expData-methods: Methods for expanded datasets

In jmpsteen/medflex: Flexible Mediation Analysis Using Natural Effect Models

Methods for expanded datasets

Description

Regression weights, residuals and residual plots for expanded datasets.

Usage

## S3 method for class 'expData'
residuals(object, ...)

## S3 method for class 'expData'
residualPlot(model, ...)

## S3 method for class 'expData'
residualPlots(model, ...)

## S3 method for class 'expData'
weights(object, ...)

Arguments

object	an expanded dataset (of class "expData").
...	additional arguments.
model	an expanded dataset (of class "expData") (for use with residualPlot and residualPlots).

Details

weights extracts regression weights (to be used in the natural effect model) for each observation of an expanded dataset.

residuals extracts residuals from the working model which is stored as an attribute of the expanded dataset. These can be used to assess normality of the residuals of the mediator working model when using the weighting-based approach (see example).

residualPlot and residualPlots are convenience functions from the car package. These can be used to assess the adequacy of the working model.

See Also

expData, neWeight, residualPlot, residualPlots, residuals, weights

Examples

data(UPBdata)

weightData <- neWeight(negaff ~ att + gender + educ + age, 
                       data = UPBdata, nRep = 2)

## extract regression weights for natural effect model
head(weights(weightData)) 

## assess normality
qqnorm(residuals(weightData))

## assess model adequacy
library(car)
residualPlots(weightData)

jmpsteen/medflex documentation built on July 6, 2023, 8 p.m.