Tips for Authors

 

The Basics of Statistics

Dr. Shandra Bipat, CVIR Statistics Editor

 

Statistics are vital to any scientific paper. They are a tool to analyze collected data and the most important aspect is to properly collect and clearly compare the data from the very beginning of the study to the end.

In essence, three types of data can be evaluated:

  • Continuous (e.g. length, age),
  • Binomial (e.g. male/female, factor present/absent), and
  • Categorical, which includes two or more categories:

o   Nominal (e.g., blood group A/B/AB/O)

o   Ordinal (e.g., mild/moderate/severe)

And basically there are two types of comparison:

  • Paired comparison (i.e., outcomes in the same patient population are compared, such as outcomes before and after treatment)
  • Unpaired comparison (i.e., different groups are compared, such as in a RCT).

Authors should be  aware of what the outcomes are, how the data should be collected for the outcomes, and which comparisons are relevant. Based on the above information, the authors of a manuscript should be able to choose their approach for statistical testing.