How to compare and make inference for odds?

[Tom Hsiung]

Similar as to compare and make inference for risks by

prtest

, is there any command to do the same thing for odds?

Tom

[Carlo Lazzaro]

Tom:
see -help epitab-.

[Tom Hsiung]

Thanks, Carlo. Powerful statistic tests for epidemiology. Do you know the logic behind these tests?

• ir - tests for incidence rates
• cs - tests for risks, same as proportional z-test (prtest)?
• cc - tests for odd ratio
• tabodds - tests for odds
• mhodds - tests for odds

Thanks.

[William Lisowski]

Have you tried reviewing the full documentation for epitab, and for the individual commands, found in the Stata Base Reference Manual PDF included with your Stata installation and accessible through Stata's Help menu?

[Tom Hsiung]

Oh! My mistake. How should have I forgotten such very important logics behind the distribution of odd ratio and risk ratio.

prtest

use risk1 - risk2 - 0 as the basic component of the proportional z-test

For odd ratio, the logOR is used as the basic component of the test statistics. The distribution of logOR and logRR are normal, and the standard errors of logOR and logRR also are able to estimate from the sample.

Unfortunately, the introductory statistics textbook by Neil Weiss did not cover the topic about logOR and logRR.

Thanks.