Open Source Statistics for Public Health | Documentation | Testing | About | Help |
Kevin M. Sullivan, Emory University
and Andrew G. Dean, EpiInformatics.com
based on code from John C. Pezzullo
Exact and maximum likelihood statistics adapted from a Pascal program by David Martin.Thanks to Ray Simons for advice and testing.
Interface
Andrew G. Dean and Roger Mir
Two by two tables are used to evaluate the association between a possible risk factor ('Exposure') and an outcome ('Disease'). Counts summarizing the occurence of the four possible combinations of events in the study population are entered into the appropriate cells. The table can be rotated or flipped so that either rows or columns represent Exposure, and the column headings (+) and (-) can be in either order to match common textbooks of epidemiology. A single table or multiple strata can be entered.
Statistics produced include the Fisher and mid-p exact tests, chi squares, odds ratio, maximum likelihood odds ratio estimate, risk/prevalence ratio (relative risk), risk difference, and etiologic fractions with confidence limits produced by several methods, with stratified analysis