In Reply

Descriptive Statistics: The Specification of Statistical Measures and Their Presentation in Tables and Graphs Part 7 of a Series on Evaluation of Scientific Publications by Albert Spriestersbach, Dr. rer. nat. Bernd Röhrig, Dr. med. Jean-Baptist du Prel, Aslihan Gerhold-Ay, Prof. Dr. rer. nat. Maria Blettner in volume 36/2009

We wish to thank Dr. Goralczyk for his additional comments, with which we fully agree. Although the Tukey algorithm is standard, there is a variety of other conventions. This is not unreasonable. The presentation suggested by Altman is particularly suitable for large case numbers with hundreds of observations, while the Tukey algorithm is intended for smaller numbers of cases, which are nevertheless too large to present the individual points (>20 to a few hundred). Within this range, Tukey box plots show no, or very few, outliers.

It is quite clear that the legend to a box plot must be read carefully and that it must mention the convention used.
DOI: 10.3238/arztebl.2010.0122b

Prof. Dr. rer. nat. Maria Blettner
Institut für Medizinische Biometrie
Epidemiologie und Informatik
Johannes Gutenberg-Universität
55101 Mainz, Germany
[e-mail]

Conflict of interest statement
The authors of both contributions declare that no conflict of interest exists according to the guidelines of the International Committee of Medical Journal Editors.

References

1. Spriestersbach A, Röhrig B, du Prel JB, Gerhold-Ay A, Blettner M: Descriptive statistics: The specification of statistical measures and their presentation in tables and graphs—Part 7 of a series on evaluation of scientific publications [Deskriptive Statistik: Angabe statistischer Maßzahlen und ihre Darstellung in Tabellen und Grafiken – Teil 7 der Serie Biometrie]. Dtsch Artzebl Int 2009; 106(36): 578–83.