2 × 2 Tables

2 × 2 tables are used to organize data for two proportions. From 2 × 2 tables2x2 table, we compare proportions as differences or as ratios. We can also compare groups using an odds ratioOdds ratio. An odds ratio is close in value to the risk ratio when the disease is rare. Odds ratios have the advantage of not being influenced by whether exposure or disease was sampled. This makes it ideal for case–control studies. We can test hypotheses about proportionProportion differences or ratios or odds ratios in 2 × 2 tables using Pearson’s chi-squarePearson’s chi-square. Test-based interval estimates can be calculated using the result of hypothesis testing. Sometimes data are paired. Paired data result from having the same individuals or very similar individuals have both exposure categories. Then data are organized in a paired 2 × 2 table2x2 table and proportions or odds are compared in that table. Hypotheses are tested using McNemar’s chi-squareMcNemar’s chi-square. When weChi-square are comparing two incidences, data are organized in a different table. In this table we record events and person-timePerson-time. Person-time is how long a person is at risk for having the event. Most often we use the actuarial method to determine time spent between when a person was known not to have the event and when they are known to have the event. This method represents that period with half its length. Incidences are compared as differences or ratios. Hypotheses are tested using a chi-squareChi-square test.