Hope you all are well.
I have a question regarding the following hypothetical example codes where I am comparing Sampsi and Samclus with the new updated Power Command in Stata 16 where I am determining number of observations per cluster:
Code:
sampsi `contmean' `tmean', sd(`contsd') power(0.8) * Estimated sample size for two-sample comparison of means * Test Ho: m1 = m2, where m1 is the mean in population 1 * and m2 is the mean in population 2 * Assumptions: * alpha = 0.0500 (two-sided) * power = 0.8000 * m1 = .170622 * m2 = .616529 * sd1 = 1.11477 * sd2 = 1.11477 * n2/n1 = 1.00 * Estimated required sample sizes: * n1 = 99 * n2 = 99 sampclus, numclus(36) rho(`rho') * Sample Size Adjusted for Cluster Design * n1 (uncorrected) = 99 * n2 (uncorrected) = 99 * Intraclass correlation = .138448 * Average obs. per cluster = 20 * Minimum number of clusters = 36 * Estimated sample size per group: * n1 (corrected) = 360 * n2 (corrected) = 360
Code:
power twomeans `contmean' `tmean', k1(36) k2(36) sd(`contsd') rho(`rho') cluster * Performing iteration ... * Estimated cluster sizes for a two-sample means test * Cluster randomized design, z test assuming sd1 = sd2 = sd * Ho: m2 = m1 versus Ha: m2 != m1 * Study parameters: * alpha = 0.0500 * power = 0.8000 * delta = 0.4459 * m1 = 0.1706 * m2 = 0.6165 * sd = 1.1148 * Cluster design: * K1 = 36 * K2 = 36 * rho = 0.1384 * Estimated cluster sizes and sample sizes: * M1 = 4 * M2 = 4 * N1 = 144 * N2 = 144
0 Response to Sampclus vs Power (cluster) for observations per cluster calculation
Post a Comment