Figure 1

Accuracy of study characteristics reported. The first column gives values for “A’”, the claims made in the study, while the following five columns provide the number of articles that made “B”, alternative claims that describe the same study characteristic, in each of five Euler classes (representing varying degrees of accuracy). Absolute numbers represent how many different articles had at least one B description in each sub-category. Each article was counted only once per sub-category (e.g., non-demented, placebo controlled) of study characteristic even if it made two or more separate claims that fell into that sub-category. This metric is captured by the term “article-claim”. Thus, the numbers in each cell refers to the number of “article-claims” made per sub-category added together for each category in each class. To calculate the total number of article-claims made per class, the numbers of article-claims per category were added together. Theoretically, a single article could account for 15 category claims since they could have made a claim about each sub-category of study characteristics. To calculate the total number of article-claims made per category, the numbers of article-claims per class were added together. Here, a single article could account for a maximum of 5 times the number of sub-category claims that is possible in each category, since they could have made a claim about each sub-category that fell into each class.