ABSTRACT

Trial design and data collection methods determine the appropriate statistical tests for analysing 2×2 tables. Data-examples are given to illustrate the use of the classical chi-square test and Fisher’s exact probability test in assessing for any association between two factors of the data or for testing differences in proportions ‘responding’ in two unrelated groups. In medical trials where two different treatments are being compared, patients may be ‘matched’ and subject to one treatment only or each patient may be exposed to both treatments. Statistical tests such as McNemar’s and Cart’s may be applicable and data examples are provided to demonstrate the relatively simple computations involved.