Skip to main content
Figure 3 | BMC Medical Research Methodology

Figure 3

From: An accurate test for homogeneity of odds ratios based on Cochran’s Q-statistic

Figure 3

Achieved levels for heterogeneous studies, N=150. Comparison of achieved levels, at the nominal level of 0.05, for the three tests Q γ (solid line), BD (dot-dash), and \(Q_{{\chi }^{2}}\) (dash) plotted against the log odds ratio θ for heterogeneous studies. Here the studies have average size 150 divided equally between arms, but the study sizes and the binomial parameters vary for each study. In the left graphs, the smallest control probabilities are paired with the smallest study sizes. In the right graphs, the smallest control probabilities are paired with the largest study sizes.

Back to article page