Issues, debates and perspectives in psychology

6.1 · Free notes

Ethical considerations and practical issues in psychological research

Key points

• The BPS Code of Ethics and Conduct (2021) is built on four principles: respect, competence, responsibility and integrity. Older guidelines (consent, deception, protection from harm, withdrawal, confidentiality, debriefing) are still used as the practical checklist.
• Informed consent: participants should know the aims and what they will do before agreeing. Children under 16 need parental consent. Where this is impossible, presumptive consent (asking a similar group) or prior general consent can be used.
• Deception is sometimes needed to avoid demand characteristics, but it breaks informed consent. It must be justified by the study's value, and participants must be debriefed afterwards, with the chance to withhold their data.
• Protection from harm: risk should be no greater than in everyday life. Examples include Milgram (stress) and Watson and Rayner's Little Albert (induced fear, never removed).
• Right to withdraw: participants can leave at any time and withdraw their data. This is undermined if they feel pressured to continue, as with Milgram's verbal prods.
• Confidentiality and privacy: data should be anonymised (numbers or pseudonyms), stored securely and protected under the Data Protection Act / UK GDPR. In case studies, identity can be revealed by unique details.
• Ethics committees carry out a cost-benefit analysis, weighing the benefits of the research against the risks to participants and society. Social sensitivity matters too, because findings can stigmatise groups.
• Animal research is regulated by the Animals (Scientific Procedures) Act 1986 and the 3Rs: Replace animals where possible, Reduce numbers, Refine procedures to lessen suffering. Critics question whether results generalise to humans.
• Practical issues affect the quality of data. Reliability is consistency (test-retest, inter-rater). Validity is whether the study measures what it claims, including ecological validity. Both are threatened by demand characteristics, social desirability bias and researcher bias.
• Sampling issues: volunteer or opportunity samples (often students or Western participants) can be biased, which limits generalisability. Lab studies give control but lower ecological validity and may lack mundane realism.
• Inter-rater reliability can be given as % agreement = (agreements ÷ total observations) × 100, or as a correlation. Reliability above about +0.80 is generally treated as good.
• There is a trade-off between ethics and validity: telling participants everything may change their behaviour, while deception improves validity but costs ethically. Evaluation must discuss this balance.

In more detail

• Worked inter-rater reliability: two observers agree on 45 of 50 observations → 45 ÷ 50 × 100 = 90% agreement, which counts as good reliability.
• Ethics in the core studies: Milgram used deception to avoid demand characteristics; Piliavin's subway passengers could not consent, withdraw or be debriefed; Simons and Chabris kept the gorilla secret but debriefed everyone afterwards.
• Dealing with problems: deception → full debrief and the chance to withdraw data; no consent possible → presumptive or retrospective consent; risk of harm → stop the study and offer support or counselling.
• Socially sensitive research: Sieber and Stanley (1988) defined it as research with possible consequences for the participants or the group they represent, such as studies of race and intelligence.
• Exam tip: name the guideline, explain how a specific study broke or kept it, then say how it was or could be dealt with. A generic list of ethical issues scores few marks.