In the third and final segment of our body & facial expression webinar series, we will talk about how to automate the analysis of facial responses to understand your user’s emotions. Facial analysis is one of the least intrusive methods for capturing real-time emotional reactions in a highly automated manner. These systems use computer algorithms that take video from a common webcam and provide frame-by-frame emotion metrics as outputs.
Current methods for collecting emotional responses often require a participant to constantly verbalize their thoughts and feelings. This approach is unnatural and may yield unreliable results. Facial metrics can be captured in the background as a participant performs a series of tasks while evaluating a website. The software can automatically pinpoint when a person is feeling certain emotions and can be tied directly to what they are doing.
30 Virtual Events Hosted & Counting
30
Watch KLI’s Webinar:
Join our Director/Principal Researcher, Andrew Schall, and CEO of Kairos, Brian Brackeen, to learn about the latest tools and methods used to analyze facial expressions as well as understand the opportunities and limitations of using this technology in user research. During this webinar we will cover:
An overview of available facial analysis tools
Demonstration of facial analysis software
How to integrate facial analysis into your existing user research methods
Discuss the pros and cons of using facial analysis
In a recent collaboration with User Experience Magazine I published an article entitled: The Future of UX Research: Uncovering the True Emotions of Our Users. The motivation for writing this article was to expose the UX community to the possibilities of using biometric and neurometric measurements to understand the emotions of our users. Current methods for understanding a user’s emotional response are at best limited, and at worst, entirely inaccurate. As the field of user experience evolves, we need to explore new methods for measuring emotional responses using technologies borrowed and refined from neuroscience and human biology. I also wanted to highlight some of the latest technologies available to user researchers, as well as the challenges of working with these tools.
I was thrilled to receive so many positive comments about the article as well as many questions about using these tools. This recap provides some of the key aspects of using biometrics for user experience research as well as answers to the most frequently asked questions about the topic.
Here are some do’s and don’ts for using biometrics in your user experience research projects:
Do:
Hire specialists on your team with a neuroscience, cognition, or experimental psychology background (preferably a Ph.D.). They will be most helpful during the study design and also during the analysis of your biometric data.
Run lots of pilot tests. Incorporating biometrics can add unforeseen challenges to your study and requires extra practice to ensure that data is collected accurately with actual participants.
Ensure that you have a sufficient sample size. Biometric studies require more participants than a typical usability test to account for the potentially large variability between participants. You will also encounter more situations where data needs to be excluded because due to improper equipment calibration, equipment failures, etc.
Don’t:
Examine biometric data in a vacuum. It is important to triangulate data across a combination of different methods including nonverbal observations, participant retrospectives, surveys & rating scales.
Assume that your data is perfect. Measuring biological and neurological responses, especially with the tools accessible to UX researchers, is not going to be 100% accurate. I would recommend looking at general trends in valence (i.e. positive/negative spectrum of emotions) changes instead of trying to pinpoint exact moments where very specific emotions were felt.
Questions & Answers
There have been a variety of tweets and posts generated as a result of publishing this article. Here are some of the most common questions with my thoughts:
Do participants mind wearing all of that equipment? Does anyone complain about it being uncomfortable?
Before making any purchases, I heavily researched all of the existing hardware currently available to make sure that it was as minimally intrusive to participants as possible. In many cases we are only using one or two measurements at a time. I try to use only one measurement that requires physical contact with the participant, such as wearing a GSR wristband combined with facial analysis that only requires a webcam. At the end of a session I always ask the participant whether wearing any of the gear was uncomfortable for them or if it impacted their experience at all. Most said that once the session began they quickly forgot that they were wearing anything.
Is it [biometrics] worth it? Wouldn’t it just be easier to interview an individual or give them a survey to complete?
We are at the very beginning of an exciting journey to uncover more about how our users are actually feeling. Biometrics is definitely not for everyone, and it is not a useful endeavor for all situations. I actually don’t think that these tools are ready for mainstream UX researchers just yet. The tools and software continue to get better every day, and will likely become more useful to UX researchers within the next few years. However, for those who are very serious about obtaining objective, quantitative measurements about your user’s emotional journey, I would recommend trying out these tools to see if they are useful for your team.
I’m very interested in learning more about biometrics! How did you learn about this topic? Would you recommend any resources?
There really isn’t one place to find out everything you would want to know. Biometrics is still a very new topic for the UX community and most of the resources out there are geared for people in the fields of human biology and neuroscience. I started reading a lot of academic papers to understand how to collect and analyze the data. There are a few worthwhile articles in the ACM Digital Library that provide case studies where biometrics were used in the context of human-computer interaction. I would also recommend looking into each biometric and neurometric measurement (e.g. EEG, GSR, etc.) individually. You can get a basic sense of how these tools work by visiting the websites of the device hardware and software manufacturers. My article provides links to many of the major vendors who make UX researcher-friendly products.
