In simple terms
A friendly intro before the formal notes — no formulas yet.
The Referee's Dilemma: Making the Right Call
Diagnosing a mental health condition is not a simple checklist — it is a judgement made by a human being working with imperfect definitions and limited information. Two skilled clinicians can look at the same person and reach different conclusions, and that fact sits at the heart of the whole topic.
Picture a referee in a fast football match. They have an official rulebook (the classification system, like the DSM-5), but they still have to interpret a split-second tackle: was that a deliberate foul or an honest slip? Their past experience (bias), the roar of the home crowd (social and cultural norms) and pressure from the touchline can all nudge the decision. Two referees can watch the same tackle and give different cards — one yellow, one red. If they usually agree, the refereeing is RELIABLE; if their calls actually match what really happened on the pitch, it is VALID. A diagnosis works the same way, and reliability and validity are not the same thing.
- 1
First, ask what counts as 'abnormal' at all — every definition of abnormality (statistical, social, functional, ideal) captures some cases and wrongly flags or misses others.
- 2
Next, learn the 'rulebook': classification systems such as the DSM-5 and ICD-11 exist to give clinicians shared criteria and a shared language.
- 3
Then separate two very different questions. Reliability = would other clinicians agree (consistency)? Validity = is the diagnosis actually correct and useful (accuracy)?
- 4
Finally, spot the 'referee' factors that push a diagnosis off course: confirmation bias, cultural bias, gender bias and the stubborn stickiness of a label once it is applied — then weigh the ethical costs and benefits of diagnosing at all.
Explore the concept
Use the live diagram and synced steps — play it or tap a step card to walk through.
Full topic notes
Formal explanation with the rigour you need for the exam.
Defining normality and abnormality — and why it is hard
There is no single agreed definition of abnormality, and the IB wants you to understand why. Each common approach captures part of the truth and then fails at the edges. Knowing the failure of each definition is worth more marks than knowing the definitions themselves, because it shows you understand the problem diagnosis is trying to solve.
The take-home point is that abnormality is a fuzzy, value-laden construct. Because clinicians cannot rely on any one clean definition, they use classification systems that operationalise disorders as lists of specific criteria — which is where diagnosis proper begins.
Statistical infrequency — behaviour is abnormal if it is statistically rare. Problem: rarity is not the same as disorder. A very high IQ is rare but desirable, while depression is common yet clearly a disorder. The approach also cannot say where on the distribution 'abnormal' begins.
Deviation from social norms — behaviour that breaks a society's rules is judged abnormal. Problem: norms are relative to culture and to their era. Homosexuality was classified as a disorder in the DSM until 1973; this approach risks pathologising difference and enforcing conformity.
Failure to function adequately — abnormality as an inability to cope with daily living (self-care, work, relationships). Problem: 'adequate' is subjective, some people function while deeply distressed, and some unconventional but harmless lifestyles would be wrongly flagged.
Deviation from ideal mental health (Jahoda) — abnormality as the absence of features such as self-actualisation, autonomy and accurate perception of reality. Problem: by such demanding criteria almost everyone is 'abnormal' some of the time, and the ideals are culturally loaded toward Western individualism.
The purpose of classification systems: DSM-5 and ICD-11
A classification system is a manual that turns vague ideas of disorder into concrete, checkable criteria: which symptoms, how many, how severe, and for how long. Its purpose is not to be the final truth about the mind but to give clinicians a shared language and shared thresholds so that a 'major depressive episode' means roughly the same thing in two different clinics. This standardisation is what makes reliable communication, comparable research and evidence-based treatment possible.
DSM-5 — the Diagnostic and Statistical Manual (5th edition, 2013; text-revised as DSM-5-TR, 2022), published by the American Psychiatric Association. Covers only mental disorders and dominates US practice and research.
ICD-11 — the International Classification of Diseases (11th revision, in effect from 2022), published by the World Health Organization. Covers all diseases and is used globally for health statistics; the two systems have been progressively harmonised but still differ in some criteria.
Categorical vs dimensional — both systems remain largely categorical (you either meet the threshold or you do not). Critics argue many conditions are dimensional (on a spectrum), and a strict cut-off can misclassify borderline cases.
Purpose, restated — to raise reliability, aid communication between professionals, guide treatment, and enable research. Judge the systems by how well they achieve these aims, not against an impossible standard of perfection.
Reliability of diagnosis (inter-rater reliability)
Reliability means consistency. The most important form for diagnosis is inter-rater reliability: if two clinicians independently assess the same patient, do they reach the same diagnosis? Consistency is often quantified with Cohen's kappa, where 1.0 is perfect agreement and 0 is chance-level agreement. Reliability matters because an inconsistent diagnosis is, by definition, partly a product of the clinician rather than the patient — and an unreliable diagnosis can never be valid.
Validity of diagnosis
Validity means accuracy and usefulness: does a diagnosis correctly identify a genuine disorder, distinguish it from other conditions, and predict its course and response to treatment? Validity is harder to establish than reliability because there is no independent 'gold standard' test for most mental disorders — we cannot open the box and check. Several lines of evidence are used to question validity.
Descriptive/labelling evidence — if a label changes how all later behaviour is read (Rosenhan, 1973), the label may reflect the context more than the person, threatening validity.
Comorbidity — disorders that are supposed to be distinct co-occur very often (e.g. depression and anxiety), suggesting the categories may not carve nature at its joints.
Predictive validity — a valid diagnosis should predict outcome and treatment response; where the same label leads to very different courses and treatments, its predictive validity is weak.
Reliability is necessary but not sufficient — clinicians can consistently apply a category (reliable) that does not correspond to a real, distinct disorder (invalid). Never treat high reliability as proof of validity.
Biases in diagnosis
Clinicians are human information-processors, and predictable biases can distort diagnosis. The IB names several; the skill is to explain the mechanism, tie it to a study, and then evaluate rather than just assert that bias exists.
Evaluation is essential here. Vignette studies like Ford and Widiger have strong internal validity (only sex was manipulated, so the difference must be caused by it) but lower ecological validity, because reading a vignette is not the same as assessing a real patient over time. Observational patterns like ethnic differences in diagnosis rates are real but harder to interpret, because true prevalence, help-seeking, social adversity and clinician bias are confounded. Good answers weigh these evidence-quality points rather than treating 'bias exists' as settled.
Confirmation bias — once a clinician forms an initial hypothesis they may seek and weight confirming information and discount the rest. Rosenhan (1973) illustrates this at the level of a whole institution: the label set the expectation and behaviour was read to fit it.
Cultural bias — behaviour normal within a patient's culture can be misread as symptomatic, and criteria developed on one population may not transfer. Two mechanisms matter: clinician bias (the diagnoser's assumptions) and reporting bias (how distress is expressed, e.g. somatisation, differs across cultures). Cochrane and others documented higher rates of schizophrenia diagnoses among African-Caribbean patients in the UK than would be predicted, a pattern widely attributed in part to cultural misinterpretation rather than true prevalence differences.
Gender (sex) bias — the same presentation can be diagnosed differently by gender. Broverman et al. (1970) found clinicians equated 'mentally healthy adult' with 'mentally healthy man', implying women faced a double standard. Ford and Widiger (1989) gave clinicians identical case vignettes altering only the patient's sex: they more often diagnosed histrionic personality disorder when the patient was described as female and antisocial personality disorder when described as male — direct experimental evidence of gender bias.
The effect of labelling — beyond the moment of diagnosis, a label can create stigma and a self-fulfilling prophecy, and it is 'sticky', shaping how the patient is treated and even how they see themselves (Rosenhan, 1973; Scheff's labelling theory).
Ethical and social implications of diagnosis
Diagnosis is not a neutral act of description; it has real consequences, and a strong Paper 2 answer weighs them on both sides rather than declaring diagnosis simply harmful or simply helpful.
Benefits — a valid diagnosis opens access to treatment, services and support; it gives patients and families an explanation that can reduce self-blame; and it enables research on comparable groups and fair allocation of resources.
Stigma and discrimination — a psychiatric label can bring social rejection, disadvantage in work and relationships, and internalised shame, sometimes outlasting the condition itself.
Self-fulfilling prophecy and the sick role — expecting to be 'ill' can shape behaviour and identity, and others may treat the person only through the lens of the label (Rosenhan, 1973).
Medicalisation and over-diagnosis — expanding criteria can turn normal distress or ordinary variation into disorder, raising questions about who benefits and who decides.
Cultural justice — because criteria and norms are largely Western in origin, applying them across cultures can pathologise difference; culturally sensitive assessment is an ethical requirement, not an optional extra.
For Paper 2, never just name a study — state its aim, procedure and finding briefly, then USE it: say what it shows about reliability, validity or bias, and how good that evidence is. Match the study to the claim (Rosenhan for validity/labelling; Cooper et al. or kappa data for reliability; Ford and Widiger for gender bias). And treat 'discuss' as a demand for balance: considerations for and against, developed, not a one-sided list.
Common mistakes examiners penalise
Confusing reliability with validity — the number-one error. Reliability = agreement/consistency; validity = correctness/usefulness. Evidence about one does not settle the other, and 'reliable but invalid' is a real and examinable possibility.
Describing studies instead of USING them — three neat paragraphs on Rosenhan's procedure with no link to a claim earns knowledge marks (B) but not use-of-research marks (C). Every study must do a job in the argument.
Treating 'discuss' as 'list problems' — a discuss question demands balance. An answer that only attacks diagnosis, or only defends it, caps critical-thinking marks (D) because it is one-sided.
Asserting bias exists without evaluating the evidence — saying 'there is cultural bias' is weak; explaining the mechanism, citing a study, and weighing confounds (true prevalence vs clinician bias) is strong.
Using outdated studies uncritically — Rosenhan (1973) and Cooper et al. (1972) are valuable but decades old; the top band acknowledges that criteria have since been tightened rather than treating old findings as the current state of diagnosis.
No conclusion, or a conclusion that just repeats the introduction — the essay must reach a reasoned, evidence-based judgement that follows from the argument built in the body.
Worked examples
See the formulas applied — reveal one step at a time, like the exam.
WORKED EXAMPLE 1 — Cooper et al. (1972), the US–UK Diagnostic Project, showed the same videotaped clinical interviews to psychiatrists in New York and London. New York psychiatrists were roughly twice as likely to diagnose schizophrenia, while London psychiatrists were roughly twice as likely to diagnose mania or depression. Using your knowledge, explain what this demonstrates about diagnosis — and state precisely what it does NOT show.
- 1
Step 1 — Identify the concept. This is evidence about the inter-rater reliability of diagnosis.
WORKED EXAMPLE 2 — In Rosenhan's (1973) 'On being sane in insane places', eight healthy 'pseudopatients' each reported a single symptom (hearing an unclear voice saying 'thud', 'empty', 'hollow') to gain admission to psychiatric hospitals. Once admitted they behaved normally. All were admitted (mostly diagnosed with schizophrenia), and normal behaviours — such as taking notes — were recorded as pathological ('patient engages in writing behaviour'). Explain how this challenges the VALIDITY of diagnosis, and note one methodological limitation.
- 1
Step 1 — Identify the concepts. The findings bear on validity and on the effect of labelling (with confirmation bias as the mechanism).
Paper 2, extended response: Discuss the reliability and validity of the diagnosis of psychological disorders. [22 marks]
- 1
MODEL ESSAY
How it all connects
The big idea sits in the middle — tap a linked idea to explore the link.
Tap a linked idea to see how it connects back to the main topic — that connection is what examiners reward.
Glossary
Try to recall each definition before you reveal it.
Quick check
Answer in your head first — then tap to check. No pressure.
Revision flashcards
Flip the card. Test yourself before the exam.
Abnormality (the problem)
There is no single agreed definition. The IB expects you to know several approaches — statistical infrequency, deviation from social norms, failure to function adequately, and deviation from ideal mental health — and to explain the weakness of each.
Key takeaways
Review these before you close the topic — retrieval beats re-reading.
- ✓
Statistical infrequency — behaviour is abnormal if it is statistically rare. Problem: rarity is not the same as disorder. A very high IQ is rare but desirable, while depression is common yet clearly a disorder. The approach also cannot say where on the distribution 'abnormal' begins.
- ✓
Deviation from social norms — behaviour that breaks a society's rules is judged abnormal. Problem: norms are relative to culture and to their era. Homosexuality was classified as a disorder in the DSM until 1973; this approach risks pathologising difference and enforcing conformity.
- ✓
Failure to function adequately — abnormality as an inability to cope with daily living (self-care, work, relationships). Problem: 'adequate' is subjective, some people function while deeply distressed, and some unconventional but harmless lifestyles would be wrongly flagged.
- ✓
Deviation from ideal mental health (Jahoda) — abnormality as the absence of features such as self-actualisation, autonomy and accurate perception of reality. Problem: by such demanding criteria almost everyone is 'abnormal' some of the time, and the ideals are culturally loaded toward Western individualism.
Practice — then mark it
The whole point: a real Cambridge question, marked mark-by-mark.
Get a Paper 2 essay marked: discuss the reliability and validity of diagnosis
Get a Paper 2 essay marked: discuss the reliability and validity of diagnosis
Extra simulations & links
PhET, GeoGebra and other curated tools — open in a new tab.
Frequently asked
Checkpoint
One marked question is worth ten re-reads — close the loop before you move on.
Reading it isn’t knowing it — prove it.
Before you move on: do Get a Paper 2 essay marked: discuss the reliability and validity of diagnosis on paper, snap a photo, and get examiner-style feedback on exactly where you win and lose marks.