Now Playing: Analysis of Covariance (ANCOVA) (2:08)
Suppose researchers are designing a study to compare a control mean to two treatment means, while accounting for the covariates: blood pressure and weight.
The control mean is assumed to be 47, and treatment means of 53 and 55 will be presumed.
For the within group standard deviation, standard deviations between 10 and 20 will be compared.
Because the exact value of R-Squared (of the response with the 2 covariates) is unknown, a range of values between 0.1 and 0.4 will be considered.
The researchers wish to know the sample sizes need for each group if the control group size is to be twice as large as the treatment group sizes. A power of 90%, and an alpha of 0.05, are to be used.
The first table of the output gives a numeric comparison of the average and total sample sizes.
The ‘details’ section shows the sample sizes and means for each line of the first table.
The chart at the bottom shows a comparison of the average sample sizes, graphically.