Skip to main content
. 2001 Jan 23;98(3):831–836. doi: 10.1073/pnas.021529998

Figure 2.

Figure 2

Scatter plots of the relative change in the pooled odds ratio as a function of the cumulative number of patients in previously published randomized trials. An outlier relative change of 0.05 for sample size = 50 is not shown in the myocardial infarction graph. A relative change greater than 1 means that the odds ratio tends to increase with the accumulation of more data. Fixed effects graphs were similar.