Sampling and Data Collection AA SL
Almost every statistical claim rests on a sample, so how the sample is chosen decides whether the result can be trusted. This page shows how to name a sampling method and judge it, how to work out a systematic interval or a stratified allocation, how to spot bias, how to tell discrete from continuous data, and how to decide whether a data value is an outlier, with worked examples and exam-style questions.
Formulas at a glance
The interquartile range, , is in the formula booklet. The sampling interval, the stratified allocation and the rule for outliers are not, so learn them.
Proportion of the population sampled from each stratum
Sampling fraction
Not in the bookletNumber sampled from a given subgroup
Stratum sample fraction stratum size
Not in the bookletGap between selected members in systematic sampling
Systematic interval
Not in the bookletWhich method to use
Choose the sampling method from how the population is arranged and whether every member must have a chance of selection, then use the matching test for bias, data type or outliers.
| The situation is | Use | Notes |
|---|---|---|
| A full list of the population is available and every member should have an equal chance | Simple random sampling | Number the list and use random numbers. It avoids selection bias but needs the complete list |
| A list is available and a fixed gap between chosen members is acceptable | Systematic sampling | Interval , a random start from to , then every th member. It can be biased if the list repeats with a pattern matching |
| The population splits into groups of different sizes and each must be represented in proportion | Stratified sampling | Sample from a group , choosing members at random within each group |
| Interviewers must fill a fixed number from each group but choose the people themselves | Quota sampling | Not random within a group, so it can be biased |
| Whoever is easiest to reach is chosen | Convenience sampling | Quick and cheap, but likely to be biased and unrepresentative |
| A judgement on whether a data value is unusually far from the rest | The outlier rule: more than beyond the nearest quartile | Work out and , then compare. Decide in context whether to keep the value |
| A variable to classify as discrete or continuous | Counted values are discrete, measured values are continuous | Ask whether the value comes from counting or from measuring, not whether it looks like a whole number |
| A survey or sampling method whose results may not be fair | Look for who could be missed or pushed towards an answer | Check the sampling method, the survey question and missing data. A larger sample does not remove bias |
Section 1 of 2 SL4.1
Sampling methods
A sample is a group taken from a population, and the sampling method decides whether it fairly represents that population. The five methods to know are simple random, convenience, systematic, quota and stratified sampling, each with its own strengths and sources of bias. Name the method, describe how it works, and judge how reliable the results are.
Working out a sample
Stratified allocation
Also not in the booklet. Apply the same sampling fraction to each stratum's own size so the sample stays proportional across all the groups.
Key terms
What is the difference between a population and a sample?
The population is the whole group you want information about, and a sample is the part of it from which you actually collect data.
e.g. If students from a school of are surveyed, the population is all students and the sample is the .
What is a simple random sample?
Every member of the population has an equal chance of being chosen. It avoids selection bias but needs a complete list of the population.
e.g. Numbering people from to and choosing different numbers using a random number generator gives a simple random sample of .
What is convenience sampling?
Convenience sampling chooses whoever is easiest to reach. It is quick and cheap, but it is likely to be biased and unrepresentative.
e.g. Surveying the first students you meet in the canteen is a convenience sample.
What is sampling bias?
Bias occurs when the method makes some members of the population more likely to be chosen than others, so the sample does not represent the population. A larger sample does not remove it.
e.g. Surveying people only at a gym to estimate how much a town exercises is biased, because gym users exercise more than most people.
Worked example, step by step
The question No calculator needed [4 marks]
A library keeps an alphabetical list of its registered borrowers. A systematic sample of borrowers is required.
(a) Find the sampling interval.
(b) The first borrower chosen is number on the list. Write down the numbers of the next two borrowers chosen.
(c) Explain why systematic sampling could give a biased sample if the list repeated a pattern every names.
Worked solution: worked example
Answer: (a) . (b) and . (c) Every chosen borrower would hold the same position in the repeating pattern, so some types would never be sampled and the sample would be biased.
Step 1 - Find the sampling interval.
M1 - Divide the population size by the sample size, A1 -
Why? The interval tells you how far apart the chosen members are on the list. Dividing the population by the sample size makes the chosen members spread evenly along the whole list, so the last one chosen is near the end.
Step 2 - Check the interval is a whole number.
, so the interval is exactly and the sample will contain exactly borrowers.
Why? If the division is not exact, the interval is rounded down and the sample size can come out slightly different, so it is worth checking by multiplying back.
Step 3 - Apply the interval from the start.
The random start is number . Add each time: , then . The next two borrowers are numbers and . A1 - and
Why? Only the starting point is random. After that every th member is chosen, so the rest of the sample is fixed once the start is known.
Step 4 - Explain the possible bias.
If the list repeated a pattern every names, every chosen borrower would sit at the same place in the pattern, so for example all of them could be one type of borrower. Other types would never be chosen, so the sample would not be representative. R1 - States that every chosen member would hold the same place in the repeating pattern, so the sample is unrepresentative
Why? Systematic sampling is not simple random sampling. A hidden pattern in the list that matches the interval makes the selection systematically unrepresentative, which is the standard weakness to name.
M1 Divide the population size by the sample size,
A1
A1 and
R1 States that every chosen member would hold the same place in the repeating pattern, so the sample is unrepresentative
Now you try: enter your answers to be marked
These are real practice questions. Type your answer in each box and press Mark it. The worked solution and mark scheme appear straight away, and your first attempt is the one that counts. Use the keyboard for powers, roots and trig, and the working box for your notes.
Worked solution: practice question 1
(a) Find the number of students sampled from each of Year 1, Year 2 and Year 3.
Worked solution: practice question 2
Worked solution: practice question 3
Your score so far: 0 out of 10 marks
Common mistakes
- Mixing up stratified and quota sampling. Stratified sampling picks members randomly within each proportional subgroup; quota sampling only fixes how many are needed from each group and leaves the interviewer to choose who fills it, so it is not random within a group.
- Applying the sampling fraction to the wrong total. The fraction (sample size divided by population size) must be multiplied by each stratum's own size - not by the overall sample size or population size again.
- Assuming a bigger sample fixes a biased method. A biased technique, such as convenience sampling, stays biased no matter how large it grows; only switching to a properly randomised method removes the bias.
Section 2 of 2 SL4.1
Population, bias, data types and outliers
This section covers what a population and a sample are, how bias and unreliable data arise, how to classify data as discrete or continuous, and how to decide whether a data value is an outlier.
Key terms
Five words worth knowing cold, each with an example showing exactly what it means.
What is the difference between a population and a sample?
A population is the entire group you want to draw conclusions about. A sample is a smaller subset of the population that is actually studied, chosen because surveying the whole population is usually too slow, costly or impossible.
e.g. All 800 students at a school form the population; a stratified sample of 40 of them is the sample.
What is stratified sampling?
Stratified sampling divides the population into subgroups (strata) and samples from each in proportion to its size, guaranteeing that every subgroup is represented fairly. The proportion sampled from each stratum equals the overall sampling fraction.
e.g. For 240 staff (120 production, 80 sales, 40 admin) and a sample of 30, the fraction is , so production contributes
What is systematic sampling?
Systematic sampling selects every th member of a list after a random starting point, where is the sampling interval. It's fast and spreads the sample evenly, but can be biased if the list has a hidden pattern matching the interval.
e.g. For a population of 800 and a sample of 40, the interval is , so every 20th person is chosen.
What is convenience sampling?
Convenience sampling selects whoever is easiest to reach, such as the first people who happen to walk past. It's quick to carry out but is likely to be biased, since it only reaches a particular kind of person rather than a representative cross-section.
e.g. Surveying people leaving a gym about exercise habits over-represents people who already exercise regularly.
What is sampling bias?
Sampling bias occurs when the way a sample is chosen makes it systematically unrepresentative of the population, so conclusions drawn from it are unreliable. It can come from the sampling method itself, or from how survey questions are worded.
e.g. "Don't you agree that taxes are too high?" is a leading question that pushes respondents toward one answer.
Data and reliability
Discrete vs continuous
Discrete data is counted (whole numbers); continuous data is measured (any value in a range).
Number of cars is discrete; height or time is continuous.
Not in the formula booklet - prior knowledgeReliability of data
Consider missing data, recording errors, and whether the sampling method itself introduces bias.
A large sample doesn't fix bias - a biased method stays biased no matter how many people you ask.
Not in the formula booklet - key exam ideaOutliers
An outlier is a data item more than IQR from the nearest quartile.
Some outliers are genuine and should stay in the data; others are recording errors and should be investigated.
Not in the formula booklet - definition to knowWorked example, step by step
The question No calculator needed [5 marks]
The daily sales, in pounds, of a cafe over a period of weeks have lower quartile and upper quartile . The smallest value is and the largest value is . An outlier is a data item more than the interquartile range (IQR) from the nearest quartile.
(a) Find the interquartile range.
(b) Determine whether the smallest value or the largest value is an outlier.
(c) The largest value was recorded on the day of a local festival. State, with a reason, whether it should be removed from the data.
Worked solution: worked example
Answer: (a) . (b) is an outlier (it is above ); is not (it is above ). (c) Keep it, because it is a valid value explained by the festival.
Step 1 - Find the interquartile range.
A1 -
Why? The IQR is the spread of the middle half of the data, so it is the yardstick for what counts as far from the quartiles. It is in the formula booklet, but it is just the difference of the two quartiles.
Step 2 - Find the outlier limits.
. Lower limit: . Upper limit: . M1 - Use beyond the nearest quartile A1 - Limits and
Why? A value is an outlier only if it is more than beyond the nearest quartile, so the limits are measured from downwards and from upwards, not from the median.
Step 3 - Compare the extreme values with the limits.
The largest value , so it is an outlier. The smallest value , so it is inside the lower limit and is not an outlier. A1 - is an outlier and is not
Why? Check each end separately against its own limit. A value only just inside a limit, like here, is not an outlier.
Step 4 - Decide in context.
The value is a genuine sales figure for a festival day, so it should be kept. It is an unusual but valid value, not a recording error. R1 - Keeps the value, with a reason that it is a valid figure explained by the festival
Why? The syllabus asks you to interpret outliers in context. Some outliers are valid data and should stay in; others are mistakes, such as a mis-typed figure, and should be investigated or corrected.
A1
M1 Use beyond the nearest quartile
A1 Limits and
A1 is an outlier and is not
R1 Keeps the value, with a reason that it is a valid figure explained by the festival
Now you try: enter your answers to be marked
Worked solution: practice question 1
Worked solution: practice question 2
Your score so far: 0 out of 7 marks
Worked examples
Two exam-style questions with full worked solutions and mark schemes. Try each one yourself before reading the solution.
Consider the following variables.
(a) Classify the number of cars in a car park as discrete or continuous.
(b) Classify the height of a tree as discrete or continuous.
(c) Classify the time to run 100 m as discrete or continuous.
Worked solution
(a) Discrete (counted). A1
(b) Continuous (measured). A1
(c) Continuous (measured). A1
A school has 600 students in three year groups in the ratio . A stratified sample of size is taken; 20 students are chosen from the largest year group.
(a)(i) Find the number of students in the largest year group.
(a)(ii) Find the number of students in the middle year group.
(a)(iii) Find the number of students in the smallest year group.
(b) Find the sampling fraction.
(c) Hence find .
Worked solution
(a)(i) Parts largest group M1
A1
(a)(ii) Middle group A1
(a)(iii) Smallest group A1
(b) Largest group and chosen: fraction M1
A1
(c) M1
A1
Common mistakes
The four slip-ups that account for most of the marks lost on this topic - worth reading before you start practising, not just after you get one wrong.
- Mixing up stratified and quota sampling. Stratified sampling selects randomly within each proportional subgroup; quota sampling only fixes the quota size and lets the interviewer choose who fills it, so it isn't random within each group.
- Applying the population's sampling fraction to the wrong total. The fraction (sample size ÷ population size) must be multiplied by each stratum's own size, not by the total sample size or population size again.
- Assuming a bigger sample automatically fixes bias. A biased sampling method - like convenience sampling - stays biased no matter how large the sample gets; only changing the method fixes it.
- Classifying data by its numerical appearance rather than how it arises. A quantity like shoe size is discrete because it's counted from a fixed set of values, even though it can include halves - the test is whether it's measured (continuous) or counted (discrete), not whether it "looks like a whole number".
Syllabus content covered
| Code | Syllabus content |
|---|---|
| SL4.1 | Concepts of population, sample, random sample, discrete and continuous data. Reliability of data sources and bias in sampling. Dealing with missing data, errors in the recording of data. Interpretation of outliers. Outlier is defined as a data item which is more than interquartile range (IQR) from the nearest quartile. Awareness that, in context, some outliers are a valid part of the sample but some outlying data items may be an error in the sample. Sampling techniques and their effectiveness. Simple random, convenience, systematic, quota and stratified sampling methods. |
Official IB source: Mathematics: analysis and approaches subject brief (PDF, ibo.org).
Quick questions
What is the difference between stratified and quota sampling?
Stratified sampling selects members randomly within each proportional subgroup (stratum). Quota sampling only fixes how many are needed from each group and lets the interviewer choose who fills it, so it is not random within a group.
What's the difference between a population and a sample?
A population is the entire group you're interested in studying. A sample is a smaller subset of the population that is actually surveyed or measured. Statistics is largely about using what you find in a sample to draw reliable conclusions about the whole population.
How do I calculate how many to take from each group in a stratified sample?
Multiply the sample size by the fraction of the population in each group. For groups of and in a population of with a sample of , the numbers are and .
When should I use stratified sampling instead of simple random sampling?
Use stratified sampling when the population has distinct subgroups (like year groups or departments) and you want each one represented in proportion to its size. Simple random sampling doesn't guarantee this - by chance it could under- or over-represent a subgroup, especially with a small sample.
Why can a larger sample still be unreliable?
A larger sample reduces random variation but does not fix a biased method. If the method favours some members of the population, the results stay unrepresentative however many people are surveyed.
Why is convenience sampling usually unreliable?
Convenience sampling only reaches people who happen to be present or easy to reach at a particular time and place, which is rarely representative of the wider population. Surveying gym-goers about exercise habits, for example, over-represents people who already exercise a lot.
Is systematic sampling a random method?
Only the starting point is random. After that, every th member is chosen, so not every group of the required size is possible and it is not a simple random sample.
How do I calculate a stratified sample size for one group?
Find the sampling fraction - the total sample size divided by the total population - then multiply that fraction by the size of the subgroup. For example, sampling 30 from a population of 240 gives a fraction of 1/8, so a subgroup of 120 contributes 1/8 × 120 = 15 to the sample.
Ready to practise properly?
Practice questions on every section, marked instantly like the real exam.