Reliability, Validity and Ethics
Judging the Quality of Psychological Research
Two ideas run through every part of the Edexcel specification: reliability and validity. A reliable study gives the same results again and again. A valid study actually measures the thing it set out to measure. The two are not the same — a study can be reliable but not valid, for example if a test consistently measures something other than what the researcher intended.
Reliability = consistency of results on repeated occasions. Validity = accuracy — whether the study truly measures what it claims to and whether findings apply beyond the study itself. Edexcel requires you to judge the reliability and validity of sampling methods, experimental designs, and quantitative and qualitative data.
Reliability and Validity of Sampling Methods
The way a sample is chosen affects how far results can be trusted beyond the people actually tested — this is generalisability. Random and stratified sampling tend to produce samples that better represent the target population, improving generalisability. Volunteer and opportunity sampling are quicker but risk a biased sample — for example, volunteers may be more motivated or confident than the average person, reducing how far the findings can be generalised.
Reliability and Validity of Experimental Designs
Repeated measures designs use the same participants in every condition, removing participant variables as an explanation for any difference found — this can improve validity. However, taking part in more than one condition risks order effects, which can reduce reliability unless controlled. Independent measures designs avoid order effects, which supports reliability, but participant variables between the separate groups can threaten validity if the groups differ in an important way before the study even begins.
Reliability and Validity of Quantitative and Qualitative Methods
Quantitative data (numbers, such as reaction times or test scores) is usually easier to check for reliability, because the same measurement can be repeated and compared directly. However, numbers alone can miss the depth of an experience, which can limit validity. Qualitative data (words, such as interview transcripts) can capture rich detail about how someone actually thinks or feels, supporting validity, but it is harder to check for reliability because different researchers may interpret the same words differently.
Edexcel expects you to weigh up studies using six issues: validity, reliability, generalisability, ethics, objectivity, and subjectivity. A structured laboratory experiment tends to score well on reliability and objectivity because procedures are standardised and measurements are numerical. A case study or unstructured interview may score well on validity, because it captures depth and detail, but poorly on reliability and objectivity, because the researcher’s own interpretation shapes the data.
Ethical Issues in Psychological Research
Because psychological research studies real people, researchers must deal responsibly with five named ethical issues: informed consent, deception, confidentiality, the right to withdraw, and protection of participants. Understanding each issue means knowing what it is and knowing a practical way a researcher deals with it.
Informed consent means participants are told what the study involves before they agree to take part. Researchers deal with this by giving a written information sheet and asking participants to sign a consent form. Where participants are children, as in Piaget and Inhelder’s Three Mountains task, a parent or guardian gives consent on the child’s behalf.
Deception happens when participants are not told the true aim of a study, or are misled about it. Researchers only use deception when it is unavoidable — for example, when knowing the aim would change how people behave — and they deal with it by fully explaining the true purpose in a debrief afterwards.
Confidentiality means personal data is kept private. Researchers deal with this by storing data anonymously, using numbers or initials instead of names, and only reporting results in a way that does not identify any individual.
The right to withdraw means a participant can leave the study at any time, and can ask for their data to be removed even after taking part. Researchers deal with this by reminding participants of this right at the start and during debriefing, and by never pressuring anyone to continue.
Protection of participants means researchers must not expose people to risks greater than everyday life. Researchers deal with this by stopping a study if a participant becomes distressed, and by offering support after any study that could cause upset.
The five named ethical issues are: informed consent (information sheet and signed agreement), deception (only when unavoidable, resolved by a full debrief), confidentiality (anonymous data storage), right to withdraw (can leave, or remove data, at any time without penalty), and protection of participants (no more risk than everyday life; support offered if needed).
Ethical issues appear throughout the compulsory topics. Piaget and Inhelder’s Three Mountains task and Gunderson et al.’s praise study both involved children, so parental informed consent and protection from distress were essential. Caspi et al.’s study of genes and depression involved sensitive personal and medical information, raising confidentiality concerns about how genetic data was stored and reported. Young’s study of CBT for internet addiction involved real clients in treatment, raising confidentiality of therapy records and protection from harm during sensitive discussions.
Haney, Banks and Zimbardo’s simulated prison study is a well-known case where protection of participants broke down: some ‘prisoners’ experienced real psychological distress, and the study was stopped early. It raises questions about whether the right to withdraw was genuinely respected when participants felt they were playing an assigned role. Piliavin et al.’s subway field study used deception — passengers did not know they were part of an experiment and could not give informed consent in advance — which is typical of field experiments and raises the question of whether a debrief was practical at all.
Ethical Issues as an Ongoing Debate
An ethical issue is any concern about how a study’s design or conduct might affect participants’ rights or wellbeing. Psychology is often described as facing a trade-off: valuable scientific knowledge sometimes requires methods, like deception or distressing scenarios, that create ethical risk. This debate runs across all five compulsory topics — from studying children’s cognitive development to studying obedience and conformity — and you need content from more than one topic to discuss it.
Ethical issues are not abstract rules — they show up in real named studies across every compulsory topic. Being able to name the issue, explain why it arose in a specific study, and state the practical way of dealing with it is what earns marks in Edexcel’s extended questions.