Descriptive statistics: the specification of statistical measures and their presentation in tables and graphs. Part 7 of a series on evaluation of scientific publications.

BACKGROUND: Descriptive statistics are an essential part of biometric analysis and a prerequisite for the understanding of further statistical evaluations, including the drawing of inferences. When data are well presented, it is usually obvious whether the author has collected and evaluated them correctly and in keeping with accepted practice in the field.

METHODS: Statistical variables in medicine may be of either the metric (continuous, quantitative) or categorical (nominal, ordinal) type. Easily understandable examples are given. Basic techniques for the statistical description of collected data are presented and illustrated with examples.

RESULTS: The goal of a scientific study must always be clearly defined. The definition of the target value or clinical endpoint determines the level of measurement of the variables in question. Nearly all variables, whatever their level of measurement, can be usefully presented graphically and numerically. The level of measurement determines what types of diagrams and statistical values are appropriate. There are also different ways of presenting combinations of two independent variables graphically and numerically.

CONCLUSIONS: The description of collected data is indispensable. If the data are of good quality, valid and important conclusions can already be drawn when they are properly described. Furthermore, data description provides a basis for inferential statistics.