(19.1)
(A) You need a reliable estimate of the variable's standard deviation. This may come from a prior study,
established norm, or "clinical Gestalt."
(19.2) n = (1.962)(.1)(.9) / (.0252) = 553.2, or about 553 people.
(19.3) The per group sample size n = (16)(202) / 52 + 1 = 257.
(19.4) Number of cases = 400. Number of controls = 800.
(19.5) Number of cases = 128. Number of controls = 256.
(19.6) According to Dallal's table: 219 in each group. According to EpiCalc 2000: 198 in each group.