Bias is critical to consider when carrying out any form of research, including market research. There are various types of bias, which generally fall into the categories of researcher bias and responder bias. Bias can result in inaccurate data, and in the end, false or even harmful conclusions being drawn. Minimising bias in market research leads to more effective business decisions.
This article serves as a comprehensive guide for handling bias in market research. You will understand what bias is and how bias affects research and surveys. Types of bias in research will be detailed as well as how to deal with types of unconscious bias. You will also learn about how SurveyMonkey can help reduce bias in your market research.
Bias is an outlook that is either inclined or prejudiced towards something, usually in a way that is unfair. Biases can be developed towards individuals, groups or beliefs. For example, a person may be biased against the idea of putting pineapple on pizza and never try the combination, leading to them maintaining their un-evidenced opinion.
Bias can be very damaging to research and surveys, distorting results and leading to invalid conclusions. It can creep in via the researcher or respondents, or both. The biased questions or answers skew the data. Some bias is inevitable, but it can be controlled and minimised for better quality data.
Five common entry points of bias in market research and surveys are:
Having identified ways bias can manifest in market research and surveys, let’s now look at the types of bias researchers and respondents are vulnerable to. Understanding types of bias in research is an important step in minimising their impact.
Scientific research suggests that bias is hardwired into our brains, biologically or psychologically. As human beings, researchers are therefore vulnerable to common types of bias. The types of bias highlighted below can affect market research if not accounted for.
Confirmation bias occurs when information is sought or interpreted in favour of pre-existing beliefs. For example, a researcher may already assume their hypothesis is true. Instead of investigating the hypothesis with an open mind, they may favour supportive information and dismiss counter-evidence – and their research becomes a self-fulfilling prophecy.
Implicit bias is when the researcher has a predisposed preference for or aversion to a group of people or person, as opposed to being neutral, without consciously knowing it. For example, they may unconsciously stereotype individuals. Therefore, implicit bias is a type of unconscious bias. It is problematic because by definition the researcher is unaware of it, making it more difficult to control and minimise.
Cultural bias applies when the researcher makes assumptions about individuals’ motivations and influences through a cultural lens. This commonly occurs as an ethnocentric bias: judging other cultures solely based on the standards and values of one’s own culture.
The halo effect is seeing something a certain way based on one positive attribute. For example, a researcher may develop assumptions about a respondent based on one positive answer. Or, by jumping between research topics, they encourage the respondent to project their opinion about one topic upon the other.
Respondents are human beings, too, and as such, vulnerable to the same common types of bias as researchers. For instance, they could exhibit confirmation bias by interpreting a question in line with their pre-existing assumptions. However, in the context of market research, respondents are also vulnerable to more respondent-specific types of bias that could affect the research if not controlled from the outset.
Yea-saying is when the respondent has a tendency to agree with the researcher. It can occur when respondents are tired, because the respondent has an innately acquiescent personality or because the respondent sees the researcher as an expert.
Social desirability bias comes into effect if the respondent answers questions in line with what they believe will be most socially acceptable, instead of giving their honest point of view. There is usually more risk of social desirability bias with research that touches on more sensitive or personal topics for the responder.
Habituation bias is when the respondent gives the same answers to similar questions. It is a biological function of the brain to conserve energy, but in so doing, the respondent does not pay as much attention as time goes on. Habituation bias can result from interview or survey fatigue, or if there are too many similarly worded questions in the research.
Sponsor bias occurs if the respondent allows their feelings or opinions about the sponsor of the research to cloud their answers. Even a suspicion of whom the sponsor may be is enough to trigger sponsor bias.
The table summarises the eight types of bias detailed above and shows where market research is vulnerable to being affected by each type of bias (researcher vs responder):
| Type of bias | Researcher | Responder |
| Confirmation bias | Yes | Yes |
| Implicit bias | Yes | Yes |
| Culture bias | Yes | Yes |
| Halo effect | Yes | Yes |
| Yea-saying/Acquiescence bias | No | Yes |
| Social desirability bias | No | Yes |
| Habituation bias | No | Yes |
| Sponsor bias | No | Yes |
Bias is most often unconscious. That is, the researcher or respondent is unaware they have the types of bias in research spotlighted in this guide. All of them can be types of unconscious bias. For example, a researcher may not realise they have confirmation bias and will have framed their questions using pre-existing assumptions about their research hypothesis. Or a respondent may not realise their answers have been impacted by social desirability bias.
As mentioned, bias can enter the market research or survey at various points: namely, the questions, the sample of respondents, the data and data analysis. At every point, it is essential to minimise the effects from the various types of bias, conscious or unconscious. Accounting for conscious bias is easier; unconscious bias, being less obvious, is harder to deal with. However, there are practices that control for types of unconscious bias.
To minimise unconscious bias, ensure that any pre-existing assumptions on the researcher and respondent sides are eliminated. Prepare questions without any hint of the researcher’s opinion on the research topic or presumptions about the respondents and their responses – and keep a close eye on the order in which the questions appear. The choice of respondents should be based on non-judgemental methods such as probability sampling. During the research interview, respondents should feel comfortable and able to give their honest points of view without fear of adverse reaction. When analysing the data, the researcher must remain open-minded and neutral to its results.
Bias is a phenomenon that can greatly weaken the validity of research. Bias is an inclination (or aversion) towards something based on pre-existing beliefs, usually unfairly. Its opposite would be absolute neutrality. Minimising bias in market research improves its quality and the accuracy of its conclusions, leading in turn to better business decisions.
There are eight types of bias that commonly impact market research:
These types of bias can be unconscious. That is, the researcher or respondents are unaware they are biased. Dealing with types of unconscious bias is challenging; however, there are practices researchers can follow to minimise the impact of bias in market research during each stage of the process.
SurveyMonkey has a number of tools and features that tackle bias and minimise its effect in market research, strengthening the quality of your data and conclusions. SurveyMonkey’s questionnaire examples and survey templates, for example, offer a ready-made bank of expertly-phrased, neutral questions. To improve your questions further, try SurveyMonkey Genius, which brings together AI, survey experts and machine learning to give you personalised recommendations. And you can take the stress out of respondent sampling with SurveyMonkey Audience, which gives easy access to your target market so you get accurate feedback in minutes.
Bias is an inclination or prejudice towards something, someone, a group of people or an idea based on pre-existing assumptions or beliefs, usually in a way that is unfair.
Eight common types of bias in research are:
● Confirmation bias
● Implicit bias
● Culture bias
● Halo effect
● Yea-saying/Acquiescence bias
● Social desirability bias
● Habituation bias
● Sponsor bias
Bias can be highly damaging in research. If it is not controlled and minimised, bias in research can result in poor quality data and, ultimately, invalid conclusions.
Unconscious bias is when a person does not consciously know they are biased. That is, they are unaware they have a bias.