predict function

Outcome Prediction

The method computes the predicted outcome for each group with standard errors and confidence intervals.

## S3 method for class 'cslseFit'
predict(object, interval=c("none","confidence"),
                       se.fit=FALSE, newdata=NULL, level=0.95,
                       vcov.=vcovHC, ...)
## S3 method for class 'slseFit'
predict(object, interval=c("none","confidence"),
                       se.fit=FALSE, newdata=NULL, level=0.95,
                       vcov.=vcovHC, ...)

Arguments

• object: Object of class cslseFit or slseFit

  created by estSLSE.

• interval: If set to "confidence", it returns the predicted values along with the lower and upper bounds of the confidence interval.

• se.fit: Should the function return the standard errors of the predicted values?

• level: The confidence interval level if interval is set to "confidence".

• newdata: A data.frame of new data. It must include values for all covariates, and for the treatment indicator in the case of cslseFit objects.

• vcov.: An alternative function to compute the covariance matrix of the least squares estimates. The default is the vcovHC.

• ...: Additional argument to pass to the vcov. function.

Returns

For slseFit objects, it returns the predicted outcome if se.fit is FALSE or a list of the following two elements otherwise:

• fit: The predicted outcome.

• se.fit: The standard errors of the predicted outcomes.

If the argument confidence is set to "interval", the predicted outcome is a matrix with the predicted outcome, and the lower and upper bounds of the confidence intervals.

For objects of class 'cslseFit', the same is returned for each treatment group in a list. The elements of the list are treated

and nontreated (until the package allows for more than one treatment).

Examples

data(simDat3)
mod <- cslseModel(Y ~ Z | ~ X1 + X2, data = simDat3)
fit <- causalSLSE(mod)

## Predicting outcome for all observations

pr <- predict(fit, interval = "confidence")

## Predicting outcome with new data

ndat <- data.frame(X1 = c(-2, 1, 2, 3), X2 = c(-4, -2, 0, 1),
                   Z = c(1, 1, 0, 0))
predict(fit, newdata = ndat)