Study on the status quo and quality evaluation of data graphs and tables based on academic medical journals in Shanghai universities

Objective To evaluate the quantity and quality of graphs and tables in medical literature. Methods We conducted a retrospective survey of 750 randomly selected original research articles, from 2011-2015 issues of 5 academic medical journals by using a stratified random sampling method. Paper characteristics, application status, quantity and quality of graphs and tables were analyzed. Results More than 48% of original research articles contained at least 1 table or graph. The highest frequency of using table and graph was in clinical research. The number of tables and the number of articles with tables were both significantly decreased during 2011-2015 (P=0.034). The most common errors in tables were the inconsistent decimal number and the presentation of numeric data without right alignment. Only 12% of the articles contained at least 1 graph. Of the 175 graphs, bar chart with intervals predominated. Compared with other domains, the completeness of graphs was excellence. Conclusions The result expression in medical research is still dominated by tables. And the quality of graphs is much better than tables. Key words: Data graphs and tables; Journal article; Quality assessment