Unit 4: Clinical psychology A2 and psychological skills

Definitions of abnormality have moved from supernatural and moral explanations, through medical accounts, towards judgments that combine evidence about behaviour, distress and everyday functioning.

Earlier supernatural accounts could lead to punishment or exclusion. Medical classification reframed mental disorder as illness and encouraged treatment, while later psychological and social approaches emphasised cognition, learning, relationships, rights and individual experience. Each change can reduce one form of stigma yet create new labels.

Classification systems organise symptoms into diagnostic categories so clinicians and researchers can communicate, select support and compare outcomes.

Revisions respond to evidence and changing concepts. DSM-5, for example, combined several former autism diagnoses into autism spectrum disorder and added cultural-formulation guidance. ICD is designed for multilingual, international health reporting. Both use recognisable symptom patterns, but categories and thresholds can differ across manuals or editions.

A diagnosis is trustworthy only when clinicians apply criteria consistently and the resulting category accurately represents the person's difficulties in their cultural context.

Culture can change how distress is expressed, whether hearing a voice has spiritual or clinical meaning, and how a clinician interprets language. Cooper et al. found New York psychiatrists more likely than London psychiatrists to diagnose schizophrenia from the same filmed interviews, illustrating possible system or cultural bias.

Schizophrenia involves disturbances of perception, belief and thought organisation; biological explanations identify risk mechanisms rather than a single inevitable cause.

The dopamine account links excess mesolimbic dopamine activity with positive symptoms and reduced mesocortical activity with negative or cognitive symptoms. Glutamate dysfunction at NMDA receptors may alter dopamine regulation and explain a wider symptom range. Drug and imaging evidence supports involvement, but changes measured after illness or treatment do not settle causal direction.

The syllabus permits unipolar depression or anorexia nervosa as the other disorder; each requires symptoms plus one biological and one non-biological explanation.

Antidepressant response and twin evidence support biological involvement in depression; cognitive bias and CBT response support a cognitive pathway. Twin evidence supports anorexia vulnerability, while Becker's Fijian television study supports a cultural-learning influence.

Drug therapy aims to reduce symptoms biologically, while family therapy changes the social environment and the family's ability to support recovery; they are often combined.

Risperidone and other antipsychotics have reduced positive and negative symptoms in treatment studies. Reviews of family interventions report improved social functioning and lower relapse; high criticism, hostility or emotional over-involvement is associated with greater relapse risk.

For the optional disorder, drug therapy targets biological symptoms while CBT identifies and tests thought-behaviour patterns that maintain distress.

Combined antidepressant and CBT treatment can help some people with persistent depression. Routine-practice evidence from Hans and Hiller supports outpatient CBT but also shows meaningful dropout. For anorexia, CBT can improve beliefs and relapse outcomes, yet physical risk may first require coordinated nutritional and medical stabilisation.

Rosenhan tested whether hospital staff could distinguish people without mental disorder from patients once a psychiatric label and setting shaped interpretation.

Eight pseudo-patients—five men and three women—sought admission to 12 US hospitals across five states. They reported an unfamiliar same-sex voice saying 'empty', 'hollow' and 'thud', while other personal details were largely truthful. After admission they stopped simulating symptoms, behaved ordinarily and recorded ward experiences in notes and diaries.

All were admitted and ordinary behaviour was often interpreted through the diagnosis. They reported depersonalisation, powerlessness, limited privacy and little staff contact. The range of public/private and old/new hospitals strengthens relevance to US inpatient care of that era; standardised presenting symptoms support comparison.

Suzuki et al. assessed whether Japanese inpatients with schizophrenia showed different weight and nutritional status from healthy controls.

The sample contained 333 inpatients aged 16–80, diagnosed with DSM-IV-TR across nine psychiatric hospitals in Niigata Prefecture, and 191 healthy volunteers matched by age and sex. Height and weight produced BMI; fasting blood samples measured total protein, cholesterol, triglycerides and plasma glucose. Physical illness and recent drug-therapy changes were exclusion controls.

Underweight was more prevalent among the inpatients than in the general comparison group, while overweight/obesity prevalence was broadly similar. The findings support monitoring physical as well as mental health during long inpatient stays and reviewing food, activity and nutritional care.

Hans and Hiller asked whether outpatient CBT for adult unipolar depression remains effective—and how often people drop out—in ordinary clinical practice.

Their random-effects meta-analysis combined 34 non-randomised effectiveness studies of individual or group outpatient CBT. It calculated standardised mean gains at treatment end and six-month follow-up for depression severity, dysfunctional cognition, anxiety, distress and functional impairment, plus a weighted dropout rate. Results were benchmarked against high-quality RCTs.

Depression improvement was large for completers (d = 1.13) and intention-to-treat samples (d = 1.06); secondary outcomes showed moderate-to-large gains. Improvement was maintained six months after completion. Weighted mean dropout was 24.63%, and depression effects were smaller than benchmark RCT effects.

Ma, Quan and Liu tested whether perceived social support partly explains the association between core self-evaluation and depression.

A sample of 538 undergraduates—281 women and 257 men—completed standardised questionnaires measuring core self-evaluation, perceived social support and depression. Core self-evaluation covers broad judgments such as self-esteem and perceived competence. Correlations were entered into a mediation analysis to test whether support carried part of the self-evaluation–depression relationship.

Higher self-evaluation and greater perceived support were associated with lower depression, and social support was a partial mediator. This is consistent with a pathway in which more positive self-evaluation is related to perceiving or seeking support, which in turn relates to fewer depressive symptoms. 'Partial' means a direct association remained.

Becker et al. examined whether prolonged exposure to television was associated with changing eating attitudes among ethnic Fijian adolescent girls in a previously media-naive setting.

A prospective multi-wave cross-sectional design compared separate schoolgirl samples in Nadroga: 63 participants in 1995, within weeks of television's introduction, and 65 in 1998 after three years of exposure. Both completed a culturally adapted 26-item Eating Attitudes Test; semi-structured interviews with a subset explored symptoms, body image and television meaning.

Indicators of disordered eating were significantly more prevalent in 1998. Interview narratives described dieting or weight loss as ways to resemble admired television characters and linked thinness with opportunity or success. Quantitative change and qualitative meaning converge on a plausible cultural-learning influence.

Reichel et al. tested whether people with anorexia nervosa show an implicit appetitive response to extremely emaciated body images even when self-report does not reveal it.

Thirty-six adolescents/young adults with anorexia and 36 controls viewed images, including extremely emaciated bodies drawn from pro-anorexia sites. Candidate images were reduced from about 8,000; 36 were rated by a 100-person pilot and 16 selected. Acoustic probes elicited an eye-blink startle response measured physiologically, while participants also gave subjective ratings.

The anorexia group showed startle inhibition—interpreted as appetitive responding—to the extremely emaciated images, whereas controls showed startle potentiation, an aversive response. The groups did not show the same difference in subjective ratings. This divergence suggests implicit and reported emotional processing can differ.

Clinical methods must separate treatment effects from alternative explanations while protecting participants whose symptoms may affect risk, consent or withdrawal.

A sound drug RCT pre-registers outcomes, obtains valid consent, records a baseline, randomly allocates through an independent process, standardises dose/contact, monitors adverse events and analyses all allocated participants where possible.

A published research report separates the question, procedure, evidence and interpretation so readers can audit how the conclusion was reached.

Before publication, editors ask knowledgeable independent reviewers to judge contribution, methods, analysis, ethics, clarity and fit with the journal. Reviewers identify specific problems and recommend accept, revise or reject; authors respond and editors decide. Peer review can catch error and improve transparency.

HCPC standards protect the public by setting threshold proficiency and ethical expectations for registered practitioner psychologists.

Practitioners must meet profession-specific proficiency standards to register and remain within the scope for which they are competent. The conduct, performance and ethics standards guide behaviour and are used when concerns are raised. Registration lets the public check professional status; failure to meet standards can trigger fitness-to-practise processes.

Clinical data analysis begins with the research question and measurement level, then separates a sample description from an inferential decision.

For CBT sessions attended and number of anxious episodes, Spearman's rank fits because the hypothesis concerns association and both paired variables can be ranked. After calculating the statistic, use sample size, direction, significance level and the correct critical value to decide whether to reject the null.

Clinical evaluation links a concrete method feature to the exact claim that becomes stronger or weaker, then states the bounded conclusion that survives.

The same feature can cut both ways: non-randomised routine CBT studies improve ecological validity but weaken causal control; a scanner offers objective output but interpretation may remain uncertain. Triangulation is strongest when methods address complementary weaknesses.

A content analysis converts words, images or audio from at least two real sources into systematic evidence about attitudes to mental health.

1. Define a focused comparison, such as tabloid versus support-organisation reporting or recent versus historical coverage. 2. Set transparent source criteria, dates and search terms; sample sources without selecting only striking cases. 3. Define mutually clear categories such as dangerousness, blame, recovery, medical explanation, support or neutral description. 4. Pilot a coding manual, then have at least two coders independently code the same material and resolve disagreements. 5. Tally frequencies or percentages and retain short contextual notes. 6. report procedure, results and a discussion of the supported conclusion.

Credibility depends on who created each secondary source, for what audience and purpose, and whether length, format and date are comparable. Public availability does not remove privacy or stigma risk: avoid exposing identifiable vulnerable people, reproduce only necessary material and frame categories without reinforcing harmful labels.

Quantitative and qualitative describe the form of data; primary and secondary describe where data come from. These are two separate decisions, so a dataset can be both qualitative and secondary, or quantitative and primary.

Choose from the inference backwards. A CBT researcher might collect primary rating-scale scores for comparable change and primary interviews for how change was experienced; a review could instead analyse secondary trial reports. Record who produced secondary data, when, how and for what purpose.

A sample represents a defined target population. The sampling frame and selection rule determine who could be included and therefore how far findings may generalise.

A complete answer names the frame, eligibility, numbers and selection action. For example, stratifying a school sample by age still requires random selection within each age group; merely choosing equal numbers is not proportional stratification.

Experimental design states how participants enter conditions. Its central trade-off is participant-variable control against order effects and practicality.

Choose by the claim and task. A before-and-after therapy comparison is naturally repeated measures, whereas exposure that permanently changes a person may require independent groups. State how the design works in the scenario, not only its name.

A hypothesis turns an aim into a falsifiable prediction by naming the population and operationalised variables: exactly how each variable will be manipulated, grouped or measured.

Directional experiment: 'Participants hearing traffic noise will correctly recall fewer words from a 20-word list than participants hearing silence.' Null: 'There will be no significant difference in the number correctly recalled...' A correlation instead predicts an association between two co-variables; it does not use an IV.

Questionnaires and interviews gather self-report data. Their structure controls comparability, while question form controls whether the response is numerical, categorical or descriptive.

Use neutral, single-focus wording, exhaustive non-overlapping response options and a time frame. Pilot for ambiguity and accessibility. Standardise instructions, setting and recording; protect privacy so social desirability is less likely. A semi-structured interview can combine a 0–10 rating with an open follow-up asking why that rating was chosen.

An experiment manipulates an independent variable (IV), measures its effect on a dependent variable (DV), and controls plausible alternatives so a causal inference may be tested.

Operationalise the IV as the exact condition difference and the DV as a replicable score, count or measurement. Hold extraneous variables constant or randomise them; if one varies systematically with the IV, it becomes a confounding variable.

Observation records behaviour using declared categories and sampling rules. Design choices affect naturalness, ethics, depth and reliability.

Operationalise each category with observable start/stop rules, pilot it, select times and locations, train two observers, independently code overlapping periods and calculate agreement. Record context without changing categories after seeing a striking case.

A correlation asks whether two measured co-variables vary together. Plot paired scores first: an upward pattern is positive, a downward pattern negative, and no consistent pattern indicates little or no correlation.

Strength describes how closely points follow a monotonic pattern, from values near 0 to values near +1 or −1. The sign gives direction, not strength. A scatter diagram also reveals outliers, restricted range and non-linear patterns that one coefficient can conceal.

Correlations are valuable when manipulating a variable is impossible or unethical, for prediction and for identifying questions for controlled or longitudinal research. Spearman's rank is suitable when paired data are at least ordinal and the prediction concerns association.

Additional methods answer different psychological questions; their labels do not confer automatic credibility. Match the evidence form to the claim.

Control is purposeful management of variables and response effects that could offer a rival explanation for the observed result.

Operationalisation converts a construct into a repeatable manipulation or measure. Randomising condition order is not the same as random allocation; standardisation does not control a badly chosen measure.

Descriptive statistics summarise the sample; they do not decide whether a population effect is statistically significant.

Build a frequency or summary table, check missing/impossible values and sample size, calculate with units and requested precision, then compare like with like. Interpret using actual values: name the group, direction and size of the pattern. Formulae need not be memorised, but the mathematical steps must be competent.

Inferential testing asks how compatible the observed sample result is with the null hypothesis. Choose the test from the research question, design and level of measurement before calculation.

State a one- or two-tailed hypothesis, calculate the observed value, identify N or degrees of freedom, select the matching significance level and critical value, then apply the table's rule. For Wilcoxon a sufficiently small T is significant; for chi-squared a sufficiently large observed value is significant; Spearman tables use coefficient magnitude and direction.

At p≤.05, a result at least this extreme would occur no more than 5% of the time under the null model. Rejecting a true null is Type I; retaining a false null is Type II. A stricter alpha reduces Type I risk but raises Type II risk when other factors are fixed.

Methodological evaluation is a chain: identify a concrete feature, explain its effect on evidence, and revise the exact conclusion.

Standardise instructions, operationalise variables, pilot materials, use inter-rater or test-retest checks, control confounds and sample across the target population. For qualitative research, retain an audit trail, compare sources, check interpretations and acknowledge researcher positioning.

Thematic analysis identifies patterned meaning across qualitative data while preserving enough context to show how the interpretation was produced.

1. Transcribe and repeatedly familiarise yourself with the data. 2. Code relevant units using clear labels. 3. Group related codes into candidate themes. 4. Check themes against coded extracts and the full dataset. 5. define, name and distinguish each theme. 6. Report the pattern with brief evidence and a conclusion tied to the research question.

Use a coding guide, reflexive notes and an audit trail. A second coder can independently code a shared subset and disagreements can refine definitions. Frequencies may convert themes into quantitative summaries, but prevalence alone does not establish meaning or importance.

Published conventions separate question, method, evidence and interpretation so another researcher can judge and replicate the work.

Editors send suitable manuscripts to knowledgeable reviewers, who examine originality, method, analysis, ethics, clarity and fit. Authors revise or answer objections and the editor decides. Replication, corrections and later synthesis continue scrutiny after publication. Preregistration and accessible materials can make selective reporting easier to detect.

When reading a paper, trace every discussion claim back to a stated result and then to the measure and sample that produced it.

Human research begins with competence, a proportionate risk assessment and respect for autonomy, dignity, privacy and scientific value.

Under the UNCRC, children have a right to express views in matters affecting them, with weight appropriate to age and maturity, alongside protection from harm. Seek child assent and appropriate adult consent, design accessible choices and make refusal real; protection should not silence participation.

UK animal procedures are bounded by the Animals (Scientific Procedures) Act 1986 and Home Office regulation. A potentially useful question does not itself justify animal use.

Researchers must justify species, numbers, procedure and expected benefit; work under the relevant licences, trained personnel and veterinary/welfare oversight. A harm–benefit assessment considers severity, duration, cumulative effects and whether reliable knowledge is realistically obtainable.

A key question for society uses concepts, theories or research from Topics A–H, excluding optional Topics F and G, to explain a real decision and its consequences.

1. Define the social question and stakeholders. 2. Select two relevant psychological routes, such as biological and learning accounts. 3. Explain each mechanism accurately. 4. apply it to the question with specific evidence. 5. compare evidence quality, ethics, feasibility and unintended effects. 6. reach a conditional judgment stating when an intervention or policy is justified.

For reducing internet addiction, operant conditioning explains how variable social rewards maintain checking and suggests changing cues and reinforcement; dopamine-reward evidence offers a biological risk mechanism but does not prove inevitability. Evaluate whether support improves functioning without pathologising ordinary use or transferring control to platforms.

Ethical evaluation asks whether a study's knowledge could justify its method, which beings bear risk, and what safeguards make participation or animal use proportionate.

Use a harm–benefit analysis that is study-specific: severity and duration of harm, vulnerability, scientific validity, alternative methods and likely value. Poorly designed research is ethically weak because burdens cannot yield reliable benefit.

A feasible investigation aligns aim, operationalisation, method, sample, controls, ethics, analysis and resources. One weak link can make the final inference unusable.

1. Convert the topic into a focused aim and justified hypothesis. 2. define the target population and recruit through a named frame. 3. choose method/design and operationalise variables. 4. pilot instructions, materials, timing and data capture. 5. control rival explanations without destroying the target behaviour. 6. complete risk/ethics review and consent materials. 7. preselect descriptive and inferential analysis. 8. standardise, record deviations, securely store data and debrief.

Time, access, equipment, researcher competence and participant burden constrain design. A larger representative sample may cost more; tighter laboratory control may reduce realism; repeated measures may save participants but create order effects. Improvements should target the most consequential threat.

Reductionism explains behaviour through smaller components; holism studies the interacting person and context. They are complementary levels of analysis, not simply bad versus good.

A drug study may isolate a receptor pathway to test efficacy, while interviews and longitudinal follow-up reveal adherence, relationships and quality of life. Strong explanation moves between levels and checks whether a component mechanism still predicts outcomes in context.

Psychological explanations select mechanisms at different levels. Comparing them means tracing what each predicts, what evidence could test it and whether their claims can coexist.

For addiction, reinforcement explains maintained behaviour, cognitive accounts explain expectancies, biology explains vulnerability and reward response, and social factors explain availability and norms. A combined model may predict more, but each link still requires evidence.

Psychology is scientific to the extent that claims are operationalised, tested systematically, exposed to falsification and revised through transparent evidence.

Laboratory control can isolate causes but reduce realism; qualitative work can be systematic and transparent without pretending meaning is observer-free. Measurement error, replication failure, publication bias and paradigm disagreement are reasons for stronger scientific practice, not automatic proof that the subject is non-scientific.

Cultural and gender bias can enter through who is studied, whose behaviour defines the norm, how constructs are measured and how findings are applied.

A cross-cultural difference may reflect sampling, language, response style, economic conditions or the construct itself. Researchers should involve communities, examine invariance, disaggregate data and avoid deficit labels.

Nature refers to inherited and biological processes; nurture to physical and social environments and experience. Modern evidence usually tests their interaction rather than allocating behaviour to one side.

Twin and adoption comparisons estimate whether greater genetic relatedness accompanies similarity, but shared environments, gene–environment correlation and unequal experiences complicate inference. Experiments and longitudinal studies can test environmental pathways, while epigenetic processes show experience can affect gene expression without changing DNA sequence.

A vulnerability–stress account predicts that inherited risk may be expressed under particular adversity or protection. People also select and evoke environments partly through heritable traits, so genes and environments are statistically entangled. Cross-fostering, longitudinal and genetically informed designs address different parts of this problem.

Compare concordance, effect size, temporal order and plausible environmental differences before deciding how much each pathway contributes in the studied population.

Development over time is shown when new methods, populations, evidence or ethical standards alter an explanation—not merely because a later study has a newer date.

Early introspection gave way to observable behaviourism; cognitive methods reintroduced mental processes as testable models; neuroscience linked cognition and behaviour to biological measures. Social explanations expanded from personality-only accounts to situational and identity processes. Clinical practice moved from moral/institutional judgments towards standardised diagnosis, evidence-based treatment, rights and cultural formulation.

Progress can mean better measurement, replication, broader samples, less harm and more accurate prediction. It may also create new problems: a diagnostic revision disrupts historical comparison, a brain image can invite biological reductionism, and formal ethics cannot eliminate power differences. Replication can reveal where an influential effect depends on context.

Use dates only to anchor the sequence; the evaluative work is to explain why the later evidence changed confidence, practice or the scope of a theory.

Social control is the deliberate or structural regulation of behaviour towards norms or goals. Psychology can make this control effective, but effectiveness does not settle legitimacy.

Assess the evidence for behaviour change, durability, side effects, fairness and who holds power. Beneficial coordination—such as a safety norm—can coexist with surveillance, stigma or unequal enforcement.

Application translates a supported psychological mechanism into action, then tests whether benefits transfer beyond the original study and whether harms are acceptable.

1. Define the social need and target outcome. 2. identify the mechanism and quality of supporting evidence. 3. co-design an intervention suited to population and culture. 4. pilot feasibility, consent and accessibility. 5. compare outcomes with an appropriate baseline/control. 6. measure adverse effects, equity and durability as well as average benefit. 7. revise, stop or scale with transparent accountability.

Memory research can improve interview questioning; learning principles can support phobia treatment; social-identity evidence can shape prejudice reduction; clinical evidence can guide treatment and anti-stigma practice. Each application changes context and therefore needs fresh outcome evidence.

Socially sensitive research concerns topics or interpretations that can affect identifiable people or groups beyond the immediate study—for example mental health, aggression, prejudice, trauma, crime, intelligence or childhood.

Sensitivity is not a reason to prohibit all research: avoiding a topic can leave harm invisible. Evaluate social value, scientific validity, power, foreseeable misuse and whether safeguards can reduce risk without silencing participants.