Power and Sample Size Requirements, Cohort Studies
Use either StatCalc* or EpiTable* to perform calculations.
The determinants of power and sample size are:
- the [allocation] ratio of unexposed to exposed subjects (n2 / n1)
- the expected relative risk (a RR "worth detecting")
- the expected incidence in unexposed subjects (p2), and
- the alpha (confidence) level of the test.
Of course, power and sample size are inversely related.
* These computer program uses formulas from Fleiss (1981, pp. 44 - 45).