Skip to main content

Table 3 Sample sizes and effect sizes in terms of a difference of proportions between two groups given a statistical power of 80%

From: Assessing the effect of an educational intervention on early childhood development among Mexican preschool children in the state of Oaxaca: a study protocol of a cluster randomized stepped-wedge trial

Sample size per group Proportion in the group without exposure to NPECDP-UKA Effect size (difference of proportions) given different design effects (DEFF)*
DEFF = 1.0 DEFF = 1.5 DEFF = 2.0
750 0.50 0.072 0.088 0.102
900 0.50 0.066 0.081 0.093
1200 0.50 0.057 0.070 0.081
750 0.25 0.065 0.080 0.093
900 0.25 0.059 0.073 0.085
1200 0.25 0.051 0.063 0.073
750 0.125 0.052 0.064 0.075
900 0.125 0.047 0.058 0.068
1200 0.125 0.040 0.050 0.058
  1. *The design effect is a factor that when multiplied by the variance of an estimator under a simple random sampling design corresponds to its variance under a complex sampling design; in this case, a design with clustering of observations within communities