Non-Experimental Methods

Overview of Non-Experimental Methods

• Psychologists often investigate behaviour without manipulating variables.
• Key non-experimental approaches covered in the CIE course: correlation studies, longitudinal studies, observational studies (e.g. Bandura et al., 1961; Piliavin et al., 1969), case studies, questionnaires and interviews.
• These methods are valuable when experiments would be impractical, unethical, or when naturally occurring variation is of interest.

Self-Reports

Self-reports obtain data directly from the participant rather than by observing or testing them.

Questionnaires

• Questions are presented in written form: paper-and-pencil or online.
• Most common formats: closed and open questions.

Evaluation of Questionnaires

Strengths
• Closed-question data are easy to total, summarise, calculate averages → efficient analysis.
• Open questions can yield rich, in-depth information (↑ validity).
Limitations
• Open-response interpretation may lack reliability; different researchers may interpret answers differently → potential low inter-rater reliability.
• Low return rate; responders may share characteristics, harming generalisability.
• Social desirability bias: participants give more acceptable answers.
• Participants may guess the study aim; filler questions can disguise true purpose.

Interviews

• Interactive self-report; usually face-to-face but can be phone, video call, or chat.
• Can use any questionnaire format but tend to include more open questions.

Interview Structures

• Structured → identical questions & order for all → high standardisation.
• Unstructured → next question depends on participant’s answer → flexible but harder comparisons.
• Semi-structured → core set of fixed questions plus tailored follow-ups → balance of comparability and depth.

Evaluation of Interviews

• Rich qualitative data; interviewer can clarify misunderstandings.
• Time-consuming; requires trained staff.
• Subjectivity risk (researcher bias); personal viewpoints may influence questioning or interpretation.
• To check objectivity, another researcher blind to aims can interpret transcripts.
• Social desirability still a concern.

Reliability & Validity in Self-Reports

Reliability
• Consistency of measurement.
• Assessed via split-half: participant completes two halves of the same questionnaire; high correlation ⟹ internal reliability.
• Improve by clarifying ambiguous items or training interviewers.
Validity
• Does the measure assess the intended construct?
• Threats: lying, leading questions, social desirability, lack of depth in closed items.
• Assess via concurrent validity: correlate scores with another established self-report on same topic.
• Improve by avoiding leading items, adding open questions, guaranteeing confidentiality.

Key Self-Report Concepts

• Inter-rater reliability: agreement between two researchers scoring the same responses.
• Social desirability bias: tendency to answer in socially approved ways.
• Filler questions: irrelevant items inserted to mask study aims.

Case Studies

Characteristics

• In-depth investigation of a single instance: individual, family, or institution.
• Uses multiple methods: interviews, naturalistic/controlled observations, tests, questionnaires, archival records.
• Valuable for rare phenomena or developmental change; often linked to therapeutic context though research aim is primary.

Data Sources & Techniques

• Direct: participant interviews, behavioural observations, psychometric tests.
• Indirect: relatives, colleagues, medical/school records.
• Triangulation: combining several techniques → cross-validate findings (↑ validity).

Strengths

• Rich, detailed qualitative information.
• Generates insights & hypotheses for further research.
• High ecological and face validity due to natural context.
• Triangulation can verify consistency across sources.

Limitations

• Low generalisability: one case may be atypical.
• Researcher bias: prolonged engagement may reduce objectivity.
• Confidentiality threat: extensive detail may reveal identity.
• Low reliability: unique context prevents replication.

Reliability & Validity Notes

Validity
• Depth & real-world context ↑ validity; triangulation further enhances.
• Close researcher–participant relationship may introduce subjectivity ↓ validity.
Reliability
• Single case + few researchers → difficult to replicate; interpretations may be unique.

Observations

Settings

• Naturalistic → participant’s normal environment, no interference.
Strength: high ecological validity.
Limitations: uncontrolled variables, behaviour may not appear, small/unrepresentative samples.
• Controlled → environment or social context manipulated (may be lab or modified natural setting).
Strengths: easier replication, quicker, larger samples possible, quantitative data easier to analyse.
Limitation: possible demand characteristics (participants act differently when watched).

Structuring the Data Collection

• Unstructured → record all behaviours (often pilot phase).
• Structured → predefined behavioural categories; improves focus, inter-observer reliability.

Behavioural Categories

• Operationalised, discrete, observable actions (not inferred states).
• Clear definitions ensure different observers record the same events.

Observer Role in Social Setting

• Participant observer → joins group; gains insight but may lose objectivity.
• Non-participant observer → remains separate (e.g. behind one-way glass).

Disclosure Status

• Overt → participants know they are watched; may alter behaviour (↓ validity).
• Covert → hidden/disguised; more natural behaviour but raises ethical issues of deception & lack of consent.

Strengths & Limitations Summary

• Naturalistic + Covert → highest ecological validity, lowest reactivity, but ethical/practical difficulties.
• Controlled + Structured + Overt → replicable, quantitative, but demand characteristics threaten validity.

Reliability & Validity in Observations

Reliability
• Assessed with inter-observer (inter-rater) reliability: two observers’ records are compared.
• Improve by clear categories & thorough observer training.
Validity
• Threats: participant reactivity, observer bias.
• Improve by clearer coding, keeping observers blind to aims, using more observers.

Correlations

Definition & Rationale

• Statistical technique examining relationship between two co-variables.
• Appropriate when variables cannot be manipulated for practical/ethical reasons (e.g. long-term violent TV exposure).
• Data may come from self-reports, observations, tests, physiological measures, etc.

Direction of Relationship

• Positive correlation → variables increase together.
Example: more exposure to aggressive models ⟹ more violent behaviour.
• Negative correlation → one variable increases while the other decreases.
Example: more years in education ⟹ lower obedience.
• Zero (uncorrelated) → no consistent pattern.

Causality Caution

• Correlation ≠ causation; a third variable may influence both.
E.g. attention in class and exam score may both depend on student dedication.

Hypotheses in Correlational Studies

• Variables must be operationalised.
• Non-directional: “There will be a correlation between the number of computer games played and A-Level grade.”
• Directional: “There will be a negative correlation between games played and grade; as games increase, grade decreases.”

Evaluation

• Validity depends on accurate, relevant measurement of both variables.
• Reliability high if scientific instruments used; lower if based on subjective self-reports or observations.
• Useful first step that can guide experimental research.

Longitudinal Studies

Definition

• Follow the same participants over weeks, years, or decades to track developmental change or effects of experiences/interventions.
• Enable within-person analysis, controlling for individual differences.

Historical Examples

• Terman study of gifted children.
• Golding’s “Children of the 90s.”
• Dutch longitudinal study of adopted children.
• Marmot’s Whitehall stress study of civil servants.

Strengths & Challenges (implicit)

• Strengths: observe real developmental trajectories; detect long-term effects.
• Challenges: time, cost, participant attrition, changes in measurement tools over time.

Key Terminology (A–Z)

• Behavioural categories, Case study, Closed question, Covert/Overt observation, Correlation (positive/negative), Filler question, Generalisability, Inter-observer reliability, Inter-rater reliability, Likert scale, Longitudinal study, Naturalistic observation, Objectivity/Subjectivity, Open question, Participant/Non-participant observer, Reliability, Social desirability bias, Structured/Unstructured interview, Structured/Unstructured observation, Triangulation, Validity.

Embedded Example Assessment Prompts (from transcript)

• Designing questionnaire items on subject choice (closed + open).
• Choosing unstructured interview to allow flexibility (Judith).
• Ethical issue: discussing phobias may cause distress; need informed consent/debrief.
• Case-study techniques: interview, observation, psychological testing.
• Debra (naturalistic, covert tree observer) vs Jin (controlled lab rat).
• Ekua & Takis’ coffee-dream correlation: measurement via daily coffee diary; discuss reliability/validity.
• Longitudinal textbook reference pp. 30–32.