Statistics can strengthen an argument, but numbers do not interpret themselves. A percentage can be accurate yet misleading when the baseline is hidden, the sample is unrepresentative or the comparison has been selected to create a particular impression.
Always find the denominator
“Complaints increased by 100%” sounds dramatic. If complaints rose from one to two, the percentage is correct but the absolute change is small. Ask: 100% of what?
Relative and absolute change
If a risk falls from 2 in 1,000 to 1 in 1,000, that is a 50% relative reduction and an absolute reduction of 1 in 1,000. Both descriptions can be correct. A careful reader asks which representation best communicates the practical magnitude.
Sample quality
A survey of 500 volunteers from one online community may not represent an entire country. Sample size matters, but selection matters too. A large biased sample can still produce a misleading estimate.
Averages can hide distributions
An average can rise because everyone improved slightly or because a small group improved dramatically. Mean, median and range answer different questions. If a persuasive text reports only one summary measure, ask what variation is hidden.
Correlation is not automatically causation
If students who read more score higher, reading may contribute to performance, but other variables may also differ. A persuasive text that turns association directly into proof of cause has strengthened the claim beyond the evidence.
Time windows can frame results
A company can choose a starting month that makes growth look unusually strong. A campaign can report improvement immediately after launch while ignoring later decline. Check the period and ask whether another reasonable period would change the impression.
Worked example
Claim: “Four out of five users prefer our new design.” Questions: How many users were asked? How were they selected? What were the alternatives? Was “prefer” measured after actual use or after viewing an image? Who conducted the survey? These questions do not reject the result; they establish what confidence it deserves.
Statistical reading checklist
- What exactly was measured?
- What is the denominator or baseline?
- How large and representative is the sample?
- Is the text reporting relative or absolute change?
- Is an association being presented as a cause?
- What time period and comparison were selected?
Continue the evidence-literacy route
Continue with Charts and Infographics, Fact, Opinion and Interpretation, and Source Credibility and Bias.