9.1.8, 2.2.13, 5.2.3, 5.2.4 - Case studies and biological methods

Depth and inference in case studies

A case study investigates one bounded case intensively, which may be a person, a therapy group or an institution rather than necessarily one participant.

A case study is a research strategy, not a synonym for an interview. Interviews, observations, test scores, medical records and brain scans can all contribute. Qualitative data capture experiences and meanings; quantitative measures can show performance or change. Combining sources is triangulation. If an interview reports improvement but a repeated test does not, a careful account preserves that disagreement and considers whether they measure different aspects of recovery.

Henry Molaison (HM) is an example of a brain-damaged patient whose memory was investigated repeatedly using several methods. His case helped distinguish aspects of memory that were impaired from abilities that remained. The detailed case is taught in the cognitive lesson. A rare pattern can challenge a theory claiming all memory is one undifferentiated ability, even though one person's history is not representative of every patient. Researchers cannot randomly create that brain damage or rule out every pre-existing difference. Squire's account of HM's contribution.

Lavarenne and colleagues (2013) provide a group case example. They explored how an existing outpatient psychotherapy group might support people experiencing psychosis. The report describes one 45-minute session attended by six people with schizophrenia or schizoaffective illness. Clinicians described speech, emotions and interactions and interpreted them through Ego boundaries, meaning the psychological distinction between oneself and others. For example, giving and rejecting gifts were interpreted in relation to connection and tolerance of closeness. The authors proposed that the group offered stability and support for firmer boundaries. This is an interpretation of clinical material, not a randomly assigned treatment comparison. Original abstract and Pearson's study summary, pp24–26.

The detailed context allows a reader to understand a possible therapeutic process that a single score might miss. However, selecting one session and interpreting it through the clinicians' framework can privilege supporting examples. There is no untreated comparison that isolates the group as the cause of improvement. Access to the therapists' contemporaneous notes, contradictory cases and a transparent account of interpretation would strengthen credibility, meaning that the account is believable and grounded in the evidence. Neither a compelling story nor a scan automatically establishes causation.

Clinical interviews and mixed evidence

Clinical research interviews investigate people's experiences and views, which should be distinguished from diagnostic interviews used to inform professional assessment.

A structured interview uses a fixed schedule, increasing comparability but limiting unexpected detail. A semi-structured interview has planned prompts and flexible follow-up questions: “What did you find useful?” can be followed by “Can you give an example?” An unstructured interview follows the conversation more freely, allowing depth while making comparisons and repetition harder. Avoid leading follow-ups such as “How much better did this excellent group make you feel?” Recording with permission and accurate transcription retain the evidence for analysis.

Vallentine and colleagues (2010) investigated a psycho-educational group at Broadmoor, a high-security hospital. Forty-two male patients were referred and 31 completed the programme. Four 20-session groups addressed understanding of mental illness, treatment and coping. Twenty-one completers gave post-group semi-structured interviews; transcripts were content-analysed, alongside questionnaires and clinical records. Interviewees valued information and sharing experiences, but also described difficulty disclosing experiences and remembering material. Group-level questionnaire comparisons did not establish significant pre/post improvements. The authors cautiously suggested usefulness and engagement benefits, with further work needed. Original article, author-uploaded reprint.

The interviews explain what patients valued and what could be improved; numerical outcome scores address a different question about measured change. Interviewing only available consenting completers leaves out important perspectives from non-completers. Patients may also offer favourable comments because of the institutional relationship. The absence of an experimental control makes it unsafe to attribute change to the programme. An independent interviewer, explicit reassurance that care is unaffected, and ethically seeking varied perspectives would improve the evidence. These are research decisions, not a classroom invitation to interview clinical patients.