Data.loss.SWD: For given information of loss sampling response of...

In trutscheld/SWDsampling: Sampling data within stepped wedge designs, power calculation and simulations

Description

Data for given numbers of individuals by given group means within a SWD model and derivations with loss of data

Usage

Data.loss.SWD(data.all, I, TP, B = "0", C = 0, D = 0)

Arguments

data.all	Sampled data (response) for given numbers of individuals by given group means within a SWD model
I	number of clusters (design parameter)
TP	number of timepoints (design parameter)
B	timepoint of cluster loss with 4 possibilities: "0": default - no cluster at no timepoint get lost, "1" - Cluster missing at random from timepoint 2 untill TP, "2" - Cluster is missing at beginning (1/3 of timepoints after the first), "3" - Cluster is missing at end (1/3 of the last timepoints).
C	number of cluster loss, by default zero. If a cluster get lost from time point i, all indiviual responses of that cluster will be deleted from timepoint i until timpeoint TP (end).
D	number of individuals loss, by default zero. If not zero, then individual responses to delete are selected at random from timepoints and clusters.

Value

Data frame with individuals intensities corresponds to the SWD model and full model parameter information and derivation information

Examples

noCl<-10
noT<-6
switches<-2
DM<-designMatrix.SWD(noCl,noT,switches)
#cross-sectional SWD (10 cluster and 6 time points)
#no derivation from perfect 100 percent effectiveness pattern
#no data loss (no missing)
data<-SWD.datasampling(I=noCl,TP=noT, mu=0,theta=1,beta.j=rep(1,noT),sigma.alpha=0.5, X.i.j.0=DM,N=10,sigma.e=1)
#no missing in data
Data.loss.SWD(data.all=data, I=noCl,TP=noT)
#missing individuals
Data.loss.SWD(data.all=data, I=noCl,TP=noT, D=5)
#missing 2 cluster at random
Data.loss.SWD(data.all=data, I=noCl,TP=noT, B="1", C=2)

trutscheld/SWDsampling documentation built on June 1, 2019, 1:52 a.m.