What you need to know
- 1.15–1.17Quantitative evidence, study evaluation, conflicting findings and risk perception.
- 1.19Applying scientific evidence to risk reduction.
By the end, you should be able to…
- Build a justified conclusion using data and method.
- Distinguish an association, a mechanism and an intervention outcome.
- Explain why a relative benefit depends on baseline risk.
Build your understanding
Match the conclusion to the question
Start by defining the claim: what exposure or intervention, in which populationAll the individuals of one species living in one area at the same time., compared with what, over what period, and measured using which outcome? A study of blood cholesterolA lipid found in cell membranes. High levels of LDL cholesterol in the blood are a risk factor for atheroma. is not automatically a study of heart-attack mortality, even when the two are related.
Evaluate sample selection, measurement quality, confounders, follow-up and uncertainty. A large observational study and a small trial have different strengths and limitations; study labels alone do not settle which provides the most useful evidence for a particular question.
Combine independent lines of evidence
Mechanistic studies can show how an exposure could act. Observational studies can identify patterns in people. Intervention studies can test whether changing a factor changes an outcome. Confidence rises when well-designed, independent approaches support the same explanation.
A systematic review uses explicit methods to find and assess relevant studies. It can reduce selective use of evidence, but its conclusions depend on the quality and comparability of the included studies. A pooled estimate does not make biased measurements unbiased.
State what the numbers mean
Suppose a fictional intervention reduces ten-year event risk from 10% to 8%. The absolute reduction is 2 percentage points and the relative reduction is 20%. In a second population, a change from 2% to 1.6% has the same 20% relative reduction but a smaller absolute reduction of 0.4 percentage points.
Report uncertainty when supplied, and avoid implying an exact estimate is guaranteed for every person. Distinguish absence of convincing evidence from proof of no effect: a short or imprecise study may miss a real difference.
Reach a proportionate conclusion
A useful evaluation states the supported finding, gives relevant numerical evidence, explains an important limitation and identifies what further evidence would strengthen the conclusion. Do not merely list every possible flaw.
Evidence can inform changes in smoking, activity and diet, but information alone does not determine behaviour. Access, cost, social context and perceived control can affect choices. Explain risk perception without blaming individuals or treating a single anecdote as a population estimate.
Make your reasoning visible
Use evidence in both directions
If asked to evaluate, consider supporting findings and important limitations before giving a reasoned conclusion.
Quote paired values
When comparing groups, give both values, the units and the time period where available. Explain what the difference means.
Avoid this misconception
The claim: A non-significant result proves the intervention has no effect.
The correction: An imprecise study may be compatible with both no effect and a meaningful effect. Interpret uncertainty and study power rather than equating a statistical label with proof.
Check your understanding
Try each question before opening the suggested answer.
Practice question
Retrieval and explanation · suggested answer
A fictional intervention changes five-year risk from 5% to 4%. Calculate absolute and relative reduction.
Show suggested answer
Absolute reduction = 1 percentage point. Relative reduction = (5 − 4)/5 × 100 = 20%.
Practice question
Retrieval and explanation · suggested answer
A study finds lower cholesterol after two weeks and claims it proves fewer deaths over ten years. Evaluate the conclusion.
Show suggested answer
It supports a short-term change in cholesterol, but does not directly measure ten-year mortality. Longer follow-up with appropriate clinical outcomes and comparison groups is needed to test that claim.
Take the idea further
Apply the same specification knowledge to an unfamiliar problem. Any extra assumptions are supplied in the question.
Stretch question
Two studies give similar effect estimates, but one has a much wider uncertainty interval. Does that mean their results conflict?
Show a hint
Distinguish the estimate from its precision.
Explore the reasoning
Not necessarily. Similar estimates can be consistent while differing in precision. The wider interval may reflect a smaller sample or more variable measurements. Compare the ranges and designs rather than treating precision alone as a contradiction.
Recall this lesson
Lesson recall questions
Question 1 of 5: Explain, in causal order, how a confounding variable can create a false impression of causation.
Illness data: correlation and causation
Practice question
Retrieval and explanation · suggested answer
Explain, in causal order, how a confounding variable can create a false impression of causation.
Lesson: 1C 1: Risk, Correlation and Cause
Build the sequence from memory first. Then reveal the model order.
Tap the steps in the order you recall. Tap a placed step to send it back.
- 01Next step
- 02Waiting
- 03Waiting
Tap to place
Put your biology to work
Explore where these ideas appear in the world. Enrichment, not extra required recall.
Turning many papers into a useful conclusion
1C · Using the EvidenceResearchers who synthesise evidence turn a collection of studies into a structured assessment of a health question. Cochrane explains how systematic reviews aim to reduce bias in finding…Specification 1.15 · 1.16Read & use the scienceClose article
Why this matters
Researchers who synthesise evidence turn a collection of studies into a structured assessment of a health question. Cochrane explains how systematic reviews aim to reduce bias in finding and evaluating that evidence. Evidence-synthesis researchers define a question and search systematically, rather than selecting the papers that fit a preferred answer. They compare designs, populations, outcomes and possible sources of bias before deciding whether results can reasonably be combined. A large but poorly controlled study and a small randomised trial do not answer exactly the same question. This makes evidence synthesis a scientific task: judge how each result was obtained and what it measures, rather than treating agreement as a majority vote.
Back to the specification
Pearson Edexcel IAL Biology · Issue 2, February 2021 · Unit 1, pp. 15–18. The application extends beyond required recall.
- 1.15 be able to analyse and interpret quantitative data on illness and mortality rates to determine health risks, including distinguishing between correlation and causation and recognising conflicting evidence
- 1.16 be able to evaluate the design of studies used to determine health risk factors, including sample selection and sample size used to collect data that is both valid and reliable
Use the science
Why is “six studies agree and four disagree” an incomplete summary?
Explore a possible response
Studies differ in size, quality, precision, populations and outcomes. Counting them equally can hide these differences and selective publication.
Explore the source
- Chapter 1: Starting a reviewCochrane
Half a million volunteers can still be biased
1C · Using the EvidenceLarge biobanks let epidemiologists investigate relationships between health, genes and behaviour. But people who volunteer may differ from those who do not. A 2024 study explored how weighting…Specification 1.15 · 1.16Read & use the scienceClose article
Why this matters
Large biobanks let epidemiologists investigate relationships between health, genes and behaviour. But people who volunteer may differ from those who do not. A 2024 study explored how weighting UK Biobank participants using population information changed estimated associations. Weighting gives different influence to records to better represent a target population; it is not a magical repair for every missing variable. This takes sample selection beyond the instruction to use more people. Increasing sample size can reduce random uncertainty while leaving systematic bias intact. An epidemiologist must ask whose lives are represented before interpreting a precise-looking result.
Back to the specification
Pearson Edexcel IAL Biology · Issue 2, February 2021 · Unit 1, pp. 15–18. The application extends beyond required recall.
- 1.15 be able to analyse and interpret quantitative data on illness and mortality rates to determine health risks, including distinguishing between correlation and causation and recognising conflicting evidence
- 1.16 be able to evaluate the design of studies used to determine health risk factors, including sample selection and sample size used to collect data that is both valid and reliable
Use the science
Why can adding more participants recruited by the same biased method narrow a confidence interval without making the estimate closer to the population value?
Explore a possible response
More observations reduce sampling variation but do not necessarily remove systematic differences between volunteers and the target population.
Explore the source
- Reweighting UK Biobank corrects for pervasive selection bias due to volunteeringvan Alten et al., International Journal of Epidemiology (2024)