What you need to know
- 1.15Quantitative health risks, correlationA relationship in which two variables change together. A correlation does not prove that one causes the other., causationWhen a change in one variable directly causes a change in another. Showing it needs controlled experiments or a known mechanism. and conflicting evidence.
- 1.17Differences between perceived and actual risk.
By the end, you should be able to…
- Distinguish illness, mortality and risk.
- Calculate absolute and relative risk from a supplied dataset.
- Identify confounding and explain misleading risk claims.
Build your understanding
Read the outcome and denominator
Morbidity concerns illness; mortality concerns deaths. Incidence counts new cases in a defined populationAll the individuals of one species living in one area at the same time. over a stated period, whereas prevalence describes existing cases at a point or over a period. These measures answer different questions: do not substitute one for another.
A risk is a probability of an outcome during a specified period. In a simple cohort with complete follow-up, risk = people developing the outcome ÷ people initially at risk. A rate may instead use person-time. Always identify the denominator, population and follow-up period before comparing groups.
Absolute and relative comparisons
Suppose 40 of 1000 people in group A and 20 of 1000 in group B develop the same outcome over five years. Their risks are 4% and 2%. The absolute difference is 2 percentage points; the relative risk A/B is 2. Group A has twice the risk, or a 100% relative increase. These are different descriptions of the same fictional data.
Raw case numbers can mislead when populations differ in size. Even rates per 100000 may need age standardisation if populations have different age structures. Use like-for-like comparisons and acknowledge uncertainty when samples are small.
Correlation and cause
correlationA relationship in which two variables change together. A correlation does not prove that one causes the other. means two variables vary together. causationWhen a change in one variable directly causes a change in another. Showing it needs controlled experiments or a known mechanism. means changing one contributes to a change in the other. A confounding variable is associated with both the exposure and the outcome and can distort the observed relationship. For example, age might influence both medicine use and disease risk.
An association can prompt a causal hypothesis, but stronger evidence comes from temporal order, plausible mechanisms, appropriate comparisons and agreement across different study designs. Reverse causation is also possible: early illness could reduce activity rather than inactivity being the sole explanation for illness.
Why risk can feel different from the numbers
Vivid stories, familiarity, perceived control and delayed consequences can influence risk perception. A rare dramatic event may seem more likely than a common gradual disease. A relative percentage without the starting risk can make a small absolute change seem large. These influences explain perception; they do not replace evidence.
Make your reasoning visible
State the time period
A five-year risk cannot be compared directly with a one-year risk. Include the populationAll the individuals of one species living in one area at the same time. and outcome where relevant.
Use percentage points correctly
A fall from 4% to 2% is 2 percentage points, but a 50% relative reduction. Show your calculation rather than giving an unexplained percentage.
Avoid this misconception
The claim: A correlationA relationship in which two variables change together. A correlation does not prove that one causes the other. is useless because it does not prove causationWhen a change in one variable directly causes a change in another. Showing it needs controlled experiments or a known mechanism..
The correction: Correlation is useful evidence for generating and testing hypotheses. Its limitations mean that additional evidence is needed, not that it should be ignored.
Does this exposure really double the risk?
Two fictional groups of 1000 people are followed for five years. More of the exposed group develop CVD. Is the exposure the cause, or is something else going on?
What the model shows
Exposed and unexposed groups with their five-year risks, split by age if you choose, a diagram of possible causes and the headline numbers.
What to look for
Whether the difference stays when you compare people of the same age, and how the absolute difference compares with the relative risk.
What it tells us
A correlation is not, by itself, a cause: a confounder such as age can create or exaggerate it. Always ask for the starting risk and the time period.
Be careful what you claim
Fictional people and risks.
Try it in the model: read the two risks, then choose Split by age.
Risk, correlation and cause
Compare two fictional groups followed for five years. Work out absolute and relative risk, then split by age to see whether the difference is real or caused by a confounder.
What you see
- Two fictional groups of 1000 people, exposed and not exposed, followed for five years. Each dot is 5 people; red dots developed CVD. Split by age, the older people are shown separately.
- A cause diagram (age, exposure, CVD), the headline numbers, and bars of each group’s risk.
How to use it
- Set the true effect, whether older people are more often exposed, and the starting risk. Switch between Combined and Split by age.
What it shows
- Risk = cases ÷ people at risk over a stated time. Relative risk and absolute difference describe the same data in different ways (1.15).
- A correlation can come from a confounder. Splitting by the confounder shows whether an association remains. A big relative risk can hide a small absolute change (1.17).
Limits of the model
- All people, risks and the exposure are fictional, and counts are exact expected values with no sampling error. Real studies need larger samples, other confounders, and other evidence (timing, mechanism, different study designs).
- Only one confounder (age) is modelled. Reverse causation is not modelled.
Guided tour · step 1 of 8
- Read the five-year risks in the simulator (40 of 1000 vs 20 of 1000).
- In Study, just below these steps, set Compare to Split by age (options: Combined, Split by age).
Combined, the exposed group has a risk of 4% against 2%: a relative risk of 2 and an absolute difference of 2 percentage points. Split by age, the risks are the same within each age group. The exposure does nothing here: age is a confounder, because older people are both more often exposed and more at risk.
Study
Questions
Check your understanding
Try each question before opening the suggested answer.
Practice question
Retrieval and explanation · suggested answer
In a fictional study, 30 of 600 exposed people and 10 of 500 unexposed people develop CVD over the same period. Calculate their risks and the relative risk.
Show suggested answer
Exposed risk = 30/600 = 5%; unexposed risk = 10/500 = 2%. Relative risk = 5/2 = 2.5. The absolute difference is 3 percentage points.
Practice question
Retrieval and explanation · suggested answer
A headline says a factor doubles risk. Name two missing details needed to assess its importance.
Show suggested answer
The baseline absolute risk and the time period are essential. The outcome, population, sample size and study design also help interpret the 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 age groups have identical disease risks in two towns, but one town has a higher overall risk. Explain how this can happen.
Show a hint
Consider the proportions of younger and older people.
Explore the reasoning
If the town with higher overall risk contains a larger proportion of older people, and older people have higher disease risk, its combined risk will be higher even though age-specific risks match. Age structure confounds the crude comparison.
Recall this lesson
Lesson recall questions
Question 1 of 5: What is meant by a correlation in a table of illness data?
Illness data: correlation and causation
Practice question
Retrieval and explanation · suggested answer
What is meant by a correlation in a table of illness data?
Lesson: 1C 1: Risk, Correlation and Cause
Try this from memory first.
Put your biology to work
Explore where these ideas appear in the world. Enrichment, not extra required recall.
When “same category” does not mean “same risk”
1C · Risk, Correlation and CauseA headline can put processed meat and tobacco in the same carcinogenicity category and leave readers with a misleading comparison. IARC’s categories describe the strength of evidence that…Specification 1.15 · 1.17Read & use the scienceClose article
Why this matters
A headline can put processed meat and tobacco in the same carcinogenicity category and leave readers with a misleading comparison. IARC’s categories describe the strength of evidence that an exposure can cause cancer, not an equal size of risk at ordinary exposures. The distinction transfers directly to the health-risk reasoning in Unit 1: identify the outcome, exposure, baseline risk and time period before comparing claims. Biological mechanisms can help make an explanation plausible, but a category label is not a probability. Scientific communication needs to preserve that difference even when a dramatic headline would attract more attention.
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
- 1.17 understand why people’s perception of risks are often different from the actual risks, including underestimating and overestimating the risks due to diet and other lifestyle factors in the development of heart disease
Use the science
What quantitative information would you need before comparing the risk from two exposures placed in the same evidence category?
Explore a possible response
Absolute baseline and exposed risks, defined doses and durations, comparable outcomes, population characteristics and uncertainty. Evidence strength alone gives none of these effect sizes.
Explore the source
- Q&A on the carcinogenicity of the consumption of red meat and processed meatInternational Agency for Research on Cancer, WHO (2015)