Two Postmodernisms in Rheumatology.

Area of Science:

• Biostatistics
• Medical Research Methodology

Background:

• Randomization is a cornerstone of robust data collection and interpretation in scientific studies.
• Misapplication of statistical methods, including randomization and p-value interpretation, can lead to erroneous conclusions.

Purpose of the Study:

• To identify and illustrate common misuses of randomization and p-values in scientific research.
• To critically examine the practice of using subgroups for validation after initial randomization.
• To introduce and define the concept of the "postmodern p-value".

Main Methods:

• Conceptual analysis of statistical principles.
• Illustrative examples from clinical research, specifically Behçet's syndrome and gout criteria.
• Discussion of common statistical pitfalls and their implications.

Main Results:

• Subgroup validation after whole-dataset randomization incorrectly validates the randomization process itself, not the outcome.
• P-values are frequently misused to force significance or, conversely, to imply no difference using values close to unity (postmodern p-value).

Conclusions:

• Researchers must exercise caution to avoid misinterpreting or misusing randomization and p-values.
• The "postmodern p-value" represents a flawed approach to demonstrating the absence of an effect.
• Adherence to rigorous statistical practices is essential for valid scientific interpretation.