Skip to main content

Glossary of Research Terms

As more and more young people take part in BrainWaves research studies, the more questions we get about what some of these research terms mean. Here are a list of the key terms students might encounter in research, with easy-to-understand definitions…


Any factor which is likely to skew the results of a study in a particular direction. This could be for the entire study, or for a specific sub-group within a study.

Cohort study
A cohort study is a particular form of longitudinal study where researchers track the change of behaviour in the same group of individuals, who share certain characteristics, across a period of time.

Data processing
Data processing occurs whenever any data is collected (from participants) and translated into usable information. This is usually performed by data scientists and researchers.

De-identified data
Data where all personally identifiable information has been removed from a data set.

Ethical approval
Ethics approval is needed for all research involving human participants, or data from which individuals could be identifiable. University of Oxford is committed to ensuring that its research involving human participants is conducted in a way that respects the dignity, rights, and welfare of participants, and minimises risk to participants, researchers, third parties, and to the University itself.

Linked to representativeness, generalisability refers to the degree to which the results of a study can be generalised to a broader population or context. For example, if taking a cold shower improves your mental health, to what extent can this be generalised to the claim that cold showers would improve other people’s mental health.

Interventions, in the context of psychiatry, are actions performed to bring about change in people. They can be targeted towards a wide range of behaviours, such as bullying, depressive symptoms, wellness etc.

The weakness of a study or ways in which a study could be improved in future. No study has no limitations, the important thing is for researchers to be open about the limitations and recognise them, not to try and hide them or claim they don’t exist.

Linkage key
A linkage key allows researchers to identify multiple data sets from the same participants. Since BrainWaves is a cohort study, it is important to know how a person’s behaviour changes over time.

Longitudinal data
Data that is collected with the same group of individuals across a period of time measuring a standard set of criteria.

Peer review
The process by which scientific studies are shared with other independent scientists who check whether the methods used and conclusions drawn are appropriate. This is often considered one of the most important features of scientific research.

Personal data
Data that is related to an individual who might be identified from other data sets.

Pseudonymisation is a data processing technique that removes all information in a data set that can identify the participant. However, using a linkage key, researchers can still identify the data sets contributed by the same individual participant.

The degree to which a measurement consistently measures the same thing. e.g. a questionnaire to measure the degree of stress felt by students might be considered unreliable if some students answered the survey right before a 3 hour exam, while others answered after breaktime.

The degree to which a sample (the people in a study) can be considered to represent the population as a whole (the entire group of people which the study is hoping to investigate). For example, a sample of 16–18-year-olds in schools in London would not provide an accurate representation of all 16-18-year-olds across England.

About the author

Charlotte Chan is an assistant on the BrainWaves research project. She holds a BSc (Hons) in Psychology from the University of Bath,