Dopamine and the mind – why good designs go wrong
Over the years, I have noticed a strange pattern: when executives (site owners) are asked to comment on design layouts, they often say there is too much text and demand larger pictures/graphics – whether these are relevant or not. These executives are disappointed and frustrated with the design proposals they see. On the other hand, if you listen to users (during usability testing, for example), they complain that these same pictures/graphics are getting in their way. Like the executives, they also exhibit frustration, but in a diametrically different way – “Why are you making me scroll past this crap to get to the information I really need?”
My question was simple: was there a scientific reason for these dramatically different reactions to essentially the same designs? And I think the answer is “yes”.I’ve included a few salient footnotes for those of you who are scientifically inclined.
Thesis in brief (1)
Why do two groups of people seem to consistently disagree regarding the “attractiveness” of a website design? Could it be that there was a physiological reason for these reactions? In short, was our brain playing tricks on us or misleading us? Were our development and presentation techniques actually encouraging inappropriate client reactions?
I have known about the functions of neurophysiological “reward chemicals” since my pre-med studies at Washington University in St. Louis 1972-1976. In late 2007, having spotted the curious reaction pattern described above, I started to do some more serious research, focusing on the limbic system (2) and the nature of reward chemicals (3).
I made the assumption that if the pattern I had identified was universal, voluntary intake of recreational reward chemicals (e.g. nicotine, caffine, cocaine, etc.) was probably not at the heart of these reactions. So I looked for chemical rewards produced by the body itself. Soon, my inquiry zeroed in on dopamine, a chemical messenger similar to adrenaline. (4)
Dopamine – friend or foe?
Dopaminergic neurons appear to code environmental stimuli rather than specific movements. (5) This, in layman’s terms, means that pretty pictures stimulate dopamine release, which perhaps explains why executives favour graphics over blocks of text in dummy design layouts.
Although this reaction seems obvious (pictures are more attractive than text), it was reassuring to know that there was a scientific reason for this.
The second part of my question dealt with why test subjects so often reacted badly to eye-candy (i.e. gratuitous pictures/graphics).
There are various viewpoints as to the role of dopamine and the task-completion process. For example, Pennartz et al. (6) asked in 2009:
“Given the parallel organization of corticostriatal circuits, the question arises how coherent behavior, requiring integration of sensorimotor, cognitive, and motivational information, is achieved.”
Perhaps part of the answer to this critical question can be found in Taizo Nakazato’s research, published back in 2005 (7):
“During the task performance, dopamine concentration started to increase just after the cue, peaked near the time of the lever press, and returned to basal levels 1–2 s after the lever press.”
By way of background, this study deals with rats pressing a lever to receive a food reward. In internet terms, I equate this behavior with humans pushing a button/clicking a link to receive an informational reward. In other words, task accomplishment produces a reward – in this case chemical.
Actually, though, it appears that the anticipation of task-completion triggers dopamine release (8). And it could be that executives about to see a proposed design for the first time may be anticipating the presence of pretty pictures.
Yet the essence of the problem seems to be that if something delays/hinders task completion, dopamine release actually causes post-action frustration. Dr. J.G. Fleischer describes this phenomenon quite succinctly: (9, 10)
“If the [subject] does not receive the reward when it expects to receive it, then there is a depression of dopamine release, which is consistent with the negative preduction error that would occur in that situation.”
In other words, if something gets in the way of task completion, dopamine doesn’t get where it’s needed (“depression of dopamine release”). I suggest that perhaps the pretty pictures and eye-candy that were anticipated and appreciated during the presentation phase, are actually getting in the way of test subjects who expect a more relevant response to their query (i.e. clicking on a promising link). If we make people scroll to get to the stuff they want (and expect to receive), they experience dopamine depression.
That said, a more recent study by Wanat et al. (11), suggests that further research is needed:
“[The] enhancement of reward-evoked dopamine signaling was also observed in sessions in which the response requirement was fixed but the delay to reward delivery increased, yoked to corresponding trials in PR sessions. These findings suggest that delay, and not effort, was principally responsible for the increased reward-evoked dopamine release in PR sessions. Together, these data demonstrate that NAcc dopamine release to rewards and their predictors are dissociable and differentially regulated by the delays conferred under escalating costs.”
In other words, the tougher it is to achieve a result, the greater the dopamine reward. This somewhat contradicts my thesis – and yet these findings also indicate that the response is situational. Hence, I feel certain that Wanat & Co. are actually looking at a different side of the problem, unrelated to task-based frustration, but that related to task-completion in a triumphal ”I just made it to the summit of Mt. Everest” kind of manner.
Drawing on my network
In late 2009, my online research led me to my grade-school best-friend, Jon Kassel. (12) Jon is now Professor of Psychology at the University of Illinois. Jon’s speciality is addiction. Naturally, the effect of drugs on emotions represents a key part of his own research.
Jon and I chatted informally about the problem with which I was wrestling. And without putting too many words in Jon’s mouth, it seems my thesis holds water – certainly from a cognitive point of view, and more and more from a clinical-psychology point of view, too. I hope that Jon and I can work on this in more detail sometime.
Please note: my conversations with Jon served merely as litmus tests and should not be construed as formal endorsement of my theories on the part of Dr. Kassel or the University of Illinois.
Of course, it could be that the pattern I thought I had detected was merely a fata morgana, Maybe my community wasn’t seeing the same things I was. So in January 2010, I published a simple survey on SurveyMonkey, which I broadcast to the interactive-design community via social media and list serves. (13) All of my questions could be answered with a simple yes/no. Here they are, along with the results of the 144 people who responded within the first week:
1. Have you ever been at a client meeting where you or your company have presented detailed page mockups for a proposed website (a “comp” complete with graphics and “greeked” text)?
Note: This may or may not represent the culmination of a longer discovery/strategic/IA process, but exactly where this presentation occurs in the overall process is not particularly important in terms of this survey.
2. If you have been to a website design presentation meeting as described above, have you ever heard the client say, “Very pretty, but there’s too much text. We need more/better/prettier graphics.” (this is when clients start talking about including pictures of their pet cat.)
I see this mostly when senior officials have not participated in an earlier discovery/IA/wireframing process.
3. Having been present at the original design presentation, have you later observed (probably through a one-way mirror during a usability session) that respondents say “Don’t make me scroll through the damned eye-candy to get to the substance. Get rid of the picture of that dumb cat!”
4. So in short, do you see any correlation between requests for more eye-candy during the layout approvals, and irritation with the same eye-candy during task-based usability testing?
About 62% of the respondents were from North America, 30% were from Europe, 8% were from the rest of the world.
Even though this is a primitive survey, the statistical results are significant; the pattern I hypothesised is recognized by others by a factor approaching 2 to 1.
Today, “dopamine” seems to have become “flavor of the month”
I first mentioned this research en passant in blogpost I published in January, 2009. (14) I talked about it again briefly at the IA Summit in Phoenix, AZ in April, 2010. Today, the subject seems to be finally taking hold – most recently at the IxDA’s conference, Interactions 11, in Boulder, CO last week (February 2011). Here, Charles Hannon, presented the subject formally (e.g. as the main subject of a talk) for the first time in our community. (15) Although the subject has also been broached tangentially at EuroIA 2010 and elsewhere, I look forward to speaking with Prof. Hannon at some point; alas, I was not able to attend the Boulder conference.
A second empirical observation
When I first suspected that comprehensive design mock-ups might be creating problems, we tweaked the development/presentation process in my own company, FatDUX. Subsequently, we spent much more effort in guiding senior management through our decision-making process prior to showing actual color design mockups. Although we had always involved our clients in the earlier stages of the development process, we had never previously insisted on top-management participation.
My empirical observation is that if C-level administrators are made part of the comprehensive design process, there is less chance they will insist on bigger pictures or cuter kittens on the website. In situations where we have not been able to obtain face-time with senior officials, our designs are more often open to challenge. Only expensive rounds of usability testing have enabled us to reinstate the graphic-design best-practices we normally espouse.
Both of my parents were scientists and the value of the scientific method and controlled studies was something I learned in parallel with my ABCs. As a pre-med student at Washington University in St. Louis, I continued my scientific studies, although I did wind up in a so-called “unrelated field” (encouraged by my father, who helped me send my first e-mail back in 1982 (no typo) to his secretary at the University of Miami). I have since been involved in the creation and/or critique of over 1500 websites and online apps.
So in closing, I encourage you to do your own research to prove or disprove my contention. And if you’d like to share your own empirical observations and/or research, I hope you’ll leave a comment here or write me directly at firstname.lastname@example.org.
Here, I use “thesis” in the literal Greek fashion: as an “intellectual proposition” (θέσις), not a “dissertation” (dissertātiō).
Twitter, plus the SIGIA list maintained by the American Society for Information Science and Technology, and the discussion list of the Interaction Design Association. The survey was published on 10 January 2010.