caceSRTBoot: CACE Analysis of Simple Randomised Trials.

In eefMLM: Analytical Package for Analysing Education Trials

Description

caceSRTBoot performs is used for CACE analysis of simple randomised trials. Intervention variable must be coded as dummy with multiple analysis for multi-arms trials.

Usage

caceSRTBoot(formula, intervention, compliance, nBoot, data)

Arguments

formula	model specification of the form posttest ~ pretests+Intervention+....the model to be analysed. It is of the form y~x1+x2+..., where y is the outcome variable and X's are the predictors.
intervention	the name of the intervention variable as appeared in formula. This must be put in quotes. For example "intervention" or "treatment" or "group".
compliance	percentages of sessions attended by pupils.
nBoot	number of bootstrap required to generate bootstrap confidence interval. This must be specified.
data	data frame containing the data to be analysed.

Value

S3 mcpi object; a list consisting of

• CACE. Estimated CACE effect size based on percentages of sessions attended by pupils. The percentage data is converted into the following grids (0, 10, 20, 30, 40, 50, 60, 70, 80, 90) and CACE effect size is calculated for each grid.

• Compliers. A summary table of the percentage of pupils in the intervention and control groups that attended more than a pre-specified percentage of sessions. The values for the control group should be zeros if there is no dilution in which a pupil or school in the control group receives intervention.

Examples

data(catcht)

############# SRT
caceOutput3<- caceSRTBoot(Posttest~ Prettest+ Intervention,
			intervention="Intervention",
			compliance = "Percentage_Attendance",
			nBoot=1000,data=catcht)

### visualising CACE ES

require(metafor)
forest(x=caceOutput3$CACE$ES, ci.lb=caceOutput3$CACE$LB, 
		ci.ub=caceOutput3$CACE$UB, xlab="CACE ES",
		slab=as.character(caceOutput3$CACE$Compliance), 
		xlim=c(-1,2), alim=c(-1,7), cex=1,lwd=1.5)

eefMLM documentation built on May 2, 2019, 5:46 p.m.