Science Education Portraits VII Statistical Methods Used in 1081 Papers Published in Year 2020 Across 12 Life Science Journals Under BioMed Central. - mauriceling/mauriceling.github.io GitHub Wiki
Citation: Kim, KD, Chua, SCH, Ling, MHT. 2021. Science/Education Portraits VII: Statistical Methods Used in 1081 Papers Published in Year 2020 Across 12 Life Science Journals Under BioMed Central. Acta Scientific Nutritional Health 5(3): 06-12.
Here is the permanent [PDF] and [data set] links to my archive.
Statistics is an integral part of biology and is required for all undergraduate life science curriculum. However, are biology students trained in statistical skills required in the field? Despite studies listing various commonly statistical methods used in specialised branches of life sciences; such as, immunology and tropical biology; there is a lack of study on the common statistical methods used in life science in general. Here, we examine 1081 articles across 12 life sciences journals under BioMed Central, published in 2020, to elucidate the common statistical methods used in current life science research, as a basis to recommend an updated syllabus to all institutions that educate biologists. 72.7% of the examined articles contains identifiable statistical methods and a total of 2431 instances were identified. Our findings show that the first 3 out of 15 categories of methods; parametric comparison of means (25.38% of instances), correlation/regression (18.88%), and post-hoc test (10.32%); accounts for 54.59% of the instances. In terms of individual methods, the top 8 methods account for 52.04% of the instances – (a) t-test (13.00%), (b) ANOVA (12.26%), (c) unspecified (likely to be Pearson’s correlation) and Pearson’s correlation (9.79%), (d) Benjamini and Hochberg’s False Discovery Rate (FDR) (4.77%), (e) Tukey's HSD (4.36%), (f) Kruskal-Wallis Test (2.96%), (g) Mann-Whitney U Test (2.80%), and (h) Chi Square Test (2.10%). These findings may have an impact on future curriculum design.