Confirmation bias in UX
Peter Wason discovered confirmation bias in 1960 and it refers to the tendency to let prior beliefs influence the way new information is perceived.
It is a cognitive bias that occurs when people analyze information in a way that fits their beliefs or ideas. The impact of confirmation bias is greater when the topics are related to deeply rooted emotions or beliefs, such as religion, race, or politics, for example.
Confirmation Bias in UX Research
Confirmation bias is a concept that must be kept in mind when we are dedicated to UX research, as it can result in disregarding options without relying on objective data, and it is very important to make decisions based on data.
We must avoid biases in user interviews, otherwise we can find ourselves facing the problem of validating hypotheses without having an objective reason, simply because we unconsciously formulated the questions to obtain the results we wanted.
It is also important to avoid biases when we do Card Sorting, as if we have many categories we will not be able to include them all in the test, and we will have to choose a sample that is representative enough to avoid biases.
In conclusion, understanding the importance of confirmation bias is key in the realm of user experience, and we must be aware of it to avoid such bias.
How Confirmation Bias Works in UX
For example, if we are investigating an e-commerce, we will have to carefully choose our words so that the questions are not biased. If we want to ask about the payment process, we could pose it in different ways, such as:
- Was it difficult for you to find the payment button?
- How was your experience with the payment process? Explain everything you liked or disliked during the process.
Both options ask the same thing, but the first one is an example of confirmation bias because the question is biased to collect evidence about the hypothesis that the payment button is not well located, although it might happen that this was not the biggest problem. Also, a negative language is used, which can lead the participants to think that there are indeed problems with that button. Furthermore, the question is formulated to answer yes or no, so there is no space for users to explain what problems they found in the payment process.
On the other hand, the second question is more objective, as it does not focus on what could be the problem. On the contrary, it offers the users the opportunity to explain how they felt while carrying out the payment process task and what real problems they encountered.
Tips to Prevent Confirmation Bias in UX
Understanding what confirmation bias is and what impact it can have on UX is key to actively working to avoid it. To overcome confirmation bias, some tips are recommended that can help avoid it:
- Obtaining data in the early stages of the project. It is very important to thoroughly research before moving on to the development stages. This way, designs will be data-based from the start, and the emotional involvement of designers and developers will not affect user research.
- Investigate instead of validate. Bias-free research should be conducted in all phases of the project. Although research may collect information contrary to the design decisions made initially, those in charge must accept it and iterate based on the data, setting aside the emotions of “loving” the product as designed.
- Ask impartial questions. As mentioned in the previous example, it is very important to properly formulate questions to avoid biased responses. We cannot ask to refute or affirm, as we may fall into cognitive biases and obtain data with low validity.
- Use triangulation. A good way to avoid confirmation bias is to obtain information through different methods, so that the data will be much more credible. The purpose of triangulation in UX is to seek different perspectives to verify results.
In the realm of user experience, it is very important to understand what confirmation bias means, as it can negatively affect the ability to perform objective research and interpret results.
If you are interested in learning UX Research don’t miss our UX Learn trainings (in Spanish):
This is a translation of the following article from our corporate website:
