BREAKING: Design Students No Longer Blindly Trust Data!
October 28, 2019

REFLECTING ON:  The misleading nature of data in the way it is presented to the public as all-encompassing evidence.

How might we make campus life more manageable, enjoyable, and productive for senior art and design students?

With this challenge in mind, my peers and I spent 19 days tracking data variables individually that we each found relevant to success, or lack thereof, in our personal and academic lifestyles. After the 19 days, we swapped data sets along with corresponding personas and proceeded to examine the collected data in relation to our HMW (how might we) question. After receiving my data set, I sifted through the student’s five variables — activity, background audio, location, caffeine consumption, and stress level — collected hourly over the 19 day period. Interested in how the stress level might have been impacted by other variables, I searched desperately for any correlation until a pattern finally appeared: a high stress level was recorded in the hours following caffeine consumption most of the time. With this small discovery from one student’s data, I developed the hasty argument that caffeine leads to higher stress levels, universalized it by applying it to all senior design students, and demonstrated it to the public with an infographic poster.

The argument is visualized with three graphs. The first documents how often caffeine was consumed in the few hours before each stress level was recorded, with the percentage clearly increasing with higher stress levels. The second graph shows how frequently caffeine was consumed at each hour compared to the average stress level recorded at each hour, suggesting that stress levels increase after the most frequent times of caffeine consumption. The final graph visualizes the total stress level recorded each day and highlights the days where caffeine was consumed, showing a clear trend in larger amounts of stress when caffeine was present and smaller amounts of stress when it was not.

The poster layout resembles a traditional newspaper with a Breaking News headline to mock the way data is presented to the public as irrefutable evidence to an argument, and how the public is so quick to accept this “news”. The data is visualized with coffee spill themed graphics to further the connection between stress and caffeine visually.

This argument seems straightforward enough, and it might even be somewhat true for this student, but there are many human-centered problems with coming to this conclusion. First, the data only represents one student’s experience throughout a very small time period, but it is generalized and applied to all design students. The argument would have been more sound if it came from a larger data set representing different extremes of design students’ lifestyles. Speaking of extremes, the persona which accompanies this student’s data shows that the student struggles with anxiety, meaning that there could be factors which contribute to the recorded stress levels which do not apply to students who don’t fit this extreme. Even in the data that was recorded, there were other environmental factors which were not considered in the argument including the background audio, the activity, the location, the amount of sleep the night before, and even the day of the week and time of the day. Each of these variables could have an impact on the student’s stress level, but they were not taken into account when patterns were found between stress and caffeine.

This data is presented with a very argumentative stance, despite the limitations of the data. This gave me insight into the way that data is distributed to the public as concrete evidence even when there is more to the story. Even if the data doesn’t blatantly lie, it often conveniently excludes key factors which don’t align with the argument, preventing people from questioning its validity. Going forward, I will consume data with more caution and scrutiny. When I use data to further my own argument, I will aim for a complete data set which considers human-centered factors such as qualitative data and personas from which the data was collected. I will also be transparent about the data I reference and its limitations so that the audience is equipped with honest information from which they can inform their opinion.

This post exists within a series of reflections on  topics and coursework from my Human-Centered Design class, taught by Marty Maxwell Lane at the University of Arkansas.