Random Sampling Methods Revision Notes | A Level AQA Biology
Random sampling removes investigator bias: every point in the population has an equal chance of being selected. Free AQA notes and 11 questions answered.
Updated 5 Sept 2026 · Revise Right Team
Random Sampling Methods is part of Investigating diversity in A Level AQA Biology.
8 min
Reading time
19
Lessons in this topic
A Level AQA
Pathway
Reviewed by Revise Right
Content review
What this lesson covers
Random sampling removes investigator bias: every point in the population has an equal chance of being selected. Free AQA notes and 11 questions answered.
Design appropriate methods to ensure random sampling within a biological investigation
Carry out random sampling within a single population using appropriate techniques
Analyse data from random samples to investigate the effect of position on leaf growth
Evaluate the reliability and validity of random sampling methods used in fieldwork
Revise Right Mocks '26
AQA A-Level Biology mock draft, 26–31 Oct. Draft coverage: Biology 7402/1 · Paper 1. Reviewed for an original-practice sitting; check student availability before registering. Marking and results will follow later, after separate approval. Eligible new accounts receive a five-day Pro trial when signup email verification completes. Previous trials keep their original expiry. After expiry, new papers need Pro unless you have an explicitly confirmed free sitting offer for this event. That offer adds no Pro time or AI learning. Saved attempts and released results remain readable. No card or automatic charge.
Use this lesson to stay inside Investigating diversity while connecting the detail back to the wider Genetic information, variation and relationships between organisms topic.
Random sampling: A sampling method in which every individual in a population has an equal probability of being selected, eliminating investigator bias. Used when investigating variation across a population or habitat.
Quadrat: A square frame of known area placed at randomly selected coordinates to define the sampling unit within a habitat. Used to count or measure organisms within a defined area.
Sampling bias: A systematic error introduced when the method of selection favours certain individuals over others, making the sample unrepresentative of the population. Eliminated by using random number generation to select coordinates.
Representative sample: A sample whose characteristics reflect those of the whole population, achieved through random selection and an adequate sample size. Required for valid conclusions about the population.
Revision notes
Read the core explanations from Random Sampling Methods before testing yourself from memory.
1. Why Random Sampling Matters in Biology
Picture a field of grass. You want to know whether leaves near the edge grow differently from those in the centre. If you walk out and pick the leaves that catch your eye, you have already introduced bias — you might unconsciously choose the largest, most visible ones. That single decision can invalidate your entire investigation.
Random sampling removes that human bias from the equation. When every individual in a population has an equal chance of being selected, the sample you collect genuinely represents the whole population — and your conclusions carry scientific weight.
At A Level, you are expected not only to carry out random sampling but to justify why your method is random, evaluate its limitations, and use it to draw valid conclusions about biological variation. This lesson builds all three skills.
2. Designing a Random Sampling Method
Designing a random sampling method means making deliberate choices at every stage so that investigator bias cannot creep in. The core principle is straightforward: the investigator must not choose where to sample.
The standard approach in fieldwork is to lay two tape measures at right angles along the edges of the study area, forming an x-axis and a y-axis. Random number pairs then give coordinates, and a quadrat or sampling point is placed at each coordinate. Because the coordinates are generated independently of the investigator, every point in the area has an equal probability of being chosen.
Lay two tape measures at right angles to create a grid across the study area
Use a random number generator (calculator, app, or random number table) to produce coordinate pairs
Place the quadrat or sampling point at each coordinate pair
Record measurements at every selected point — do not skip points that seem inconvenient
Repeat for a sufficient number of sampling points to give a representative sample
The number of samples matters. Too few and chance variation between individuals will dominate your results; too many and the investigation becomes impractical. In most A Level fieldwork contexts, a minimum of 10 sampling points is expected, though the appropriate number depends on the variability of the population and the size of the study area.
One subtlety worth noting: a random method does not guarantee a perfectly spread sample. By chance, several coordinates might cluster in one corner of the study area. This is not a flaw in the method — it is an expected property of true randomness. If you deliberately spread your quadrats to avoid clustering, you have introduced systematic sampling, which is a different (and sometimes valid) approach, but it is no longer random.
3. Carrying Out Random Sampling Within a Single Population
Knowing the theory is one thing; executing the method correctly in the field is another. When sampling within a single population — for example, all the leaves on a single plant species growing across a meadow — you need to define the population boundaries clearly before you begin.
Start by marking out the study area with tape measures. The area should encompass the full extent of the population you are investigating. If you only mark out part of the habitat, individuals outside that boundary have zero probability of selection, which means your sample does not represent the whole population.
A critical point about what you record: you must measure every leaf (or every individual of the target species) within each quadrat, not just the ones that look typical. Selecting only 'average-looking' leaves within a quadrat reintroduces bias at the measurement stage, even if the quadrat placement was genuinely random.
4. Investigating the Effect of Position on Leaf Growth
One of the most instructive applications of random sampling at A Level is using it to test whether position within a habitat affects a measurable characteristic — in this case, leaf growth. The underlying biological question is whether environmental gradients (light, water availability, wind, soil nutrients) vary across the study area and whether those gradients produce measurable differences in leaf size or mass.
To investigate this, you collect leaf measurements from randomly placed quadrats and then compare measurements from different zones — for example, the edge of a woodland versus the interior, or the north-facing slope versus the south-facing slope. The random sampling ensures that any difference you detect reflects a genuine biological pattern rather than an artefact of where you chose to look.
Zone
Mean leaf length (mm)
Number of quadrats sampled
Range (mm)
Edge (0–2 m from boundary)
42
10
28–61
Intermediate (2–5 m)
55
10
38–74
Interior (>5 m)
68
10
52–89
The table above shows a pattern typical of woodland edge investigations: leaves in the interior tend to be larger than those at the edge. A plausible biological explanation is that interior leaves experience lower light intensity and respond by increasing their surface area to maximise light capture — a shade adaptation. However, you must be cautious: correlation between position and leaf size does not prove causation. Other variables (soil moisture, competition, herbivory) may also differ between zones.
When writing up an analysis of positional effects, the command word 'explain' requires you to link the observed pattern to a biological mechanism. Stating that 'leaves are bigger in the interior' earns a describe mark. Adding 'because lower light intensity in the interior results in the development of larger leaf surface area (a phenotypic response) to maximise light absorption for photosynthesis' earns the explain mark.
Worked example: Comparing Leaf Lengths Between Two Zones Using Means
A student samples 5 quadrats in the edge zone and 5 in the interior zone of a woodland. The leaf lengths (mm) recorded are: Edge: 38, 41, 35, 44, 42. Interior: 65, 70, 62, 68, 75. Calculate the mean leaf length for each zone and the difference between the means.
Mean (interior) = 340 ÷ 5 = 68 mm — Divide the total by the number of quadrats sampled (5). The mean interior leaf length is 68 mm.
Difference = 68 − 40 = 28 mm — The interior leaves are on average 28 mm longer than edge leaves. This difference suggests a positional effect, though a statistical test would be needed to confirm it is not due to chance.
Answer: Mean edge leaf length = 40 mm; Mean interior leaf length = 68 mm; Difference = 28 mm
Calculating means is a starting point, not a conclusion. At A Level, you are expected to recognise that a difference between two sample means could arise by chance, particularly with small sample sizes. A statistical test — such as the Student's t-test for normally distributed data, or the Mann-Whitney U test for non-parametric data — allows you to determine whether the observed difference is statistically significant.
5. Evaluating Random Sampling: Reliability, Validity, and Limitations
Designing and carrying out a random sample is only part of the skill. Evaluating it critically — identifying what could go wrong and how to improve it — is where A Level marks are won and lost.
Reliability refers to the consistency of your results: would you get similar data if you repeated the investigation? Increasing the number of sampling points and taking repeat measurements at each point both improve reliability. Validity refers to whether your method actually measures what you intend to measure: are you genuinely sampling randomly, and is your dependent variable a fair measure of leaf growth?
Potential limitation
Effect on data
How to address it
Small sample size (e.g. fewer than 10 quadrats)
Results dominated by chance variation; means unreliable
Increase the number of randomly generated coordinates
Investigator selects which leaf to measure within a quadrat
Introduces measurement bias; larger or more accessible leaves favoured
Measure all leaves of the target species within each quadrat, or use a secondary random selection within the quadrat
Study area boundary excludes part of the population
Individuals outside the boundary have zero probability of selection
Ensure tape measures span the full extent of the population
Environmental conditions change during data collection
Measurements taken at different times of day may not be comparable
Complete all measurements within a single session or standardise the time of measurement
When evaluating in an exam, structure your response around the specific method used. Vague statements such as 'the sample size was too small' earn limited credit. Precise statements such as 'sampling only 5 quadrats in each zone means the means are heavily influenced by individual outliers, reducing reliability; increasing to 20 quadrats per zone would reduce this effect' demonstrate the analytical depth the mark scheme rewards.
6. Exam Technique: Scoring Full Marks on Sampling Questions
AQA questions on sampling methods tend to follow predictable patterns. Recognising those patterns — and knowing the precise language the mark scheme expects — is the difference between a 3-mark answer and a 6-mark answer.
When asked to 'describe a method to ensure random sampling', examiners expect you to name the tool (random number generator), explain how it is used (to generate coordinate pairs), and state how those coordinates determine sampling locations (placed on a grid formed by two tape measures at right angles). All three elements are typically needed for full marks.
Random sampling is the methodological foundation of any valid investigation into biological variation. Whether you are comparing leaf sizes across a woodland, counting species in a meadow, or measuring growth rates across a population, the principle is the same: let the numbers decide where you look, measure everything you find there, and let the data tell the story.
Sources and content standards
Published by Revise Right. Use the official specification reference to check the requirements for your course. Content updated: 2026-09-05.
What is the difference between random and systematic sampling?
In random sampling, coordinates are generated by chance, so some samples may cluster together, which is an expected property of true randomness. Deliberately spreading quadrats evenly to avoid clustering is systematic sampling, which can be a valid approach but is no longer random.
Why is random sampling important?
Random sampling removes investigator bias: every individual or point in the population has an equal chance of being selected, so the sample genuinely represents the whole population. If an investigator picks leaves that catch the eye, they may unconsciously choose the largest ones, which can invalidate the investigation.
How do you carry out random sampling in a field?
Lay two tape measures at right angles along the edges of the study area to form a grid, use a random number generator or random number table to produce coordinate pairs, and place a quadrat at each coordinate. Record measurements at every selected point, including inconvenient ones, across a sufficient number of points.
How can you improve the reliability of random sampling?
Reliability improves by increasing the number of randomly generated sampling points and taking repeat measurements. Validity improves by measuring all leaves of the target species in each quadrat, making the tape measures span the full extent of the population, and completing all measurements in a single session.
Does position affect leaf size?
In woodland edge investigations, leaves in the interior tend to be larger than at the edge, for example a mean of 68 mm in the interior compared with 40 mm at the edge. A plausible explanation is that lower light intensity in the interior leads to a larger leaf surface area to maximise light absorption, but correlation does not prove causation.
Is it a problem if random samples cluster together?
Clustering of random coordinates is not a flaw: by chance several samples may land in one corner, which is an expected property of true randomness. Deliberately spreading quadrats to avoid clustering turns the method into systematic sampling, which is no longer random.
Why should every leaf in a quadrat be measured?
Every leaf of the target species in a quadrat should be measured because choosing only 'average-looking' leaves reintroduces bias at the measurement stage, even when quadrat placement was random. A secondary random selection within the quadrat can also be used.
Why are leaves bigger in the middle of a woodland?
Leaves in a woodland interior tend to be larger because lower light intensity leads to a larger leaf surface area to maximise light absorption for photosynthesis, a shade adaptation. Other variables such as soil moisture, competition and herbivory may also differ, so correlation does not prove causation.
Which statistical test compares the means of two samples?
The Student's t-test is used for normally distributed data, and the Mann-Whitney U test for non-parametric data, to decide whether a difference between two sample means is statistically significant. A difference between means alone could arise by chance, especially with small samples.
Why must the whole population area be marked out before sampling?
The study area must cover the full extent of the population, because individuals outside the boundary have zero probability of being selected. If only part of the habitat is marked out, the sample does not represent the whole population.
What is the difference between reliability and validity in fieldwork?
Reliability is the consistency of results, improved by more sampling points and repeat measurements. Validity is whether the method measures what it intends, for example whether sampling is genuinely random and leaf length is a fair measure of leaf growth.