The best way to see how these table options work is to try them out. I suggest you find the Oswego data set (oswego.rec) and use it.

Say you have 3 variables, a, b, c where c is the outcome (0,1) and a and b are exposures (also coded 0,1).

tables c a b

standard: - gives 2x2 tables for c by a, stratified by b

first by all: - gives 2x2 table for c by a and 2x2 table for c by b

compact (with the /oa option) - use this to get a food-specific attack rate table (in oswego), for example

You will have to explore the options for compact tables to see what works best for you. You can get RR or OR

Jamie

On 2013-05-12, at 10:13 PM, epidata-list@lists.umanitoba.ca wrote:

I see that there are 3 types of table in Epidata analysis: standard, compact, first x all; however, I have not yet distinguished among them: How/when use them, especially compact type. Besides, If I want to know whether data is normal distribute then what command takes on in epidata analysis?