Chapter 4: Research Methods in Abnormal Psychology

Examining Abnormal Behavior

• Behavioural scientists investigate abnormal behaviour using the scientific method – same rigor as volcanology or cellular biology.
• Challenges:
  • Complex bio-psycho interactions; no simple answers (e.g. hallucinations, suicidality).
  • Many relevant processes are inaccessible (we can only infer mental states indirectly).
• Practical importance for consumers:
  • Knowing when childhood aggression matters; sunlight vs Hawai‘i trip for depression; ECT fears; therapy stagnation; early memory problems.
• Being an informed consumer requires understanding research methodology to distinguish fad from evidence-based practice.

Important Concepts & Student Learning Outcomes

• Three fundamental questions in abnormal psychology:
  1. What problems cause distress/impairment?
  2. Why do people behave unusually? (Etiology)
  3. How can we help them behave more adaptively? (Treatment evaluation)
• APA Undergraduate SLOs addressed:
  • Describe research methods (advantages/disadvantages).
  • Define key research concepts (hypothesis, operational definition, etc.).
  • Recognise sociocultural, theoretical & personal biases.
  • Apply ethical standards.

Basic Components of a Research Study

• Hypothesis: educated, testable guess.
• Research Design: plan to test hypothesis.
• Variables:
  • Dependent Variable (DV) – measured outcome (e.g. cognition scores).
  • Independent Variable (IV) – manipulated/influencing factor (e.g. MDMA use).
• Validity:
  • Internal Validity – confidence that IV causes DV change.
  • External Validity – generalisability beyond study.
• Table 4.1 summarises these elements.

Example: MDMA & Memory (University of Cologne)

• Research Q: Does using MDMA for 1 yr decrease cognitive performance?
• Users = ≥10 pills/yr.
• Tests: digit span, Stroop, Trail-making.
• Found poorer visual learning ⇒ practical warnings for potential users.

Confounds & Control Strategies

• Confound: extra variable making results uninterpretable (e.g. differing IQ among MDMA users).
• Strategies to boost internal validity:
  1. Control Group – identical except no IV exposure.
  2. Randomisation – equal chance assignment.
  3. Analogue Models – lab imitation of phenomenon (e.g. lab binge-eating).
• Trade-off: Internal ↑ often means External ↓ (generalisability). Need balanced programme of studies.

Statistical vs Clinical Significance

• Statistical Significance: probability effect occurred by chance is low (p<.05).
• Clinical Significance: practical, meaningful impact.
• Effect size statistics quantify how large the change is.
• Social Validity (Wolf, 1978): ask clients & significant others if changes matter.
• Example: Positive-mood video ➔ anorexia patients drank 75 ml vs 38 ml after neutral clip – statistically sig yet still far below 199 ml of controls, thus limited clinical meaning.

The “Average” Client & Patient Uniformity Myth

• Group means hide individual variability.
• Patient Uniformity Myth (Kiesler, 1966): treating all participants as homogeneous may produce inaccurate conclusions.

Types of Research Methods

1. Case Study Method

• Intensive observation of one/few individuals; historically important (Freud, Masters & Johnson, Wolpe).
• Lacks scientific control, high risk of coincidences & media sensationalism.

2. Correlational Research

• Measures natural relationships; no manipulation – cannot infer causality (directionality problem).
• Correlation Coefficient r ranges -1\le r \le +1.
  • Positive: variables move together.
  • Negative: inverse relation.
• Epidemiology: correlational approach mapping incidence, prevalence, consequences.
  • Uses prevalence (total cases) & incidence (new cases) figures.
  • Historical example: Niacin-deficiency psychosis; modern example: PTSD reactions post-9/11.

3. Experimental Research

• Manipulate IV, observe DV – allows causal inference.
• Group Experimental Designs:
  • Clinical Trial: formal experiment testing treatment efficacy & safety.
  • Sub-types: Randomised Clinical Trial (RCT), Controlled Clinical Trial, Randomised Controlled Trial (gold standard).
  • Control Conditions: No-treatment, placebo, comparative treatment.
  • Placebo Effect vs Frustro Effect (expectancy disappointment).
  • Double-Blind procedures prevent allegiance bias.
• Comparative Treatment Research: evaluate multiple active treatments; analyse process (why it works) & outcome (whether it works).

4. Single-Case Experimental Designs (Skinner)

• Systematic manipulation within one participant; improves internal validity beyond simple case study.
• Repeated Measurement assesses:
  1. Level
  2. Variability
  3. Trend
• Withdrawal Design (A-B-A): baseline ➔ treatment ➔ return to baseline.
  • Ethical/practical limits when treatment can’t be removed.
• Multiple Baseline: staggered treatment across behaviours, settings, or individuals—avoids withdrawal.
• Example: Functional Communication Training study – 5 autistic children; challenging behaviour ↓ only after intervention in each setting.

Genetics & Behaviour

• Behavioural genetics disentangles genotype (gene code) & phenotype (observable traits).
• Endophenotypes: internal, heritable traits (e.g. working-memory deficits in schizophrenia).
• Research progression (Kendler, 2005):
  1. Basic Genetic Epidemiology – Is trait heritable?
  2. Advanced Genetic Epidemiology – How do genetic influences work (sex, age)?
  3. Gene Finding – Linkage & association locate genes.
  4. Molecular Genetics – What do the genes do biologically?

Family, Adoption, & Twin Studies

• Proband: family member with trait.
• Higher concordance in first-degree relatives suggests genetic role.
• Adoption designs separate genes vs environment.
• Twin studies: MZ> DZ concordance ⇒ genetic influence; e.g. height h^2\approx 0.9.
  • Vietnam Era Twin Registry: Adult antisocial behaviour shows stronger genetic effect than juvenile.

Linkage & Association Studies

• Genetic Markers with known loci help map unknown disorder genes.
• Linkage: co-inheritance within families; Association: compare markers in cases vs controls.
• Replication essential – early bipolar-chromosome-11 findings not replicated.

Studying Behaviour Over Time

• Prevention Research Categories:
  1. Positive Development / Health Promotion (e.g., Seattle Social Development Program).
  2. Universal Prevention – whole population risk factors.
  3. Selective Prevention – at-risk subgroups (e.g., bereaved children).
  4. Indicated Prevention – early symptom individuals.

Cross-Sectional vs Longitudinal Designs

• Cross-Sectional: compare different age cohorts simultaneously; quick but cohort effects (age confounded with era).
  • Teenage drinking attitudes example: 12-yr-olds 36 % drink-to-get-drunk belief vs 64 % (15) vs 42 % (17).
• Longitudinal: follow same individuals; no cohort effect, shows individual trajectories; costly; cross-generational effect (findings may not generalise to newer cohorts).
• Sequential Design: combines both (e.g., Chassin’s 10 cohorts on smoking beliefs over decades).

Cross-Cultural Research Considerations

• Culture can act like an IV; but groups differ in genetics, history & measurement equivalence.
• Symptom expression varies (e.g., Nigerian depression = somatic heat/heaviness; U.S. = worthlessness; Chinese = fewer anhedonia complaints).
• Treatment models reflect cultural values (Japanese family-style hospitals; Saudi blend of religion & medicine).
• Requires culturally informed instruments & interpretation.

Power of a Programme of Research & Replication

• No single design suffices; converging evidence from multiple methods (e.g., Durand’s work on autism: single-case ➔ longitudinal ➔ RCT).
• Replication across studies, labs, populations builds confidence.

Research Ethics

• Balance scientific rigour vs participant welfare.
• Informed Consent components: competence, voluntarism, full information, comprehension.
• IRB oversight mandatory in universities/medical centres.
• APA Code stresses confidentiality, debriefing, minimising harm; special guidelines for children (age ≥7 assent + caregiver consent).
• Participatory Action Research: consumers involved in design, conduct, interpretation – enhances relevance & respect.

Key Numerical & Statistical References

• Anorexia smoothie study: 75 ml vs 38 ml vs 199 ml controls.
• Binge drinking prevalence in U.S. college students ≈ 40\%.
• Familial aggregation of blood–injury phobia \approx 60\%.
• Correlation strength examples: behaviour problems & marital distress r\approx +0.50; social support & illness r\approx -0.40; strangers r\approx 0.00.
• Perfect positive correlation r=+1.00; perfect negative r=-1.00.

Practical/Philosophical Implications

• Misinterpreting statistical vs clinical significance can lead to premature adoption of ineffective therapies.
• Patient uniformity myth cautions against one-size-fits-all interventions.
• Cross-cultural variation underscores need for culturally adaptive diagnostics and treatments.
• Genetic findings must consider environment – ethical issues in predictive testing and stigma.
• Preventive focus aims to shift paradigm from treatment to early intervention & health promotion.