How to Answer Research Methods and Study Design Questions in A Level Psychology (9990)

Research methods in Cambridge International AS & A Level Psychology (9990) are not confined to memorising definitions. Under the current 2024-2026 syllabus, Paper 2 is titled Research Methods and includes short, scenario-based and planning questions, while Paper 4 is titled Specialist Options: Application and Research Methods and also requires candidates to design investigations in a specialist-option context. Cambridge's own guidance repeatedly emphasises application: a candidate must take methodological knowledge and make it work in the investigation described in the question.

Know where the research-method marks come from

Paper 2 tests research methods directly across areas such as experiments, self-reports, case studies, observations, correlations, hypotheses and aims, variables, experimental design, controls, sampling, ethics, validity, reliability and data analysis. The specimen Paper 2 is 1 hour 30 minutes and worth 60 marks, and the revised format includes a Section B planning question. At A Level, Paper 4 adds research methods to the specialist options. The paper no longer has the old essay section. Instead, candidates apply psychological knowledge and research methodology within one of the specialist contexts they have studied. The 2024 Paper 4, for example, asked candidates to plan a postal questionnaire about medical appointments and a structured observation of workplace sabotage.

Click to enlarge

Read the method before you invent the design

Study-design questions often specify the broad method. If the question says 'using a postal questionnaire', do not redesign it as an interview because you think interviews are better. If it specifies a structured observation, your task is to make that observation workable. The method is a constraint, not a suggestion. Then underline every feature the question explicitly requires. In June 2024 Paper 4, the postal-questionnaire plan had to include question format and sampling technique. The structured-observation plan had to include behavioural categories and steps for making the study reliable. Missing one named feature immediately limits the completeness of the plan.

Make every required feature explicit

Cambridge's specimen Paper 2 planning grid rewards responses that describe all required features explicitly and in detail, using appropriate research-method terminology. The top band is not reached by implying that something will happen. If the sample is random, say how the random sample will be obtained. If the dependent variable is a score, explain how it will be measured or calculated. The same principle applies in Paper 4. A strong plan names the research method features and applies them to the actual investigation. Do not write a generic paragraph on what a questionnaire, interview or observation usually looks like.

Click to enlarge

Operationalise the variables

One of the clearest differences between a vague plan and a replicable plan is operationalisation. 'Measure stress' is not enough. State how stress will be measured: a score on a specified rating scale, a physiological measure, a behavioural count or another method that fits the investigation. For an experiment, define exactly how the independent variable changes between conditions and exactly how the dependent variable is measured. Cambridge's own AO2 teaching guidance says learners need to practise identifying and operationalising independent and dependent variables in novel situations.

Choose the experimental design deliberately

If the task involves an experiment, decide whether the design is independent measures, repeated measures or matched pairs and explain how participants are allocated. The design affects participant variables, order effects and practicality. Do not confuse experimental design with experimental type. Laboratory, field and other experiment settings describe where and how control operates; independent measures, repeated measures and matched pairs describe how participants are used across conditions.

Control the variables that could actually change the result

A good plan does not produce a long list of constants for the sake of it. Identify plausible participant or situational variables that could affect the dependent variable, then state how they will be controlled. Cambridge's teaching guidance specifically connects control of variables with validity and replicability. For a memory experiment, the word list, exposure time and testing instructions may need to be standardised. For a consumer study, the time of day, retail environment or stimulus material may matter. The relevant controls depend on the study.

Sampling technique needs a procedure, not only a label

If you choose a random sample, explain how the sampling frame is obtained and how selection is made randomly. If you choose an opportunity sample, identify where the accessible participants will come from. If you use self-selection, explain how volunteers are invited. Paper 4 often follows the plan with a reason, strength or weakness question about sampling. That means your initial sampling decision should be one you can justify in context rather than a technique selected because you memorised it.

Write questionnaire and interview features as actual questions

If the method is a questionnaire or interview, specify the question format and give at least one suitable example when relevant. Closed questions can generate quantitative data that are easy to compare; open questions can provide richer qualitative detail. A Likert scale may measure degree of agreement, but the response options and scoring need to be clear. For interviews, distinguish the format and the technique the question requires. A face-to-face interview may allow clarification and rapport but can increase interviewer effects. A telephone interview may be more practical but removes visual cues. These points become stronger when tied to the actual topic.

For observation, define the behavioural categories

A structured observation needs categories that are observable, objective enough to code consistently and relevant to the research question. 'Aggressive behaviour' is too broad if two observers could interpret it differently. A category such as 'hits machinery', 'removes a component' or 'deliberately stops the line' is clearer if those behaviours fit the scenario. Cambridge's June 2024 report specifically says candidates need a clear understanding of the required features for each research method. For observation, reliability steps such as operationalised categories, observer training or inter-observer agreement need to be tied to what the observers will actually record.

Reliability and replicability need concrete procedures

Do not write 'make it reliable' as an instruction. State what will make the procedure consistent: standardised instructions, the same timing, the same questionnaire wording, clear behavioural categories, a pilot study, trained observers or another relevant step. Cambridge's AO3 materials explicitly connect reliability with replicability. A study that is described precisely enough for another researcher to repeat gives you a stronger basis for checking whether the results are consistent.

Ethics should alter the procedure

Ethics are not a list to attach at the end. If informed consent is needed, explain how it will be obtained. If the study involves sensitive information, explain confidentiality or anonymity. If deception is necessary, explain debriefing. If withdrawal could affect the procedure, state how participants are told they may leave. The ethical step should fit the method. An anonymous postal questionnaire needs different practical protections from a face-to-face interview about health behaviour.

Connect the plan to psychological knowledge

Paper 4 explicitly asks candidates to identify psychological knowledge on which the plan is based and explain how features of that knowledge were used in the new design. This is application, not retelling. First describe the relevant theory, finding or previous research accurately. Then identify a feature and explain how you adapted it. For example, if previous research used a particular type of rating scale or experimental comparison, you might keep that feature because it provides a workable measure, or modify it because the new population or setting requires a different version. The mark comes from explaining the link.

Do not describe when the command word says explain

Cambridge's June 2024 examiner report says 'explain' questions often scored poorly because candidates described rather than justified. A reason question needs the consequence of the design decision. 'I used closed questions because they are easy to analyse' begins the explanation, but a stronger response says what kind of data they create and why that matters for the comparison in this study. The same applies to strengths and weaknesses. 'Opportunity sampling is biased' is a label. Explain who may be overrepresented, why that matters for this target population and how it may affect generalisability.

Click to enlarge

Use validity and reliability accurately

Reliability concerns consistency. Validity concerns whether the method or measure captures what it claims to measure and whether the conclusion is justified. A standardised procedure may improve reliability, but standardisation does not automatically make the study valid. Likewise, a realistic setting may improve ecological validity but can reduce control. Do not turn methodological terms into automatic strengths or weaknesses. Explain the trade-off created by the actual design.

Original example: planning a questionnaire

Imagine an original question asking you to plan a questionnaire investigating whether sleep habits are related to perceived academic stress. You could use a self-selecting sample of sixth-form students recruited through a school notice. Sleep could be measured using closed questions about average hours per night, while perceived stress could be measured with a short Likert-scale score. Standardise the instructions and define the scoring system in advance. Keep responses anonymous because stress may be sensitive. A weakness is that self-reported sleep and stress may be affected by social desirability or inaccurate recall. A strength of closed questions is that scores can be compared consistently across participants. Each point becomes creditworthy because it is applied to the actual investigation.

Original example: planning a structured observation

Suppose an original task asks you to investigate helping behaviour in a school common room using structured observation. Define categories before data collection, such as 'picks up another person's dropped item', 'offers directions' or 'assists with carrying an object'. Decide the observation period and location, train observers and use the same coding sheet. To check reliability, two observers can independently code the same period and compare agreement. A limitation is that participants may change their behaviour if they know they are being observed, depending on whether the observation is overt. Again, the method term matters less than the applied procedure.

Common research-method mistakes

1. Ignoring the method named in the question and designing a different study.
2. Missing one of the features the question explicitly tells you to include.
3. Naming a sampling technique without explaining how participants are actually selected.
4. Writing an independent or dependent variable without operationalising it.
5. Listing controls without saying how they are controlled.
6. Giving a generic questionnaire or interview plan without example questions or scoring where relevant.
7. Using vague observational categories that different observers could interpret differently.
8. Writing 'repeat it' or 'standardise it' without explaining what is repeated or standardised.
9. Adding ethics as labels rather than changing the procedure to deal with the issue.
10. Describing a strength, weakness or reason when the command word requires explanation.
11. Using validity and reliability as interchangeable terms.
12. Describing the previous psychological study without explaining how its features informed the new plan.

A reliable study-design routine

1. Underline the research method named in the question and every feature you must include.
2. State the aim and identify the variables or outcome the study needs to measure.
3. Operationalise the independent and dependent variables or method-specific categories.
4. Choose a sample and explain exactly how participants are obtained.
5. Write a replicable procedure in the order participants or researchers would follow it.
6. Add relevant controls, reliability steps and ethical safeguards.
7. State how the data will be recorded, scored or interpreted.
8. For Paper 4, link features of relevant psychological knowledge directly to the design.
9. For reason, strength and weakness questions, explain the consequence in this study rather than giving a generic methodological fact.
10. Before finishing, check that every named feature from the question is explicit on the page.

Put it into practice

Take one Psychology (9990) Paper 2 planning question and one Paper 4 study-design question. Before writing either answer, list the required features in the margin and tick them only when each one is explicit and applied. After marking, classify every lost mark as missing feature, weak operationalisation, generic method knowledge, poor application or unexplained justification. Inside NeuraGeek, you can use Psychology (9990) past-paper practice alongside the existing Paper 3 evaluation guide, keeping the skills separate: evaluation judges existing psychological evidence, while research-method questions ask you to build or justify the method itself.