A Threshold in a Word: Teaching Pupils to Read 'Significant' as Evidence

In a BBC News report, a new teaching intervention may be described as producing a 'significant improvement'. In an A-level biology mark scheme, a correlation may be called 'significant' only after a critical value has been consulted. In a GCSE physics answer, a pupil may write that a difference is 'significant' because it looks large on a graph. The same word is doing three different jobs, and readers are expected to know which one is in play.

That is not a trivial vocabulary problem. 'Significant' is one of the places where ordinary English and scientific argument quietly diverge. In everyday usage it means important, notable or worth attention. In statistical usage it means that an observed result has met a pre-specified criterion for being unlikely under a stated null hypothesis. The second meaning does not guarantee the first. A result can be statistically significant and practically trivial; it can be important and not statistically significant. Reading science requires pupils to notice the difference.

A word with two lives

The everyday sense of 'significant' is evaluative. It tells the reader that something matters. The statistical sense is procedural. It tells the reader that a comparison has crossed a threshold. That threshold is usually a probability, such as p < 0.05, or a critical value in a statistical test. At A-level, pupils meet critical values in tests such as Spearman's rank correlation coefficient and the chi-squared test. At GCSE, they may meet simpler versions of the same idea: whether a difference is likely to be due to chance, or whether repeats overlap enough to make a conclusion uncertain.

The statistical sense is also conditional. It depends on sample size, variability, the chosen threshold and the test used. A very large sample can make a tiny difference statistically detectable. A small sample can leave a substantial difference statistically uncertain. Neither outcome tells the reader, by itself, whether the difference is educationally, medically or scientifically important. That requires a further judgement about effect size, context and consequences.

Where the slide happens

The slide is easiest to see in three places.

First, textbooks. A glossary may define 'significant' as 'important' early in a chapter, then use 'statistically significant' later without revisiting the distinction. Pupils who read the glossary faithfully may carry the everyday meaning into the statistical sentence.

Second, news reports. A headline may say that a study found a 'significant link' between two variables. The underlying paper may report a statistically significant association with a small effect. The headline is not necessarily wrong, but it invites the reader to hear importance where the evidence may only support detectability. UK science journalism often compresses this distinction because headlines have little room; the reader has to restore it.

Third, exam questions. A question may ask pupils to 'suggest whether the difference is significant' or to 'explain whether the correlation is significant'. The expected answer usually requires a comparison and a criterion, not a general assertion that the result seems large. Pupils who read 'significant' as emphasis may answer with emphasis: 'the difference is very significant because the line goes up a lot.' The command word has been read, but the evidential structure has not.

Comparison, threshold, claim

A useful classroom routine is to treat 'significant' as a hinge word that opens three questions.

What is being compared? A significant result is always a result about a comparison. It may compare two groups, two conditions, an observed value with an expected value, or a sample statistic with a hypothesised population value. Pupils should be able to name the comparison before they interpret the word.

What threshold marks the boundary? Statistical significance is not a property that floats free of criteria. There must be a stated threshold, whether a p-value, a critical value or an uncertainty rule. If the threshold is absent, the reader should ask what has been omitted. In school practical work, the threshold may be informal: whether the ranges of repeats overlap, or whether a difference is larger than the variation within groups. Even an informal threshold should be made explicit.

What claim is being made? 'The difference was significant' is one claim. 'The intervention works' is another. 'The effect is large enough to justify changing practice' is a third. Pupils need practice in separating the statistical finding from the practical or scientific interpretation. This is not a reason to distrust statistics. It is a reason to read them as evidence with a scope.

Classroom-ready questions

The following questions can be used with a textbook extract, a news article or a past paper. They work best when pupils annotate the text rather than answer orally.

• Underline every use of 'significant'. For each one, substitute 'important' and then 'statistically detectable'. Which reading fits the sentence, and what tells you?
• Find the comparison. Is it between groups, conditions, observed and expected values, or a sample and a claim about a population?
• Find the threshold. Is a p-value, critical value, confidence level or uncertainty rule stated? If not, what is missing?
• Separate the finding from the interpretation. Write one sentence that states only what the significant result supports. Then write a second sentence that states what the author or pupil wants to conclude. What has been added?
• Rewrite a headline without the word 'significant'. What changes: the information, the emphasis, or both?
• In an exam answer, include the comparison and the threshold before using the word 'significant'. Does the answer still make sense if you replace 'significant' with 'met the stated criterion'?

These questions are not aimed at making pupils cynical about the word. They are aimed at making them precise. A pupil who can say 'the difference was statistically significant at the 5% level, but the effect was small' is reading more carefully than one who says 'the difference was significant' and means 'the difference was big'.

Reading the word as evidence

'Significant' is not the most difficult word in science. It may be one of the most instructive. It shows pupils that scientific prose often carries ordinary words with technical weights, and that reading for evidence means asking what kind of claim a word is licensed to make. In the UK classroom, where pupils move between news articles, textbooks, practical write-ups and exam questions, that lesson is worth teaching directly.

The word should not be banned or replaced with jargon. It should be read. When pupils locate the comparison, the threshold and the claim, 'significant' stops being a vague intensifier and becomes what it is in good scientific writing: a marker of a conditional, evidence-bound judgement.