Professional expertise in medical writing and editing includes fluency in the statistical methods most commonly used in one’s area of specialization. Sometimes, though, we can get really wrapped up in our statistics and get laser-focused on significance.
As readers of this blog know, a major pet peeve of mine is when someone uses “trending toward significance.” (Ick, no.) The converse can also be true–a statistically significant result is not always a meaningful one.
That distinction matters in medical writing because it is easy for language to imply more than the data support. A study may meet a statistical threshold, but that alone does not tell us whether the effect is large enough to matter in practice, noticeable to patients, or important enough to change decisions.
A small difference between groups can produce a low P value, especially in a large study. That result may be real in a statistical sense, but still modest in practical terms. On the other hand, a clinically important effect may fail to reach statistical significance in a smaller or noisier study. The P value does not carry the whole interpretive burden.
This is one reason effect size, confidence intervals, endpoint selection, and clinical context all matter. What changed? By how much? Is that difference meaningful for the patient, clinician, or decision-maker? Is it consistent with the rest of the evidence?
Significance, in other words, is not the same as importance.
This has real consequences for word choice and framing. A statistically significant finding should not automatically be described in language that suggests a major advance, a clear benefit, or a meaningful improvement—unless the broader context supports that conclusion. The goal is accurate interpretation, not just accurate arithmetic.