Fig. 4

Represents how the groupings are done and are based on similarity between studies. Studies that are more ‘similar’ to each other will appear jointed earlier. The Y-axis (height) is based on the inverse of the ‘similarity’ with studies (or groups of studies) joining up at greater heights representing studies (or group of studies) that are less similar than those that join up at lower heights. The overall spread of the studies with roughly equal numbers in groups appear to show a lack of clustering structure (at least in relation to YES/NO/MAYBE) and is consistent with the silhouette graph