AQA A-level Psychology (7182): Paper 2 – Research Methods – Study Guide

Research methods is the part of AQA A-level Psychology (7182) that teaches how psychologists design studies, collect and analyse data, and judge whether findings are trustworthy. It is compulsory content examined on Paper 2: Psychology in Context, and research methods skills are also assessed across all three papers. You need to know each method, the scientific processes behind good research, descriptive statistics, and how to choose and interpret inferential tests.

This guide follows the AQA AS and A-level Psychology (7181, 7182) specification, Version 1.4 (1 September 2025), for teaching from September 2025. Research methods is section 4.2.3 of the specification (with 4.2.3.1 Scientific processes, 4.2.3.2 Data handling and analysis and 4.2.3.3 Inferential testing) and topic 7 in the A-level subject content. Use it with the Research methods revision notes and Research methods practice questions. The other Paper 2 topics are in the Approaches in Psychology guide and the Biopsychology guide.

How is research methods examined?

Paper 2 is a 2-hour exam worth 96 marks and 33.3% of the A-level, assessing compulsory content 5 to 7. Its specification lists Section A and Section B at 24 marks each and Section C at 48 marks, all “multiple choice, short answer and extended writing”; Section C, the largest, follows Approaches and Biopsychology in the content order. The specification also states that knowledge and understanding of research methods, practical research skills and mathematical skills are assessed in every paper, and the specification’s mathematical requirements annex sets the maths you must be able to do.

Coverage table

Section	Content
4.2.3 Methods	Laboratory, field, natural and quasi-experiments; naturalistic, controlled, covert, overt, participant and non-participant observation; questionnaires and structured and unstructured interviews; correlations; content analysis; case studies
4.2.3.1 Scientific processes	Aims and hypotheses; sampling; pilot studies; experimental designs; observational design; questionnaire and interview design; variables; control; demand characteristics and investigator effects; ethics and the BPS code; peer review; implications for the economy; reliability; validity; features of science; reporting investigations
4.2.3.2 Data handling and analysis	Quantitative and qualitative data; primary and secondary data and meta-analysis; mean, median, mode, range and standard deviation; percentages; correlations; graphs and tables; normal and skewed distributions; correlation coefficients; levels of measurement; content analysis, coding and thematic analysis
4.2.3.3 Inferential testing	The sign test; probability and significance; critical values; Type I and Type II errors; choosing between Spearman’s rho, Pearson’s r, Wilcoxon, Mann-Whitney, related t-test, unrelated t-test and Chi-squared

Which research methods do you need to know?

Experiments. An experiment manipulates an independent variable (IV) and measures its effect on a dependent variable (DV), while controlling other variables.

Type	Definition	Strength	Limitation
Laboratory	IV manipulated in a controlled, artificial setting	High control; cause and effect; easy to replicate	Low ecological validity; demand characteristics
Field	IV manipulated in a natural setting	Higher ecological validity; participants may be unaware	Less control; ethical issues of consent
Natural	IV varies naturally (e.g. a policy change); researcher records the effect	Studies variables that could not ethically be manipulated	No random allocation; event may be rare
Quasi	IV is a characteristic of participants (e.g. age, gender)	Allows comparison of existing groups	Participant variables confound; no random allocation

Observation. Naturalistic observation happens where the behaviour normally occurs; controlled observation sets up some conditions. Covert observation is hidden; overt observation is known to participants. In participant observation the researcher joins the group; in non-participant observation they watch from outside. Covert observation reduces participant reactivity but raises consent issues; participant observation gives insight but risks losing objectivity.

Self-report. Questionnaires use written questions; they reach many people quickly but can suffer social desirability bias. Structured interviews use fixed questions in a fixed order, which makes them easy to compare. Unstructured interviews follow the conversation, which gives rich detail but makes analysis harder and depends on the interviewer’s skill.

Correlations. A correlation measures the relationship between two co-variables. Unlike an experiment, nothing is manipulated, so correlation cannot show cause and effect; a third variable may explain the link.

Content analysis turns qualitative material (texts, media, interview transcripts) into data by coding it into categories. Case studies investigate one person, group or event in depth, often over time; they give rich, idiographic data but cannot easily be generalised.

What are the scientific processes?

Aims and hypotheses. An aim is a general statement of what the study investigates. A hypothesis is a precise, testable prediction. A directional hypothesis states the direction of the difference or relationship (“participants who sleep 8 hours will recall more words than those who sleep 4 hours”); a non-directional hypothesis predicts a difference without stating which way. Directional hypotheses are usually used when previous research suggests the outcome.

Sampling. The population is the group the researcher wants to study; the sample is the people who actually take part.

Technique	How it works	Main issue
Random	Every member of the population has an equal chance (e.g. random number generator)	Needs a full list; may still be unrepresentative by chance
Systematic	Every nth person from a list	Unbiased if the list has no pattern
Stratified	Sub-groups sampled in proportion to the population	Representative but time-consuming
Opportunity	Whoever is available	Quick but biased
Volunteer	People who respond to an advert	Easy, but volunteer bias

Bias limits generalisation from sample to population.

Pilot studies are small-scale trial runs that check procedures, materials and timings, so problems can be fixed before the main study.

Experimental designs. Repeated measures: each participant does every condition; controls participant variables but risks order effects, handled by counterbalancing. Independent groups: different participants in each condition; no order effects but participant variables can confound, reduced by random allocation. Matched pairs: participants are matched on a key variable and one of each pair goes into each condition; controls key participant variables but matching is time-consuming.

Observational design. Behavioural categories break behaviour into observable, measurable units. Event sampling records every time a behaviour occurs; time sampling records behaviour at set intervals.

Questionnaire and interview design. Open questions give qualitative answers in the participant’s own words; closed questions give fixed choices and quantitative data. Good design avoids leading, double-barrelled and jargon-filled questions. Interviews need a schedule, a suitable setting and a plan for recording.

Variables. The IV is manipulated; the DV is measured. Extraneous variables are any other variables that might affect the DV; when they do affect it systematically, they become confounding variables. Operationalisation means defining variables in measurable terms, such as “number of words recalled out of 20”.

Control. Random allocation to conditions, counterbalancing (e.g. ABBA), randomisation of materials or order and standardisation (same instructions and procedures for all participants) reduce bias.

Demand characteristics are cues that let participants guess the aim and change their behaviour. Investigator effects are any influence of the researcher on the result, such as tone of voice. Single-blind and double-blind procedures reduce them.

Ethics. The British Psychological Society (BPS) code of ethics guides research. Key issues are informed consent, deception, protection from harm, privacy and confidentiality, and the right to withdraw. Ways of dealing with them include presumptive or prior general consent, full debriefing, cost–benefit judgements by an ethics committee, and anonymising data.

Peer review is the assessment of a study by other experts before publication. It checks validity and quality and prevents fraudulent or poor research being published, but reviewers may be biased towards findings that fit current theories, and positive results are more likely to be published (publication bias).

Implications for the economy. Research can have economic benefits, such as treatments that help people with mental health problems return to work, or findings that reduce absence or improve productivity.

Reliability is consistency. Test–retest reliability gives the same test to the same people on two occasions and correlates the scores. Inter-observer reliability compares two observers’ records; a correlation of +0.80 or above is usually taken as good. Reliability can be improved by clearer behavioural categories, observer training and standardised procedures.

Validity is whether a study measures what it claims to measure. Face validity: the measure looks as if it measures the right thing. Concurrent validity: results match those of an established measure. Ecological validity: findings generalise to real-life settings. Temporal validity: findings remain true over time. Validity is improved by control groups, standardisation, blind procedures, anonymous questionnaires and triangulation.

Features of science. Science relies on objectivity and the empirical method (knowledge from direct observation and experiment). It requires replicability (repeating studies to check findings) and falsifiability (Popper: a theory must be capable of being proved wrong). Theory construction and hypothesis testing move between theory and evidence. Kuhn argued that a science has a shared paradigm, and that a paradigm shift happens when contradictory evidence builds up enough to replace it.

Reporting. A scientific report has an abstract (short summary), introduction (literature review and hypotheses), method (design, sample, materials, procedure, ethics), results (descriptive and inferential statistics), discussion (interpretation, limitations, implications) and referencing (sources in a standard format).

How do you handle and analyse data?

Types of data. Quantitative data are numerical; qualitative data are descriptive (words, meanings). Primary data are collected first-hand for the study; secondary data were collected by someone else. A meta-analysis combines the results of many studies to find an overall effect size.

Descriptive statistics. The mean adds all scores and divides by the number of scores; it uses all data but is distorted by extreme values. The median is the middle value in order; it is not affected by outliers. The mode is the most frequent value. The range is the highest minus the lowest score (often plus 1 in some conventions; follow your teacher’s guidance); it is easy to calculate but affected by extremes. The standard deviation measures average spread around the mean; a larger value means more spread. You must be able to calculate mean, median, mode, range and percentages.

Displaying data. Tables summarise; bar charts show discrete categories with gaps between bars; histograms show continuous data with no gaps; scattergrams show correlations; line graphs show change over time.

Distributions. In a normal distribution, the curve is symmetrical and the mean, median and mode are at the same point. In a positive skew, most scores are low with a long tail to the right, and the mean is higher than the median, which is higher than the mode. In a negative skew, most scores are high with a tail to the left, and the mean is lowest.

Correlation coefficients range from -1 (perfect negative) through 0 (no correlation) to +1 (perfect positive). The size shows strength; the sign shows direction.

Levels of measurement. Nominal data are categories (counts in groups). Ordinal data can be ranked but intervals are not equal (ratings). Interval data use units of equal size (time, temperature).

Content analysis, coding and thematic analysis. In coding, the researcher develops categories and counts how often each occurs, producing quantitative data. In thematic analysis, the researcher identifies recurring themes and keeps the data qualitative.

How does inferential testing work?

The sign test is used for a difference, with related data (repeated measures or matched pairs) at nominal level. To calculate it: record a + or - for each participant’s change, ignore any zero differences, count the less frequent sign (this is S), and compare S with the critical value for the number of non-zero differences (N). If S is equal to or less than the critical value, the result is significant.

Probability and significance. Psychologists usually use a significance level of p ≤ 0.05: a 5% or lower probability that the result occurred by chance if the null hypothesis is true. A stricter level such as 0.01 is used when a mistake would be costly. Statistical tables give critical values; you need to know whether the hypothesis is one-tailed (directional) or two-tailed (non-directional), N (or degrees of freedom) and the significance level. For some tests the calculated value must be equal to or greater than the critical value; for others (including the sign test, Wilcoxon and Mann-Whitney) it must be equal to or less.

Type I and Type II errors. A Type I error is a false positive: rejecting the null hypothesis when it is true, more likely with a lenient level such as 0.10. A Type II error is a false negative: accepting the null hypothesis when it is false, more likely with a strict level such as 0.01.

Choosing a test. The specification states that the choice depends on factors including level of measurement and experimental design.

	Unrelated design (independent groups)	Related design (repeated measures, matched pairs)	Correlation
Nominal	Chi-squared	Sign test	Chi-squared
Ordinal	Mann-Whitney	Wilcoxon	Spearman’s rho
Interval	Unrelated t-test	Related t-test	Pearson’s r

The three t-tests and Pearson’s r are parametric tests: they need interval data, a roughly normal distribution and similar variance between groups.

Common errors

• Writing hypotheses that are not operationalised (“memory will be better”).
• Confusing opportunity and volunteer sampling.
• Saying counterbalancing removes order effects; it balances them across conditions.
• Confusing ecological validity with generalisability to other people.
• Using a bar chart for continuous data, or a histogram for categories.
• Forgetting that for the sign test, Wilcoxon and Mann-Whitney the calculated value must be equal to or less than the critical value.

Frequently asked questions

What is the difference between reliability and validity? Reliability is consistency: getting the same result again under the same conditions. Validity is accuracy: whether the study measures what it claims to and whether results apply beyond it. A measure can be reliable without being valid.

When do you use a directional hypothesis? When previous research suggests the direction of the result. Otherwise use a non-directional hypothesis, which affects whether you read one-tailed or two-tailed critical values.

How do I remember which statistical test to use? Ask three questions: is it a difference or a correlation; is the design related or unrelated; and is the data nominal, ordinal or interval? The answers point to one cell of the table above.

Is research methods only on Paper 2? Its content is listed under Paper 2, but the specification states research methods, practical skills and maths are assessed in all papers, so expect methods questions on Papers 1 and 3 too.

Where to go next

Condense this guide with the Research methods revision notes and practise with the Research methods practice questions. See the AQA A-level Psychology hub and the AQA A-level Psychology checklist.

Official syllabus

AQA AS and A-level Psychology (7181, 7182) specification, Version 1.4 (1 September 2025), section 4.2.3 Research methods (4.2.3.1 to 4.2.3.3) and the mathematical requirements annex — aqa.org.uk.