You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
This is a skill lesson rather than a content lesson. Its job is to teach you how to critically review and evaluate a study — the core competence that Edexcel 9PS0 Paper 3 tests when it presents you with an unfamiliar piece of research, or asks you to evaluate one of the classic and contemporary studies you have met across the course. Reviewing a study is not a matter of remembering a list of "good things" and "bad things"; it is a disciplined interrogation of a study's aim, method, sample, findings, reliability, validity, ethics and generalisability, in which each judgement is justified by reference to how the study was actually conducted and followed through to its consequences for what the study can and cannot tell us. The reason this skill matters so much is that it is transferable: once you can dismantle a study systematically, you can evaluate any study in the specification — Milgram, Baddeley, Raine, Watson and Rayner, Rosenhan — and you can respond to novel research you have never seen before, which is exactly what the highest-tariff Paper 3 items demand.
By the end of this lesson you should be able to work through a structured review framework for any study; define and correctly apply reliability, validity, generalisability and the ethical criteria; distinguish evaluation of a study's method from evaluation of its findings; apply methodological evaluation to the classic and contemporary studies met across the Edexcel course; and write a developed critical review that reaches a justified overall judgement.
Key Definition: To evaluate a study is to make a justified judgement about the trustworthiness and value of its conclusions, by examining how well its design supports the claims made from it — never merely to describe what was done.
Edexcel 9PS0 — Paper 3: Psychological Skills. This lesson develops the methodological review skills that Paper 3 rewards across the specification, applied both to novel research and to the named classic and contemporary studies. The skill is the deliberate management of the assessment objectives: describing a study accurately (AO1), applying evaluation criteria to its specific features (AO2), and building a developed critical judgement of its reliability, validity, ethics and generalisability (AO3).
| Assessment Objective | What it looks like on this skill |
|---|---|
| AO1 — knowledge & understanding | Accurately outlining a study's aim, method, sample and findings; defining reliability, validity, generalisability and ethical criteria. |
| AO2 — application | Applying each evaluation criterion to the specific features of the study under review — its actual sample, its actual controls. |
| AO3 — analysis & evaluation | Judging how far the design supports the conclusions; weighing strengths against limitations; reaching a justified overall verdict. |
Connects to…
A recurring examiner theme is that an evaluation is only worth marks when it is tied to the specific study and developed to a consequence: "the sample was biased" earns little, while "the all-male sample means the findings may not generalise to women, so the general claim about obedience is really a claim about male obedience" earns the AO3.
A reliable review interrogates a study under eight headings, in a logical order that moves from what the study did to whether we should believe it.
graph TD
A["1. Aim & hypothesis<br/>What was the question?"] --> B["2. Method & design<br/>How was it tested?"]
B --> C["3. Sample<br/>Who took part?"]
C --> D["4. Findings & conclusions<br/>What was claimed?"]
D --> E["5. Reliability<br/>Is it consistent / replicable?"]
E --> F["6. Validity<br/>Does it measure what it claims?"]
F --> G["7. Ethics<br/>Was it conducted responsibly?"]
G --> H["8. Generalisability<br/>Who do the findings apply to?"]
H --> I["Justified overall judgement"]
style I fill:#27ae60,color:#fff
The discipline is to ask a question at each heading and then draw a consequence, rather than simply labelling the study. The sections below take each heading in turn.
Begin by identifying what the study set out to discover and, where relevant, its directional or non-directional hypothesis. This matters for evaluation because a study can only be judged against its own aim: a finding that seems limited may perfectly answer the narrow question posed, while an over-broad conclusion may exceed what the aim licensed. Watson and Rayner (1920) aimed to show that an emotional response — fear — could be conditioned in a human infant; judging the study against that specific aim keeps the evaluation focused.
Identify the method (laboratory experiment, field experiment, natural experiment, observation, self-report, case study, correlation) and, for experiments, the design (independent groups, repeated measures, matched pairs) and the operationalised variables. Each method carries a characteristic trade-off that drives the evaluation:
| Method | Characteristic strength | Characteristic limitation |
|---|---|---|
| Laboratory experiment | High control → cause-and-effect; replicable | Artificial → low ecological validity; demand characteristics |
| Field experiment | Higher ecological validity | Less control of extraneous variables; ethical issues (consent) |
| Natural experiment | Studies variables that cannot be manipulated ethically | No random allocation → confounding; rare, hard to replicate |
| Observation | Captures natural behaviour | Observer bias; no causal inference |
| Self-report (questionnaire / interview) | Access to thoughts and feelings; large samples possible | Social desirability; demand characteristics |
| Case study | Rich, in-depth, studies rare cases | Not generalisable; researcher interpretation bias |
Examine who took part — the sampling method (random, opportunity, volunteer, stratified), the size, and the characteristics (age, sex, culture). The evaluative payoff comes from asking whether the sample supports the breadth of the conclusion. Milgram's original sample of 40 American men supports a claim about the obedience of American men; extending it to "people" imports an assumption the sample does not justify — the classic route into a generalisability critique.
Separate the findings (the actual data) from the conclusions (the interpretation drawn from them). A frequent flaw — and therefore a rich source of evaluation — is over-interpretation: a conclusion that outruns the data. Distinguishing the two lets you accept a study's data while challenging its interpretation, which is a more sophisticated move than rejecting the whole study.
Key Definition: Reliability is the consistency of a measure or procedure. A study is reliable if the same procedure produces the same results when repeated (external reliability, checked by replication) and if a measure is internally consistent (internal reliability).
Reliability is evaluated by asking whether the procedure is standardised enough to be repeated. A tightly controlled laboratory experiment (Baddeley, 1966) is highly replicable; a one-off case study or a naturalistic observation is far harder to repeat identically. Inter-rater reliability — the extent to which two independent observers agree — is the relevant form for observational studies. High reliability matters because a finding that cannot be reproduced cannot be trusted, which is exactly the concern the replication problem raises for the whole discipline.
Key Definition: Validity is the extent to which a study measures or reflects what it claims to. Internal validity concerns whether the effect was genuinely caused by the independent variable (rather than a confound); external validity concerns whether the findings generalise beyond the study — across settings (ecological validity), people (population validity) and time (temporal validity).
Validity is often the most productive evaluation heading, because it can be attacked from several angles:
| Threat to validity | Question to ask | Illustration |
|---|---|---|
| Confounding variables | Could something other than the IV explain the result? | A natural experiment with no random allocation |
| Demand characteristics | Did participants guess the aim and change their behaviour? | Any lab study where the purpose is transparent |
| Investigator effects | Did the researcher unintentionally influence the result? | Interviewer tone shaping self-report answers |
| Ecological validity | Would the behaviour occur outside the artificial setting? | Learning word lists (Baddeley) versus everyday memory |
| Social desirability | Did participants answer to look good? | Self-report of sensitive attitudes or behaviour |
Evaluate the study against the ethical criteria: informed consent, deception, the right to withdraw, protection from harm (physical and psychological), confidentiality and debriefing. Ethical evaluation is not merely moral bookkeeping — it connects to validity, because the measures used to protect participants (such as gaining fully informed consent) can themselves introduce demand characteristics, and the deception used to preserve validity raises the ethical cost. Milgram's study is the standard case: its deception and the psychological distress it caused are the central ethical criticisms, though Milgram's own defence (thorough debriefing, participants' retrospective consent) is part of a balanced evaluation.
Key Definition: Generalisability is the extent to which the findings from a study's sample can be applied to the wider population and to other settings and times.
Generalisability draws the sample and external-validity threads together into the final question: to whom, where and when do these findings actually apply? A study high in internal validity but conducted on a narrow, unrepresentative sample in an artificial setting may support a conclusion that is true but local. This is where a review reaches its overall judgement — not "good study / bad study," but a calibrated statement of what the study can and cannot establish.
The point of the framework is that it works on every study. The table below shows the single most productive evaluation line for a range of the Edexcel classic and contemporary studies — but a full review would develop several headings for each.
| Study (topic) | Method | Standout evaluation line |
|---|---|---|
| Milgram (1963) — obedience (Social) | Controlled observation / lab | Ethics (deception, distress) and low population validity (40 US men) versus high internal control |
| Baddeley (1966) — coding in memory (Cognitive) | Laboratory experiment | High reliability and internal validity, but artificial word-list task → low ecological validity |
| Raine et al. (1997) — brains of offenders (Biological) | Quasi/natural experiment (PET scans) | Objective biological measure and control group, but correlational → cannot infer that brain differences cause violence |
| Watson & Rayner (1920) — Little Albert (Learning) | Controlled observation (case study) | Demonstrates conditioning of fear, but N=1 → not generalisable, and serious ethical concerns (harm, no protection) |
| Rosenhan (1973) — being sane in insane places (Clinical) | Covert participant observation (field) | High ecological validity and a powerful challenge to diagnosis, but deception of staff and questionable replicability |
Two disciplines make this table into marks. First, tie the criterion to the specific feature: not "Baddeley lacks ecological validity" but "learning lists of unrelated words is unlike everyday memory for meaningful material, so the coding findings may not extend to real-world remembering." Second, balance the evaluation: Milgram's ethics are a serious criticism, but the study's high level of control and its striking, replicated findings are genuine strengths — a review that only prosecutes, or only defends, is weaker than one that weighs.
A subtle but high-value distinction separates two things students often blur: evaluating the method and evaluating the findings/conclusions. They are different activities with different targets.
Evaluating the method asks whether the design was sound — was the sample adequate, were variables controlled, was the measure valid and reliable? Evaluating the findings asks whether the conclusion drawn is justified by the data even granting that the method was fine — is the interpretation the only one available, does it over-reach, does it fit other evidence? Raine et al. (1997) illustrates the difference cleanly: the method can be praised (objective PET imaging, a matched control group) while the conclusion is challenged (because the design is correlational, the finding that offenders showed reduced prefrontal activity cannot establish that the brain difference caused the violence rather than resulting from it or from a third factor). The strongest reviews separate these strands explicitly, because it lets you say "the study is methodologically strong but its causal conclusion outruns its correlational design" — a far more precise judgement than a blanket verdict.
The weakest reviews end with a bare summary ("so there are strengths and weaknesses"). A strong review reaches a calibrated conclusion that states what the study establishes and what it does not, and why. The move is to weigh the most consequential strengths against the most consequential limitations and to say where the balance falls for the specific claim being made. For Milgram, a substantiated conclusion runs: "Milgram's high control and replicated findings establish that situational pressure can produce startlingly high obedience — a robust and important effect; however, the narrow sample means the specific rate should not be read as a universal human constant, and the ethical costs, while partly mitigated by debriefing, are real. The study is therefore best read as a compelling demonstration of a general phenomenon whose precise parameters and cross-cultural generality require the further research it inspired." That conclusion is earned from the evaluation, it is specific to the study, and it calibrates rather than simply rejecting or accepting — which is exactly the top-band move.
It is worth seeing the framework run end-to-end on a single famous study, because the value of the method only becomes clear when every heading is developed rather than listed. Take Milgram's (1963) obedience study.
Aim and hypothesis. Milgram set out to investigate whether ordinary people would obey an authority figure's instructions to administer what they believed were harmful electric shocks to another person — a response to the post-war question of how atrocities are carried out by seemingly ordinary individuals. Judging the study against this aim matters: the study is a test of situational obedience, so evaluations that fault it for ignoring dispositional factors are partly answering a question Milgram did not pose.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.