7 Biases to avoid in qualitative research
Qualitative research is an exploratory scientific method of observation to gather non-numerical data. Rather than determining ‘counts or measures’ as in quantitative research, qualitative research involves a description of things, related characteristics and meanings, and basic observations and interpretations. Some of the common approaches to conducting qualitative research include interviews, participant observation, and focus group discussions. Often, this approach is widely used in the political science, education, and social work fields and may also be applied to market research, business, and journalism.
Although scientific or academic research needs to be handled objectively, the subjective nature of qualitative research may make it difficult for the researcher to be detached completely from the data, which in other words means that it is difficult to maintain objectivity and avoid bias. Bias, defined as the “inclination or prejudice for or against one person or group, especially in a way considered to be unfair,” can be extremely detrimental to scientific progress as it can lead to the distortion of reality and thereby affect the validity and reliability of research findings.
As a result, qualitative research is often criticized for lacking transparency and scientific rigour. It is censured for being a compilation of impressions that are subjected to researcher bias and that may not necessarily be reproducible. However, qualitative research has the benefit of being flexible and allowing creativity; thereby providing additional insights that cannot be collected through quantitative research.
What then are the types of biases faced by qualitative researchers? And are there any means of avoiding them. Once you understand and identify the different types of biases, it will become easier to take measures to avoid them as well.
The two main types of biases
Broadly, biases can be of two types – participant bias and researcher bias. Participant bias stems from the respondents or participants responding to the questions based on what he or she thinks is the right answer or what is socially acceptable rather than what he or she really feels. Another aspect that may introduce participant bias is if the participants are opinionated about the sponsor of the interview, which could lead them to agree either to everything or nothing proposed to them.
On the other hand, bias from a researcher’s end may get introduced if researchers unknowingly interpret data to meet their hypothesis or include only data that they think are relevant. They might ask questions in an order that may affect the participant's response to the next question or ask leading questions that may prompt a certain response.
While completely avoiding bias is impossible, there are several means to reduce it. Most importantly, identification of a potential bias can aid in taking appropriate preventive measures. The table below lists out the potential types of bias to look out for as well as appropriate preventive measures for each type.
Types of biases and how to avoid them
Type of biases
How to avoid bias
Acquiescence bias or friendliness bias occurs when the participant chooses to agree with the moderator or researcher. Some participants may agree just to complete the interview, and this typically happens once fatigue sets in.
Frame questions that are open-ended to prevent the participant from simply agreeing or disagreeing, and guide him or her to provide a truthful and honest answer. If the answers do not sound true, ask the question in different ways. Alternatively, use direct questions that allow the participant to select from a range of potential choices rather than a ‘Yes’ or ‘No’ choice.
Social desirability bias or social acceptability bias. Many participants respond inaccurately just so that they could be liked or think that they would be better accepted, especially in cases of sensitive questions or questions on personal or controversial topics.
Again, phrase the questions in a manner that allows the participant to feel accepted no matter what the answer is. You can also opt for indirect questions that ask what a third party would do in a particular situation. This will help the participant to project his or her own feelings onto others and provide accurate, truthful, and more representative answers.
Habituation bias occurs when participants provide the same answers in response to similarly-worded questions.
Ensure that different questions are worded differently and that the questions are engaging throughout the interview.
Sponsor bias. Such a bias can occur if a participant is opinionated about the sponsor of the research or gets influenced by the sponsor’s reputation or mission statement.
It is important for you, as researchers, to maintain neutrality so as to not influence the participants’ responses. So you should not give away any details about the sponsors such as the company logo or provide your role or the goal of the study.
Confirmation bias. This most common and highly recognized bias occurs when a researcher interprets the data to support his or her hypothesis. Researchers may also omit data that does not favour their hypothesis.
Consider all the data obtained and analyse it with a clear and unbiased mind. Continually re-evaluate the impressions and responses, and ensure that pre-existing assumptions are kept at bay.
Question-order bias. Some questions may influence the responses to subsequent questions. Participants may compare and judge subsequent questions based on their response to the first question resulting in a biased and inaccurate answer.
Consider potential bias while constructing the interview and order the questions suitably. Ask general questions first, before moving to specific or sensitive questions.
Leading questions and wording bias. Questions that lead or prompt the participants in the direction of probable outcomes may result in biased answers.
Keep the questions simple and be careful to avoid words that could introduce bias. Do not use leading questions that can prompt the participant to respond in favour of a particular assumption.
While research that is completely unbiased is ideal, it may not always be possible. But it is important that researchers consider and address every detail at the study design stage to avoid bias. As you must have noted, most biases can be avoided by framing the questions and structuring the interview skilfully. It may also be a good practice to have the questions and the data reviewed by a colleague or mentor for a second, unbiased opinion.
Thus, the attempt must be to stay objective and try to minimize bias throughout the entire research process. Further, for added transparency, while writing the manuscript, indicate the efforts taken to avoid bias while recruiting participants, qualitative questioning, and also while analysing and reporting.