predict.listExperiment: Predict method for the list experiment

In misreport: Statistical Analysis of Misreporting on Sensitive Survey Questions

Description

Obtains predictions from a fitted list experiment model of the class listExperiment.

Usage

## S3 method for class 'listExperiment'
predict(object, newdata = NULL,
  treatment.misreport = 0, par.control = NULL, par.sensitive = NULL,
  par.misreport = NULL, ...)

Arguments

object	Object of class "listExperiment"
newdata	An optional data frame from which to calculate predictions.
treatment.misreport	Value of the treatment variable covariate in the misreport sub-model (if included in the model). 0 treatment indicator in the misreport sub-model is set to 0 for all individuals (default). 1 treatment indicator in the misreport sub-model is set to 1 for all individuals. "observed" treatment indicator in the misreport sub-model is set to the observed treatment value.
par.control	An optional set of control-items sub-model parameters to use in place of those from the fitted model.
par.sensitive	An optional set of sensitive-item sub-model parameters to use in place of those from the fitted model.
par.misreport	An optional set of misreport sub-model parameters to use in place of those from the fitted model.
...	Additional arguments

Details

If newdata is omitted, predictions will be made with the data used for model fitting.

Slots

z.hat

Predicted probability of answering affirmatively to the sensitive item in the list experiment.

u.hat

Predicted probability of misreporting (assuming respondent holds the sensitive belief).

References

Eady, Gregory. 2017 "The Statistical Analysis of Misreporting on Sensitive Survey Questions."

Examples

data(gender)

## Not run: 
# Note: substantial computation time
model.gender <- listExperiment(y ~ gender + ageGroup + education +
                                       motherTongue + region + selfPlacement,
                                   data = gender, J = 4,
                                   treatment = "treatment", direct = "direct",
                                   control.constraint = "none",
                                   sensitive.response = 0,
                                   misreport.treatment = TRUE)
predict(model.gender, treatment.misreport = 0)

## End(Not run)

misreport documentation built on May 2, 2019, 11:27 a.m.