Question

An essential part of scientific thinking isn't only how to use statistics correctly, but also how to identify the misuse of statistics. A primary reason for the misuse of statistics is

Answer 1

…people (or organizations) use numbers to persuade rather than to test hypotheses — driven by poor understanding of statistical principles plus cognitive and motivational biases. In practice that shows up as cherry-picking data, bad sampling, p‑hacking/multiple comparisons, confusing correlation with causation, misleading graphs/percentages, and hiding denominators or exclusions.

How to spot it (quick checklist)
- Who chose the sample and how was it selected? (Nonrandom or self‑selected samples are red flags.)
- Is the effect based on correlation only? Are causal claims justified or is confounding likely?
- Are results statistically significant but tiny in practical size?
- Were multiple outcomes or subgroups tested without correction (p‑hacking)?
- Are denominators, base rates, or time frames hidden or changed?
- Do graphs use truncated axes or distorted scaling to exaggerate differences?
- Is important data excluded or were outliers removed without explanation?
- Could incentives (political, commercial, reputational) explain a biased presentation?

If you keep these points in mind you’ll be better at recognizing misuse even when numbers look impressive.