What is Extreme Response Bias?

There is no denying the value of survey research, especially online survey research. It is the only way to gather information on a wider population in a way that is analyzable and less prone to bias, so that you can make decisions that are in the best interests of your company.

But of course, “less prone” does not mean “immune to.” There are still some biases that can be introduced to surveys. One of these biases is known as “extreme response bias.”

What is Extreme Response Bias?

Extreme response bias is the tendency of a respondent to answer in the extreme, even if they do not have an extreme view. For example:

“How much do you like the New Pepsi?”

Where the respondent answers with a 5 out of 5 even if they really only liked it a 3 or a 4. Some respondents have a tendency to pick the extreme response rather than the middle. They either “love” or “hate” everything, choosing the least or highest response even if that is not their true stance.

Why Does Extreme Response Bias Occur?

There are many different theories for extreme response bias. A few include:

• General Indifference: Someone that is not passionate about answering a survey honestly may be more prone to answering in an extreme, especially a positive extreme. Interestingly the opposite is also true – a respondent may start entering the middle response in an effort to speed up the survey when their answers are actually more extreme. But there is no denying that a person that is indifferent to the survey and feels generally positive may be responding extremely positive.
• Willingness to Please: It may also be a willingness to “please” the company. Many people have the desire to please others, and (especially for those that are not passionate about answering honestly), a respondent may answer the way that they think the company wants them to answer. In the “New Pepsi” example, the respondent may be expecting that the company wants them to love New Pepsi, so they answer in the extreme affirmative. (Note: The opposite extreme may also be true. The respondent may not want change, so they answer in the extreme negative to indicate their dissatisfaction, even if their true belief lies somewhere in the middle).
• Wording: Question wording may also be to blame. If the question is worded in a way that the respondent feels that an answer is desired, they may give the desired answer. This is one of the reasons that it is always important to analyze every question for any introduced biases, to make sure that there is no pushing of the respondents towards a single, specific answer.

Research has also shown that there are characteristics that lead to this type of extreme response bias. Some cultures tend to provide more extreme responses, and those of lower intelligence tend to be more likely to provide extreme responses as well. There are potentially many similarities, and these may be considered with your sample.

How to Reduce Extreme Response Bias

The most important thing you can do to reduce extreme response bias is make sure that there are no bias questions and that your survey is online and not given by a specific person. People are more likely to respond in the extreme when they are afraid of being judged. An anonymous online survey should reduce extreme response bias.

In addition, providing either fewer responses or cutting down on total survey time may prevent some people from responding in the extreme. Figuring out new ways to ask your questions may as well. Some suggest putting a “feeler” question that is purposely bias to see how the person answers. Those that answer in the affirmative to a bias question may be more likely to provide extreme responses.

Nevertheless, this is still something to watch out for in your surveys. It doesn’t necessarily mean that your data isn’t ready to be analyzed. It will be. But it does mean that you should at least look at your data and your sample carefully to ensure that extreme response bias isn’t altering the results.