What is acquiescence bias and how can you stop it?
Acquiescence bias, also known as the agreement bias, is the tendency for survey respondents to agree with research statements, without the action being a true reflection of their own position or the question itself.
Typically, the issue will come up when you ask a participant to confirm a statement, or if the question is answered with opposing pairs, such as ‘Agree / disagree’, ‘True / false’ and ‘Yes / no’.
For example, in this Yes / no question, a participant might answer ‘yes’ if they have acquiescence bias:
That means that, regardless of the importance of the question that’s being asked, the participant’s results might be skewed to favor the ‘Yes’ answer.
This brings serious accuracy concerns to the data quality of research. It can lead to incorrect information being used by teams, which leads to uninformed actions and decisions.
What causes acquiescence bias?
Participants have different backgrounds
There are some elements in a person’s background that may influence whether they may succumb to acquiescence bias. One study found that 15% of acquiescence variance came from country-level variations in corruption levels and collectivism, showing that external factors can influence submission to the acquiescence bias.
Education is also a factor. The same research found that where a participant has a low education level and a low degree of conservatism, the responses had a higher tendency towards acquiescence bias.
Participants are impacted by their ideal version of themselves
Participants can be influenced to change their behavior, by agreeing to be part of a survey. The very act of being part of a survey could alter first how they think of themselves, and then how they answer the questions.
Another research study found that participants who took a survey ended up using it as a way to share the views of their own ideal version of themselves. In that way, the survey was “transformed, from an inquiry about ‘what I do, to ask about ‘who I am.’”
Participants are influenced by the researcher
It is expected that participants would show a higher acquiescence bias on interviewer-administered surveys, than on the self-administered questionnaires (like those surveys sent by email or website). However, this isn’t the case.
Participants on self-administered surveys may want to see behind the questions to the researcher’s intent when creating the survey. This effect could be increased if they are given prior knowledge about the survey, or contribute to its making. This could result in unnatural responses that please the researcher, or aim to fulfill what they believe that the research wants.
(American psychologist Lee Cronbach opposed this theory, instead saying that a participant uses their own memories to form an opinion on the statement. Cronbach’s theory explains why there could be two contradictory answers in a survey response: One answer is based on pleasing the researcher, while the other is gained from the participant’s position based on their own memories – the two answers are truthful, but just happen to be at odds with each other.)
Participants lack the motivation to engage with the survey
Some respondents, including those who are not highly motivated to think through the questions, take mental shortcuts when they are responding to questions. This can occur if the survey is:
- Not aimed at the right audience
- The questions are not clear enough
- Does not engage the participant enough to want to complete it
- The survey is too long
- The participant does not have the time to complete the survey
This tendency to answer positively is one of those common shortcuts, along with selecting ‘middle-ground’ answers without variance or thought.
Participants want to keep to socially desirable attitudes
When participants are aware that there will be someone looking at their answers, or that the results can be tied back to them, this could cause anxiety about how their data will be used.
In these cases, where there are questions that ask confirmation questions about a socially unpopular research statement, the answer which appears more agreeable to the social norm will be selected.
For example, asking ‘would you consider yourself to be a sociable person?’ could get a stronger ‘yes’ reaction, than ‘no’, even if the answer isn’t true. This would be because the participant thinks being sociable is better than the opposite.
Participants don’t believe a ‘middle-ground’ answer exists
There are some questions that can bring out extreme beliefs of a participant, who may hold a strong view on an issue. For example, a question asking ‘Do you agree that reducing greenhouse gas emissions is everybody’s responsibility?’ is leading the participant to answer either ‘agree’ or ‘disagree’, depending on their view – even if there was a ‘slightly agree’ or ‘neither agree nor disagree’ option available.
The acquiescence bias occurs when a participant holds such a strong belief towards one view, that it overcomes the rating stage of the decision-making process. They don’t see a middle-ground existing, which produces skewed results in favour for the ‘extreme’ answers at each pole.
5 ways to avoid acquiescence bias in survey
1. Reformulate the question
By reformulating the response formats and options to correspond more closely with the subject of question, you also make interpretation easier for the respondents. For example, read this question:
The risk of acquiescence response bias can be reduced if the question was reformatted to simply ask how satisfied they feel about their experience.
Here is the question reformulated in a way that avoids asking for agreement below:
2. Introduce measures to help the participant’s focus
It’s possible to include more information about the question itself, by including subtitle text under the question. This helps participant to focus on what’s being asked.
This method raises understanding and removes the likelihood that they will skim over the content of the question, or fail to understand what is being asked.
3. Plan survey participant group to include the right survey participants
Responses to acquiescence bias can vary if participants from more than one country are surveyed at the same time, due to cross-national differences. When looking at creating survey, think about which participants would be best to include and where they are located.
Being conscious of the different demographic information about participants, can give you more accurate results in the context of participant’s backgrounds. This helps you at the time of analyzing results, as you can judge the results with knowledge of the background factors.
4. Be sensitive in role as the researcher
Bias created by the researcher is a core ethical consideration for anybody, and it’s still possible to occur in self-administered surveys.
Be careful in asking for confirmation on overly emotional positions. This can avoid scenarios where participants feel cornered into a single position, where several exist. Instead, consider how question can be positioned in a neutral and non-alarming way, so that each answer will be given due consideration
If you need to ask sensitive questions, try out a different research format to collect data. This can be through video interviews or in-person survey collection, which gives the participant the opportunity to expand their answer and explain the rationale behind their answers.
5. Reduce anxiety about the survey with transparency
The social norm of appearing agreeable and polite can bias respondents, so consider if you can keep participants anonymous to gain a high level of truthfulness. This might be easier with topics that are considered socially taboo, or if there is likely to be social harm done to the participant if their answers are made public.
You can also prevent anxiety about the survey, by clearly stating how the participant’s data will be used, and confirming that they can opt-out if they don’t feel comfortable completing the survey. This can prevent the results from being diluted by inconsistent or incorrect data, and give participants a choice about whether they want to go ahead. In choosing to continue, they’ll more likely be willing to answer honestly as they’ll feel in control.