Skip to main content
Fig. 4 | BMC Medical Research Methodology

Fig. 4

From: The impact of iterative removal of low-information cluster-period cells from a stepped wedge design

Fig. 4

An illustrative example of the changing contributions of different cluster-period cells to estimation of the treatment effect in stepped wedge designs with four clusters (rows in each grid) and five periods (columns), with 90 subjects measured in each cluster-period, under a discrete-time decay correlation structure with \(\rho =0.15\) and \(r=0.95\). The information content of each cell pair is displayed in each cell, with cell shading also representing the information content, ranging from light yellow (small values) to dark red (large values). The complete stepped wedge design is shown in the top left and the designs become progressively reduced as cluster-period cells are removed from the design, moving from left to right along each row, top row to bottom row

Back to article page