Two Postmodernisms in Rheumatology
Yazici, H.
Bulletin of the Hospital for Joint Disease 75(2): 91-92
2017
ISSN/ISBN: 2328-5273 PMID: 28583052 Document Number: 692186
While randomization is an essential tool of data collection and interpretation, it can be misused. One such misuse is using a subgroup of a data set for validation purposes after randomizing the whole data set into two groups after which we attempt to validate the outcome data from one group with the outcome in the other. This surely validates only the randomization process. Two examples are given from disease criteria reports from Behçet's syndrome and gout. P-values are misunderstood and misused, and this is usually in the direction of investigators trying to show significant differences in their studies. Another, less criticized, misuse of the p-value is when the investigators try to use an insignificant and close to unity p-value as evidence for no difference. This author calls this a postmodern p-value.