Policies for Manuscript Writing

Authorship

A frequently asked question is whether biostatistical consultants should be co-authors on scientific papers. We feel that decisions about authorship should be independent of consideration of funding sources. As recommended in published guidelines (Parker RA, Berman NG: Criteria for authorship for statisticians in medical papers. Statistics in Medicine 17: 2289-2299 (1998)), "The basis of financial support should be the time/effort spent on a project and the basis for authorship should be whether the statistician has made a scientific contribution to the project." Examples of scientific contributions are the following.

1. The statistician has to develop new statistical methods to meet the project's needs, or she/he has to combine existing techniques in a novel way.
2. The statistician has a major role in designing the study.
3. The statistician writes part of the manuscript other than a standard paragraph or two describing which statistical methods were used.
4. The statistician is asked to critique an initial draft and the statistician spends a considerable amount of time suggesting alternative wording and presentation of results.
5. The statistician provides data analysis along with interpretation of results.

By JAMA's criteria for authorship for statistical experts involved in the analysis and interpretation of data used in a manuscript; a statistician is a co-author if (1) he/she took part in the drafting of the manuscript or (2) he/she was involved in a critical revision of the manuscript for important intellectual content.

It is our policy that MS and PhD biostatisticians may not have their names listed under acknowledgements.
If assistance was given through the Biostatistics Clinic, we appreciate acknowledgements of the clinic itself, e.g. "We would like to acknowledge the Biostatistics Clinic, Vanderbilt School of Medicine Dept. of Biostatistics, for guidance in statistical methods."

Above all, it is important for the medical researcher and biostatistician(s) assigned to the project to agree on criteria for authorship early in their collaboration.

Manuscript Preparation

Whether or not the statistician is an author, it is important to allow sufficient time for the statistician to check statistical results and descriptions of statistical methods that appear in a manuscript. We frequently find inconsistencies between analyses we perform and basic descriptive statistics computed by the investigator. For example, the primary statistical comparison may emphasize differences in medians while the investigator quotes mean values elsewhere in the manuscript. When different analyses are carried out by different personnel, it is beneficial for the investigator and statistician to map out the entire analysis together in advance.