An important aspect of reading critically is considering what forms of bias may show up in the materials you encounter, whether those are journal articles, book chapters, practitioner reports, government documents, business analysis, theatre reviews or industry podcasts.

The tables below outline some types of bias you might encounter, that you would be expected to acknowledge in your work:

Sample bias

Sample biases are those related to a study sample:

Selection bias	Selection bias is where the people in the study are not representative of the overall population they are supposed to represent.
Performance bias	Performance bias occurs when there are differences in the way that groups are treated during research, distinct from the difference which is being studied. For example, if a trial compares a new treatment with a current treatment, the only difference between the two groups should be that one group gets the new treatment and the other gets the current treatment. Any other differences such as how often people are seen or time spent with researchers could affect the results.
Detection (or observer)	Detection or observer bias can occur where researchers see the answer they want to see. This can be particularly prevalent when the researchers know which group individuals belong to. Blinding (when either the patient or the researcher do not know which treatment the patients have been assigned to) or double blinding (when the patient and the researcher do not know which treatment the patients have been assigned to) a trial can help to prevent this. However, blinding or double blinding a trial is not always possible.
Attrition	Patients often drop out of a study for many reasons. However, if too many participants drop out of a study this can lead to attrition bias. This tends to be even more of a problem if the drop-out rate is unequal between groups.
Response	Response bias occurs when participants respond to researcher questions with answers they think the researcher wants to hear, rather than honest responses.

Interpretation bias

Interpretation biases are those related to the researcher's interpretation and reporting of the data:

Confirmation	Confirmation bias can occur when researchers interpret the evidence in a way that fits their existing beliefs or theories and overlook the evidence which doesn’t fit.
Publication or reporting	Publication or reporting bias occurs when only positive results are published or reported.

Funding/sponsorship bias

Funding or sponsorship biases are where the incentives of the funders introduce bias to the research. For example, a funder may have a vested interest in finding favourable results to advance their business interests.

Acknowledgement: Content adapted from Cardiff University Library Service (2024) 'Critical evaluation: reading and critically analysing a journal article'. Available at: https://xerte.cardiff.ac.uk/play_18649 (Accessed: 26 June 2024). Licensed under CC4.0.