")

JointModelEq[,"Y_var_sel"] <- c(0,1,1,1)
JointModelEq[,"Y_var_out"] <- c(0,1,1,0)

arg <- MNARargument(data=simul_data,varMNAR="Y",JointModelEq=JointModelEq)
arg$method["Y"] <- "hecknorm2step"

## Not run: imputation3 <- mice(data = arg$data_mod,
                 method = arg$method,
                 predictorMatrix = arg$predictorMatrix,
                 JointModelEq=arg$JointModelEq,
                 control=arg$control,
                 maxit=1,m=5)

analysis3 <- with(imputation3,lm(Y~X1+X2+X3))
result3 <- pool(analysis3)
summary(result3)
## End(Not run)

miceMNAR documentation built on May 2, 2019, 8:31 a.m.