Skip to main content
Fig. 1 | BMC Medical Research Methodology

Fig. 1

From: Prevalence of evidence of inconsistency and its association with network structural characteristics in 201 published networks of interventions

Fig. 1

Plot of p-values (fourth-root scale) of the DBT model against ratios of network structural characteristics (logarithmic scale). a Explores the ratio of the number of studies to the number of interventions in a network (correlation coefficient [p-value] = − 0.07 [0.31]); b Explores the ratio of the number of loops to the number of interventions in a network (correlation coefficient [p-value] = 0.02 [0.72]); c Explores the ratio of the number of loops to the number of studies in a network (correlation coefficient [p-value] = 0.08 [0.28]); d Explores the ratio of the number of unique designs to the number of studies in a network (correlation coefficient [p-value] = 0.07 [0.28]); e Explores the ratio of the number of unique direct intervention comparisons to the number of studies in a network (correlation coefficient [p-value] = 0.06 [0.40]); f Explores the ratio of the number of multi-arm studies to the number of total studies in a network (correlation coefficient [p-value] = 0.09(0.19)). All analyses have used the REML estimator for heterogeneity. The horizontal green and purple lines represent the cut-off p-value = 0.05 and p-value = 0.10, respectively. The blue diagonal line is the regression line. * Networks with no multi-arm studies were treated as networks including a single multi-arm study, to avoid excluding them from the plot, since the logarithm could not be calculated. Abbreviations: DBT, design-by-treatment interaction model; REML, restricted maximum likelihood

Back to article page