|
All epidemiologic studies benefit from sample size and power
calculations in the pilot phase of the research. This graph
shows that the sample size needed for a cohort study depends
upon, among other things, the percent of the cohort having the
outcome of interest. Injury studies with people who do heavy
manual materials handling may need a smaller number of subjects
than a study of office workers. Where the outcome is frequent,
such as 30% of an elderly cohort falling over a six-month period,
cohort studies become easier and less expensive to conduct.
|