Skip to main content
. 2011 Nov 30;13(2):274–288. doi: 10.1093/biostatistics/kxr044

Fig. 2.

Fig. 2.

Estimated bounds for power to detect θx = 1.3, based on a case–control design, as a function of case–control sample size n for various scenarios for confounding. Estimates were obtained using the algorithm in the supplementary material (available at Biostatistics online) with R = 10000.