The Chromozone / IAL Psychology / Topic A: Social psychology
Directed study: exam technique
Exam technique: sampling questions. Answer about this study
Teach these first
This page builds on these lessons. Teach them before students start here:
How this page works
One exam skill, in four steps: learn the skill, see it done, practise it with help, then do it on your own.
Step 1: the skill 5 min
Most sampling questions in the exam are about a study in the question, not about sampling in general. Before you write a sentence, use the "any study" test: could this sentence be pasted into an answer about a different study? If yes, it is generic. Add a detail from this study: who, where, how many, how they were chosen.
| Question | What a full answer needs |
|---|---|
| Identify the technique (1) | The name only, for example "volunteer (self-selected) sampling". |
| Describe how to take the sample (2 or 3) | The steps, in order, using this study's population, list and numbers. |
| Explain a strength or a weakness (2) | The point, and why it matters here: who is in the sample, who is missing, and so what. |
| Suggest an improvement (2) | A change, and why it helps in this study: which people it brings in. |
| Evaluate the technique in this study (4) | Strengths and weaknesses, each linked to the study, and a short judgement. |
Key words: populations and samples
- Target population
- The whole group the researcher wants to draw conclusions about.
- Sampling frame
- A complete list of everyone in the target population, for example a staff list.
- Representative
- A sample that matches the target population in its key features, such as age, gender or job.
- Volunteer bias
- Volunteers often share traits, such as strong interest or free time, so the sample is not representative.
Common misconceptions and exam errors
These come up again and again in the reports on real exam answers. Each one costs marks.
- "Asking random people" is not random sampling. Random sampling needs a full list of the target population, so that everyone has an equal chance. Stopping whoever walks past is opportunity sampling.
- Generic strengths and weaknesses. "Volunteer samples are biased" earns little. Say what kind of person would volunteer for this study, and what that does to the results.
- "Use a bigger sample." On its own this is not an improvement. Say who is missing from the sample, and how your change brings them in.
- Equal strata instead of proportional ones. A stratified sample has each group in the same proportion as in the target population, not the same number from each group.
- Answering a different question. If the question asks which technique the study used, describe that one. Do not suggest the technique you would have used.
Step 2: worked example 6 min
Scenario: Dr Jayasuriya wants to find out how stressed nurses at a large hospital in Kandy feel during night shifts. She leaves 200 questionnaires in the nurses' staff room, with a note asking nurses to fill one in and put it in a box. 48 nurses return one.
Question (a): Explain one weakness of the sampling technique used in this study. (2)
Answer 1 1 of 2 marks
It is a volunteer sample, which causes volunteer bias, so it is not representative (1).
Why: Correct, but it passes the "any study" test: the same sentence fits every volunteer sample. No detail about the nurses.
Answer 2 2 of 2 marks
Only nurses who use the staff room and chose to fill in the questionnaire took part, perhaps the most stressed ones, who want to complain (1). So the 48 may not represent all the hospital's nurses, and the level of stress may be overestimated (1).
Why: Who is in the sample, why, and so what for this study's results.
Question (b): Suggest one improvement to the sampling in this study, and justify your suggestion. (2)
Answer 3 0 of 2 marks
She should use a bigger sample, so the results are more reliable.
Why: A bigger sample of the same kind of volunteers would have the same bias. It does not say who is missing.
Answer 4 2 of 2 marks
She could use the hospital's staff list to choose nurses at random from each ward, in the same proportions as in the hospital (a stratified sample) (1), so nurses who never visit the staff room, or who are less stressed, are included too, and the results apply to all the nurses (1).
Why: A clear change, and the people it brings in.
Step 3: guided practice, fix the mistake (10 marks) 13 min
Each answer makes one of the five mistakes in the purple box. Find it, then rewrite the answer.
Question: A researcher stands outside a library in Galle and asks people who walk out to fill in a questionnaire. Identify the sampling technique he used, and explain your answer. (2)
A student's answer: "Random sampling, because he asked random people who were passing."Rewrite the answer so that it earns both marks. (1.2.5) (2)
Question: Explain one weakness of the sample in Haun et al. (2014). (2)
A student's answer: "The sample was small, so it cannot be generalised."Rewrite the answer so that it earns both marks. (1.2.5, 1.3.4) (2)
Question: Suggest one improvement to Asch's (1951) sample, and justify your suggestion. (2)
A student's answer: "Asch should have used more participants."Rewrite the answer so that it earns both marks. (1.2.5, 1.1.5) (2)
Question: A school has 500 students: 300 travel by school bus and 200 walk or cycle. Calculate how many of each group should be in a stratified sample of 50. (2)
A student's answer: "25 bus students and 25 walkers, so that both groups are equal."Rewrite the answer so that it earns both marks. Show your working. (1.2.5) (2)
BusWalk or cycleQuestion: Describe the sampling technique used in Milgram's (1963) study. (2)
A student's answer: "He should have used a stratified sample so that women were included."Rewrite the answer so that it earns both marks. (1.2.5, 1.1.2) (2)
Reveal the mark scheme: guided practice
3(a) Random or not? (2)
AO2 (2 marks)
Credit one mark for identifying the technique and one mark for a reason linked to how people were chosen.
- Opportunity sampling (1)
- Only people leaving the library then could be chosen; random needs a full list, with an equal chance for everyone (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
3(b) Haun's sample (2)
AO3 (2 marks)
Credit one mark for identifying a weakness and one mark for justifying it, in relation to Haun et al. (2014).
- Weakness: a detail: 18 two-year-olds from Leipzig; 12 chimpanzees from one sanctuary; 12 orangutans from one centre (1)
- So the findings may not generalise to children in other cultures, or to apes living in the wild (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
3(c) Asch's sample (2)
AO3 (2 marks)
Credit one mark for identifying an improvement and one mark for justifying it, in relation to Asch (1951).
- Improvement: include women, non-students or people from other countries, for example a stratified sample of adults (1)
- So the 32% could be generalised beyond young American men (1)
No marks for "more participants" with no reason.
Generic answers score 0 marks.
Look for other reasonable marking points.
3(d) Stratified sample of 50 (2)
AO2 (2 marks)
Credit one mark for the working and the answer for each group.
- Bus: (300 ÷ 500) × 50 = 30 (1)
- Walk or cycle: (200 ÷ 500) × 50 = 20 (1)
Reject all other answers.
3(e) Milgram's sample (2)
AO1 (2 marks)
Credit up to two marks for an accurate description of the sampling technique used.
- Volunteer (self-selected) sampling (1)
- Men answered a newspaper advert or a letter for a study of memory: 40 men aged 20 to 50 from New Haven, paid $4.50 (1)
Look for other reasonable marking points.
Step 4: exam-style questions on your own (17 marks) 18 min
No help now. You need a calculator. Every answer is about the study in the question.
Scenario: a psychologist wants to study whether workers at a bank's head office in Colombo follow unwritten rules about dress and lunch breaks. The office has 800 workers: 480 customer service staff, 240 IT staff and 80 managers. She has a list of every worker, and wants a sample of 40 for a questionnaire.
Calculate how many workers from each group should be in a stratified sample of 40. Show your working. (1.2.5) (2)
Customer serviceITManagersDescribe how the psychologist could take this stratified sample. (1.2.5) (3) (about 60 words)
Explain one strength of using a stratified sample in this study. (1.2.5) (2)
Her assistant suggests a quicker way: hand out the questionnaire in the staff canteen at lunchtime instead.
Explain one weakness of the canteen sample for this study. (1.2.5) (2)
Explain one problem the psychologist could face when she chooses workers at random, and suggest how she could deal with it. (1.2.5) (2)
She finds that 9 of the 40 workers say they always follow the dress rules.
Explain why she cannot conclude that 22.5% of office workers in Sri Lanka always follow dress rules. (1.2.5) (2)
Scenario: another researcher puts a post on a university's student website, asking for volunteers for a study of whether students change their answers in an online quiz to match their friends' answers. 120 students reply.
Evaluate the use of volunteer sampling in this study. (1.2.5) (4) (about 90 words)
Reveal the mark scheme: questions 1 and 2
1(a) The stratified sample (2)
AO2 (2 marks)
Credit one mark for the working and one mark for all three correct answers.
- Method: (workers in the group ÷ 800) × 40 (1)
- Customer service 24; IT 12; managers 4 (all three) (1)
Check: 24 + 12 + 4 = 40.
Reject all other answers.
1(b) Describe how (3)
AO2 (3 marks)
Credit up to three marks for an accurate description in relation to the scenario.
- Split the staff list into the three groups (strata) (1)
- Number the people in each group and pick 24, 12 and 4 at random, for example with a random number generator (1)
- So each group is in the sample in the same proportion as in the office (60%, 30%, 10%) (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
1(c) A strength (2)
AO2 (1 mark), AO3 (1 mark)
Credit one mark for identifying a strength and one mark for justifying it.
- Strength: all three job groups are in the sample in proportion: the 80 managers cannot be left out or over-represented by chance (1)
- So the results are more likely to apply to the whole office of 800 (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
1(d) The canteen (2)
AO2 (1 mark), AO3 (1 mark)
Credit one mark for identifying a weakness and one mark for justifying it.
- Weakness: only workers who eat in the canteen at that time take part, for example not managers who eat at their desks (1)
- The study is about lunch-break rules, so canteen users may follow them more: biased results (1)
The second mark must link to this study.
Generic answers score 0 marks.
Look for other reasonable marking points.
1(e) A problem and a fix (2)
AO2 (1 mark), AO3 (1 mark)
Credit one mark for a problem and one mark for a way to deal with it.
- Some workers chosen may refuse or be away, so the sample becomes partly self-selected (1)
- Fix: replace each one with another worker chosen at random from the same group; or keep it short and anonymous (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
1(f) Generalising (2)
AO3 (2 marks)
Credit one mark for identifying a limit and one mark for justifying it.
- The target population is the 800 workers of one bank's head office (1)
- Other offices and companies may have different rules or cultures, so the 22.5% may not apply to them (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
2 Evaluate volunteer sampling (4): max 2 strengths, max 2 weaknesses; max 2 with no link to the quiz study
AO3 (4 marks)
Credit one mark for each accurate point, up to four marks. Answers must relate to the volunteer sample in the quiz study.
- Strength: quick: 120 students replied to one post (1)
- Strength: ethical: students chose to take part, which supports informed consent (1)
- Weakness: only students who read the website and have time, perhaps psychology students who may guess the aim (1)
- So the 120 may not represent all the university's students, and conformity may be under- or overestimated (1)
- A judgement: fine for a first study, but a random sample from the student list would be more representative (1)
Generic answers score 0 marks.
Look for other reasonable marking points.
Self-marking checklist 5 min
Before you hand in, check that:
- every answer passes the "any study" test: it uses the bank, the canteen or the quiz study;
- your stratified numbers add up to 40, and are in proportion to the groups;
- your describe answer gives the steps in order, with the staff list and the numbers;
- any improvement says who is missing and how the change brings them in.
After you hand in, the mark schemes open. Then:
- tick each mark point you earned; a generic point earns at most one mark;
- find the mistake from the purple box that cost you most, and rewrite that answer in a different colour;
- write your total out of 27, and one target, for example: "I will name the people who are missing."