Statistical presentation of data in biomedical publications

Scientists rely upon statistics to report their findings accurately and to interpret the published findings of others. The proper application of this discipline provides confidence in statements based upon experimental results. Professional statisticians may help at some phases of planning experiments even before results are obtained, but not all scientists avail themselves of such services. As a result, errors can occur in presentation of variability of data, and more often than not, such errors are ultimately published in biomedical journals. This paper examines the most common descriptive statistics for quantitative and categorical data. Standard deviation (SD), standard error of the mean (SEM), confidence intervals (CI), and various technical details, including how to present data in publications, and when to use particular statistical tools, are discussed as well.