SteppedPower (version 0.3.4)

construct_DesMat: Construct the Design Matrix

Description

Constructs the design matrix with one column for every (fixed) parameter to be estimated and one row for every cluster for every timepoint. This function calls `construct_trtMat` to construct a matrix that indicates treatment status for each cluster at each timepoint. This is then transformed into the first column of the design matrix. `construct_CovMat` further calls `construct_timeAdjust` to get the fixed effect(s) of the timepoints.

Note: Unlike the usual notation, the treatment effect is in the first column (for easier access by higher level functions).

Usage

construct_DesMat(
  Cl = NULL,
  trtDelay = NULL,
  dsntype = "SWD",
  timepoints = NULL,
  timeAdjust = "factor",
  period = NULL,
  trtmatrix = NULL,
  timeBlk = NULL,
  N = NULL,
  incomplete = NULL,
  INDIV_LVL = FALSE
)

Value

an object of class DesMat

Arguments

Cl

integer (vector), number of clusters per sequence group (in SWD), or number in control and intervention (in parallel designs)

trtDelay

numeric (possibly vector), `NA`(s) and/or value(s) between `0` and `1`. `NA` means that first (second, ... ) period after intervention start is not observed. A value between `0` and `1` specifies the assumed proportion of intervention effect in the first (second ... ) intervention period.

dsntype

character, defines the type of design. Options are "SWD", "parallel" and "parallel_baseline", defaults to "SWD".

timepoints

numeric (scalar or vector), number of timepoints (periods). If design is swd, timepoints defaults to length(Cl)+1. Defaults to 1 for parallel designs.

timeAdjust

character, specifies adjustment for time periods. One of the following: "factor", "linear", "none", "periodic". Defaults to "factor".

period

numeric (scalar)

trtmatrix

an optional user defined matrix to define treatment allocation

timeBlk

an optional user defined matrix that defines the time adjustment in one cluster. Is repeated for every cluster.

N

numeric, number of individuals per cluster. Either a scalar, vector of length #Clusters or a matrix of dimension #Clusters x timepoints. Defaults to 1 if not passed.

incomplete

integer, either a scalar (only for SWD) or a matrix. A vector defines the number of periods before and after the switch from control to intervention that are observed. A matrix consists of `1`s for observed clusterperiods and `0`s or `NA` for unobserved clusterperiods.

INDIV_LVL

logical, should the computation be conducted on an individual level? This leads to longer run time and is mainly for diagnostic purposes.

Examples

construct_DesMat(Cl=c(2,0,1))
construct_DesMat(Cl=c(2,0,1), N=c(1,3,2))

## manually defined time adjustment (same as above)
timeBlock <- matrix(c(1,0,0,0,
                      1,1,0,0,
                      1,0,1,0,
                      1,0,0,1), 4, byrow=TRUE)
construct_DesMat(Cl=c(2,0,1), timeBlk=timeBlock)