1.6 - Issues in social psychology

Ethics in research and its implications

Ethics concerns both the treatment of research participants and the consequences of the knowledge produced.

Obedience research illustrates a tension between studying behaviour convincingly and protecting participants. If people know a learner cannot be harmed, their response may differ from behaviour under a believed harmful order. Deception can therefore increase realism, but it reduces fully informed choice and can create distress. Milgram’s apparent shocks, learner protests and authority prods raised concerns about protection from harm and whether participants felt able to withdraw. The scientific value of an unsettling result does not automatically outweigh those concerns.

Burger’s partial replication shows that ethical constraints can reshape the question. Screening, repeated withdrawal assurances, a weaker sample shock, monitoring and immediate debriefing reduced risks. Stopping just after the decision at 150 volts also removed the later, more stressful stages. Yet participants still believed they were affecting a learner, and a prod implying that there was no choice could conflict with withdrawal assurances. It is reasonable to judge the procedure less risky while retaining ethical reservations.

Prejudice research presents a related problem. Sherif’s camp staff deliberately produced competition and hostility among boys who did not know they were research participants. The study yielded evidence relevant to conflict reduction, but engineered conflict, deception and limited informed consent remain concerns. A modern learner can examine supplied evidence or use a low-stakes questionnaire; useful learning does not require repeating the harmful manipulation.

Findings also have ethical implications beyond the sample. Saying that a particular group is especially obedient or prejudiced can stigmatise its members. Saying that situations influence harmful behaviour can improve prevention, but may be misused to excuse individual actions. Research reports should distinguish explanations of behaviour from moral or legal conclusions about responsibility and avoid generalisations unsupported by the sample.

Practical design and psychology as a science

Practical research decisions determine what can be measured, while scientific reasoning determines what those measurements can support.

A questionnaire about obedience is feasible and can collect many responses without exposing people to harmful orders. Closed items permit comparison; open items and interviews can explain participants’ reasoning. But the researcher must make terms such as “obey” specific, avoid leading wording and provide questions participants can interpret consistently. An interviewer’s approval or authority may alter an answer. Unstructured exploration can reveal unexpected ideas but makes responses harder to standardise than a structured interview.

Social desirability is a practical and validity problem: people may say they would resist an unjust instruction because that answer is socially approved. A private anonymous questionnaire can reduce this pressure, yet cannot turn self-reported intentions into observed behaviour. Conversely, highly realistic behaviour may be difficult to study safely. There is no single method that maximises feasibility, control, realism and protection simultaneously.

Psychology acts as a science when explanations generate testable predictions, measures are explicit, procedures and analyses are transparent, and claims can be challenged or replicated. Random allocation in Burger’s comparison reduced systematic differences between his conditions; a standard task and script improved comparability. These features make a prediction about peer refusal open to evidence. They do not guarantee that every finding generalises beyond the laboratory.

Distinguish reliability, consistency of measurement, from validity, whether a measure or conclusion is appropriate. A questionnaire can produce stable socially desirable answers and therefore be reliable while failing to capture harmful obedience. Covert observation at Sherif’s camp captured meaningful behaviour but left room for researchers’ expectations to influence interpretation. The use of independent recording and more than one measure strengthened checks without making the observation entirely objective.

An appropriate judgement therefore concerns the particular claim. Quantified answers, replicable procedures and converging measures support scientific inquiry; uncertainty about constructs and observer influence limits some conclusions. Qualitative accounts can contribute systematically when coding decisions and interpretations are transparent. Counting responses alone does not make an invalid measure scientific evidence for the claimed behaviour.

Reductionism and competing explanations

An explanation can identify an important mechanism while becoming reductionist if it treats that mechanism as the whole explanation.

Reductionism explains a complex phenomenon in terms of fewer or simpler components. In social psychology, explaining every harmful act as obedience to authority, or every prejudice as a contest for resources, risks leaving out identity, personality, interpretation and institutions. Simplification is useful when it yields clear predictions, but its usefulness does not establish completeness.

For obedience, agency theory proposes that people may see themselves as agents carrying out an authority’s wishes and transfer responsibility to that authority. Social impact theory instead focuses on the strength, immediacy and number of sources influencing a target. A close, credible expert may exert more influence than a distant source. These social mechanisms differ from a personality account proposing that some people are more disposed to submit to authority. Situational effects are important evidence, yet a person who refuses in the same setting shows why an account of uniform mechanical responding is too simple.

For prejudice, realistic conflict theory explains how incompatible interests over valued resources can create hostility and how cooperative superordinate goals can improve relations. Social identity theory explains how categorisation, identification and favourable comparison can produce in-group favouritism even without substantial resource competition. Both address group-based relations, but their proposed starting points differ. A shared-goal intervention targets interdependence; an inclusive identity targets who is considered part of “us”.

Sherif’s competition and cooperation sequence supports the importance of goal relationships. However, derogation began before the formal tournament, so it does not establish competition as a necessary cause of all hostility. Tajfel’s minimal-group allocation evidence supports a role for categories without a prior hostile history; it does not prove that identity alone explains entrenched social inequality. The two explanations can complement each other when a real situation contains both resource rivalry and meaningful group boundaries.

A non-reductionist judgement does not mean adding every possible factor without evidence. It asks which factors a particular study distinguishes and which remain alternatives. For example, an intervention that improves both access and shared identity may work, but an uncontrolled improvement cannot reveal which mechanism contributed most. Comparisons should relate predictions and evidence, rather than merely list two descriptions.

Culture, gender and nature–nurture

Group differences need careful interpretation because cultural learning, individual experience and study design can be entangled.

Culture includes shared practices, norms and meanings. It can influence whether an authority is trusted, whether dissent is approved and which group categories are socially important. A behaviour that looks like deference in one context might reflect trust or role obligations rather than an inborn submissive character. Cultural explanations concern environmental influences learned through families, education, peers and institutions; they are not fixed descriptions of every member of a nation.

Cross-cultural comparisons face measurement problems. Translating the same question does not ensure the same meaning, and recruitment or status of the researcher may differ. If one group interprets “obey a teacher” as following a reasonable instruction and another imagines an abusive order, their scores do not isolate a cultural difference in harmful obedience. Compare equivalent situations and sampling, consider within-culture variation and report the limits of small groups.

Gender differences also need evidence rather than stereotypes. Burger included men and women and detected no significant difference in their continuation rates. This does not demonstrate exact equality in every setting, and does not justify assumptions that women are naturally more obedient or men naturally less empathic. Social expectations about assertiveness, care and authority can develop through environmental experience. Sampling, power relations and the specific task can change whether a difference appears.

The nature–nurture debate asks about inherited or biological influences and environmental influences. A personality explanation is not automatically a nature explanation: personality can develop through both predispositions and experience. Agency and social-impact accounts emphasise circumstances, while accounts of authoritarian tendencies emphasise individual differences. Those emphases are distinguishable without treating nature and nurture as mutually exclusive.

Burger’s measured empathy and desire-for-control findings illustrate the caution. Some associations with early reluctance or stopping appeared in particular analyses, but traits were not randomly assigned. They did not consistently predict final continuation. One cannot infer that an innate trait caused obedience, or that a strong situation proves that individuals have no influence. A defensible conclusion recognises interaction: learned expectations and personal characteristics may shape how someone interprets an authority situation, whose features also matter.

How psychological understanding changes

Psychological knowledge develops when later research tests earlier claims, revises explanations and changes methods in response to new questions or standards.

Milgram’s research demonstrated more obedience than many people expected and prompted sustained discussion of situational influence. Burger later retained important features of the task while shortening it and adding safeguards. His 70% base-condition continuation rate refers to beginning the next item after the 150-volt protest, while the comparable Milgram rate at that decision point was 82.5%. The nonsignificant difference is relevant to continuity, but Burger did not observe behaviour at 450 volts.

This is both continuity and change. Continuing willingness under authority supports investigating the persistence of situational processes. The ethical changes and narrower endpoint show how acceptable methods and what can be directly concluded developed. Samples, place and procedures also differed, so the two studies do not isolate “generation” as an independent variable. A historical comparison cannot show that all social attitudes were unchanged simply because two rates were broadly similar.

Prejudice research also shows conceptual development. Sherif’s camp highlighted competitive and cooperative goal relationships. Minimal-group research later demonstrated that relatively insignificant categories could influence allocation even without prior hostility, helping widen the explanation beyond serious resource rivalry. Social identity theory supplied a framework for categorisation, identification and comparison. This did not erase the evidence that real competition can intensify conflict; later evidence changed the range of mechanisms considered.

Development is not a rule that the newest study is automatically the best or that a theory is proved by its supporting study. Evaluate whether a later method measures the same construct and improves the relevant inference. A safer study may answer a narrower question, while an older realistic setting may have serious ethical and control limits. Psychological understanding improves through this explicit adjustment between claims and evidence, including failures to replicate or ambiguous findings.

Social control, societal use and sensitive claims

The same psychological knowledge can support beneficial cooperation or influence people in ways that restrict their choices.

Social control concerns attempts to regulate behaviour. Knowledge that a credible, nearby authority can encourage obedience may help a workplace communicate a safe evacuation instruction. It could also help a manager pressure workers into an unfair demand. A scientific explanation of influence does not determine whether the goal, power relationship or means of obtaining compliance is acceptable. Questions about transparency, dissent and who benefits are therefore necessary.

Using psychological knowledge in society can include reducing prejudice, improving safety and promoting constructive resistance to harmful orders. Sherif’s evidence suggests that genuinely interdependent shared goals can improve group relations; identity reasoning suggests that a wider shared membership can reduce favouritism. A community project can combine those ideas with fair access and shared decision-making. Its benefits need evaluation: apparent harmony could conceal pressure on an excluded group to remain silent. An ethical project offers voice and influence, rather than making cooperation conditional on obedience to the majority.

Socially sensitive research has possible consequences for the groups represented, including stigma, discrimination or political misuse. A small study comparing groups’ obedience can be publicised as evidence that one culture or gender is inferior. That extrapolation may ignore sampling and context and encourage the very prejudice being studied. Sensitivity is not a reason to make difficult questions unaskable; it is a reason to design and communicate them responsibly.

Explain measured variables, distinguish within-group variation from group averages and avoid treating nationality or gender as a biological essence. Give affected people a meaningful voice in research where appropriate, protect confidentiality and show the boundaries of the conclusion. Anonymity protects an individual’s identity; it does not automatically prevent a report from stigmatising an entire group.

These three strands are related but distinct. The evacuation example is a use of knowledge and a form of social control; whether it is legitimate depends on the aim and implementation. A prejudicial headline concerns the social sensitivity of a claim. A reasoned response therefore identifies the mechanism, the evidence and the people affected, then judges the particular use rather than assuming that all control is harmful or all application beneficial.