Fig. 8
From: The impact of iterative removal of low-information cluster-period cells from a stepped wedge design
Precision loss for incomplete designs with three different percentages of cluster-period cells removed (\(20\%\), \(50\%\), \(80\%\)). Each panel corresponds to a different combination of periods and subjects per cluster-period. They consist of three prespecified correlation parameters: within-period ICC \(\rho\) \(\left( 0.01, 0.05, 0.15\right)\) and cluster autocorrelation r \(\left( 1, 0.95,0.8\right)\). In total there are 36 \(\left( 2 \times 2 \times 3 \times 3\right)\) different trial configurations