IB DP Psychology: Approaches to researching behaviour – Revision Notes

Course note: these notes follow IB DP Psychology for first assessment 2019, whose last examination session is November 2026. The replacement course is first assessed in May 2027.

Approaches to researching behaviour is the research-methods strand of IB DP Psychology: quantitative and qualitative methods, sampling, ethics, and how far findings can be generalised. It runs through every approach and option – you evaluate studies with it in every essay – and at HL it is examined directly on Paper 3, where you answer questions about an unseen summary of a study.

Related: Approaches to researching behaviour study guide and practice questions.

Syllabus coverage (first assessment 2019)

Area	What to know
Quantitative methods	True (laboratory) experiments, field experiments, natural and quasi-experiments; correlational studies
Qualitative methods	Interviews (structured, semi-structured, unstructured, focus groups), observations (participant, non-participant, overt, covert, naturalistic, controlled), case studies
Sampling	Random, stratified, self-selected (volunteer), opportunity (convenience), purposive, snowball
Ethics	Informed consent, deception, debriefing, right to withdraw, protection from harm, anonymity and confidentiality; ethics in data reporting
Quality of research	Validity, reliability, credibility, bias, triangulation, reflexivity
Generalisation	Sample-to-population generalisation, theoretical generalisation, case-to-case generalisation (transferability)
HL Paper 3	Unseen study summary; static questions: method, sampling and alternative method (3 x 3 marks), ethics (6), then one of generalisation, credibility or bias (9)

Experiments compared

Type	IV manipulated?	Random allocation?	Setting	Main strength	Main limitation
True (lab) experiment	Yes	Yes	Controlled	Cause and effect; replicable	Low ecological validity; demand characteristics
Field experiment	Yes	Usually	Natural	Higher ecological validity	Less control over extraneous variables
Natural experiment	No – occurs naturally	No	Natural	Studies variables that cannot be manipulated ethically	No causal conclusions with confidence
Quasi-experiment	No – pre-existing groups (e.g. gender)	No	Either	Compares groups that could not be assigned	Confounding participant variables

Designs: independent samples (different participants per condition; risk of participant variables), repeated measures (same participants; risk of order effects, controlled by counterbalancing), matched pairs (participants matched on key variables).

Correlational studies and hypotheses

• A correlational study measures two co-variables without manipulating either. A positive correlation means both rise together; a negative correlation means one falls as the other rises.
• Correlation is not causation: a third variable may cause both, or the direction may be the reverse of what is assumed (bidirectional ambiguity).
• An alternative hypothesis predicts a difference or relationship; the null hypothesis predicts none. Both must name operationalised variables, e.g. “hours of sleep” rather than “tiredness”.
• A confounding variable varies with the IV and offers another explanation of the result.

Qualitative research quality

Concept	Meaning	How it is improved
Credibility	Findings reflect participants’ realities (qualitative equivalent of internal validity)	Triangulation, member checking, rapport, reflexivity
Researcher bias	The researcher’s beliefs affect data collection or interpretation (e.g. confirmation bias)	Reflexivity, researcher triangulation
Participant bias	Acquiescence, social desirability, dominant respondent in focus groups, sensitivity	Rapport, neutral questions, anonymity
Triangulation	Using several methods, data sources, researchers or theories	Method, data, researcher and theory triangulation
Transferability	How far findings apply to other settings	Rich, thick description of the context and sample

Sampling quick table

Method	How	Generalisation
Random	Every member of the target population has an equal chance	Best for sample-to-population generalisation
Stratified	Sample mirrors subgroups’ proportions	Representative on chosen variables
Opportunity	Whoever is available	Weak; often students
Self-selected	Volunteers respond to an advert	Volunteer bias
Purposive	Chosen because they have a characteristic of interest	Qualitative; theoretical generalisation
Snowball	Participants recruit others	Hard-to-reach populations; not representative

Worked example: identifying a method

Summary: researchers interviewed 12 nurses about burnout, using open questions, then coded the transcripts into themes.

• Method: semi-structured interview (qualitative). Characteristics: flexible, open-ended questions; data are words analysed by thematic coding.
• Sampling: likely purposive – participants had to be nurses with relevant experience.
• Ethics: informed consent, confidentiality of identifiable workplace details, right to withdraw, debriefing; protection from distress when recalling stressful events.
• Generalisation: not sample-to-population; possible transferability to similar health-care settings if the context is richly described; theoretical generalisation to models of burnout.

Common mistakes / exam traps

• Calling every study an experiment. No manipulated IV means no true experiment.
• Quantitative language for qualitative studies. Use credibility, transferability and reflexivity, not “statistical significance”.
• Listing ethics without application. Say how each consideration was, or should have been, applied in the study described.
• “Small sample, so not generalisable” as the whole answer. Distinguish the three types of generalisation.
• Forgetting the context. Paper 3 answers must refer to the specific study summary.
• Causal claims from correlations. Say “is associated with”, not “causes”, unless an IV was manipulated.

Frequently asked questions

Is research methodology examined at SL? SL students are not set Paper 3, but they need research methods to evaluate studies on Papers 1 and 2 and to design the internal assessment experiment.

What is Paper 3? A 1-hour HL paper worth 24 marks and 20% of the HL grade: one unseen study summary followed by three compulsory questions. Question 1 has three 3-mark parts (research method, sampling method, an alternative method); Question 2 (6 marks) is on ethics; Question 3 (9 marks) asks about one of generalisation, credibility or bias.

Do I need inferential statistics? For Paper 3 you need to understand significance and generalisation conceptually; descriptive and inferential statistics are applied in the internal assessment.

Self-test

1. What distinguishes a quasi-experiment from a true experiment? Participants are in pre-existing groups, not randomly allocated.
2. Define reflexivity. The researcher’s reflection on how their own beliefs and role may have affected the research.
3. Which sampling method best supports sample-to-population generalisation? Random (or stratified random) sampling.
4. Name two types of triangulation. Method and researcher triangulation.
5. What is counterbalancing? Varying the order of conditions to control order effects in repeated measures designs.