Not long ago, a biostatistics colleague of mine told me an interesting story. She was working with a (now former) collaborator to analyze some clinical data that had some limitations (small sample size, inadequate power, large variability in the primary outcome). After a brief discussion, he said to her: “I don’t care what statistical test we use; I just need the one that’ll show my results are significant.” This comment was probably met with a heavy sigh, eye roll, and/or fit of rage.
There are a lot of misperceptions about statistics and data interpretation. For those of you with an interest in numbers, statistics, and what they all mean, this paper highlights and clarifies some common statistical misperceptions in medical and health research.