UX Design for the Public Good

How can UX design bring out the best in us, rather than the rest in us?

My work examines website features that impact the way we think about controversial topics. These topics evoke strong reactions from ourselves, our social networks, and our wider societies.

UX design must take into account certain predictable psychological reactions to website features. Factors I’ve published research on are highlighted below (See #Overview for a chart presenting different relationships). Other important factors will be included, too. Updates will be announced on the Posts and Updates blog page.

For now, this page covers #exemplification; #self-expression (voting, reactions, commenting); #citation


Did you know that we tend to judge the whole by our recent experience with its parts? Like with the proverbial blind men and the elephant, unfortunately, this tendency can lead us astray.

Illustrates exemplificaiotn using parable of blind men and the elephant

Recent and memorable experiences are judged to be more common, even though these experiences are typically memorable because they are uncommon. This is not a conscious process. It’s hard to estimate base-rates – hard to estimate population-level information. We rely on numbers when we can, but we also rely on our own experiences. Note, this reliance on personal experience may have served us well when the only populations we had to be concerned about were local and the only experiences we had were face to face – unmediated by any technology, temporal distance, or physical distance.

The modern media environment, however, allows us to access a wide array of views, at any time, day or night. This in turn allows our media selection decisions to influence our perceptions of the wider world.

For example, if you read a news article taking a conservative {liberal} stance on a political issue, you may, probably unconsciously, revise your estimates of the number of Americans sharing the article’s stance accordingly (Sude, Knobloch-Westerwick, Robinson, & Westerwick, 2019; Westerwick, Sude, Robinson, Knobloch-Westerwick, 2020). Voluntarily read a liberal article? The longer that you spend reading it, the more Americans you may estimate take a liberal stance on the article’s topic. Swaying peceptions of public opinion is one function of the fake social media profiles created by trolls.

Even the number of arguments you see in favor of a stance on a controversial issue can impact your estimates of a view’s popularity (Zerback, Koch, & Krämer, 2015)!

However, as Zeback et al. (2015) found, if you yourself are convinced by the arguments, you may further engage in projection: a tendency to assume that other people share your views.

In online spaces designed for the public good, it is important to overcome this tendency towards exemplification by giving context. As my own research, and that of others, has found, perceptions of public opinion can have social consequences. For example, the more that people perceive that others agree with them, the more they may feel licensed to hold extreme views; the less they feel that other people agree with them, the more moderate they become (Sude et al., 2019; Westerwick et al., 2020; Neubaum & Krämer, 2016).

Further, while understanding that other people disagree with them can encourage your user to be more moderate, feeling that they are in a minority can discourage your user from speaking up, out of fear of social sanctions. This means that your user may never express their view, stopping it from influencing others’s perceptions of public opinion. This can lead to a spiraling process in which one side speaks up boldly because they see people who share their views speaking up and another hides their true opinion because they falsely believe they are in the minority.

So, provide context!

How common are people taking that article’s/blog post’s/meme’s stance, really?
How common are people taking as extreme a stance as that message?
How common are people making the arguments in that message?
How common are people citing that message’s sources?

For each of the questions above, try to provide population-specific information. Common on your platform? Common in terms of polling data?

Present this information and while you won’t overcome exemplification and projection processes entirely, you can at least prevent them from being the only processes that impact a user’s perceptions of the big wide world.

Note, however, that you may be happy if some views remain quiet. For example, alerting neo-nazis that your platform harbors sympathizers may not be in line with either your brand or the public good! Perhaps, however, with attention to how the content of your site shapes users’ perceptions of public opinion, you can identify “trouble spots” among your users, and intervene appropriately.

Social identity cues

Section in process.

Self-expresssion (Reactions, Votes, etc.)

Interactivity that allows self-expression is a fundamental feature of the online information environment.

Sude, Pearson, and Knobloch-Westerwick (in press, Computers in Human Behavior) found that the mere act of up- versus down-voting an article impacted users’ political attitudes. Self-expression through voting strengthened users’ expressed attitudes, even if they did not read actual article content! Similar results have been found for sharing articles (see Johnson and colleague’s work).

Why? While mechanism was not tested directly in my study, drawing on Pingree’s Bidirectional Message Effects model and the Theory of Interactive Media Effects (TIME), my colleagues and I suspected that the act of self-expression would shift users’ focus from the content of the article. Instead of thinking about what others had to say, they would focus on their own thoughts and feelings.

This increased focus on the self, combined with self-expression behavior, would give users a sense of having clearly defined political attitudes. Rather than dwelling on nuance, or taking a balanced stance, they would remember their final decision to like or dislike, to up- or down-vote.

The act of voting on the article (the behavior) would in turn increase users’ commitment to those attitudes. Both clarity and commitment would translate to reporting a more extreme attitude.

Our predictions were supported.

Notably, people often voted without reading content at all! Further, the mere ability to vote made content less persuasive. When users could express themselves, they were not focused on what the article’s authors had to say.

Strength of initial attitudes mattered. Users who had moderate attitudes on the political topic to begin with were more likely to read attitude-challenging content, up-vote it, and become even more moderate. Extremists, on the other hand, voted in line with their initial attitudes and became even more extreme.

The overall pattern of findings suggest that self-expression affordances should be carefully managed on social media sites. Popular online community host Lace on Race, for example, discourages any use of reaction buttons on her Facebook page, requesting that people leave thoughtful comments instead. Her goal is to increase the quality of intellectual, emotional, and social engagement around a contentious issue.

While encouraging deliberation is important, both academic research and personal experience suggest that the more contentious an issue is, the more deliberative quality may suffer. Simply put, when people are passionate, they have trouble overcoming their own internal biases, let alone treating people who disagree with them respectfully.

In these case, instead of encouraging commenting, UX designers could offer a self-expression tool. I recommend offering the option of publicly answering a survey question. The user could experience the reward that comes from self-exression but also be prompted towards greater thoughtfulness. These questions should appeal to head and heart, as well as prompting more careful attention to the content.

Questions could include:

“You could learn a lot from this [content] just by skimming through it, quickly”

“This [content] taught me new things.”

“This [content] made me more certain about my own views”

“This [content] helped me to understand other people think”

“This [content] helped me to appreciate how other people feel”

“This [content] was important”

“This [content] made me want to share information with others”

“This [content] made me want to become more politically active”

Appropriately designed, these survey-question style self-expression tools could help to create a more “normatively desireable” user. Of course, these should be legitimate survey questions, not ones designed to increase bias frequently employed in political fundraising.


As a teacher, one of the hardest things to convince undergraduates of is the value of citation!

A generation that grew up immersed in social media has a tendency to equate being convincing to others with being correct. Citation – which is necessary to make broadly, objectively or at least inter-subjectively, defensible arguments – is often ignored.

Further, even adults on social media ignore citation! How many times have you seen a political meme shared presenting so-called facts with no source? How many times have you wondered whether those facts were accurately drawn from a source? How many times have you made a claim in face to face conversation that you later realized was at least partially wrong (e.g. mis-remembered or lacking context)?

Social media platforms may want to feel even more frictionless and rewarding than face to face conversation! Having to defend your claims can be a real drag 🙂 . This makes the work of Russian troll armies so much easier!

Can UX designers make citation a frictionless experience? Can a “cite this” tool be embedded in every platform, utilizing search algorithms and (relatively) vetted sources (Wikipedia; ResearchGate; google scholar, Crossref)? Can reward be built in, where people get badges for the quality and quantity of the external sources they are citing?

Ideally, the search engine associated with the “cite this” tool would be able to establish a hierarchy of badges, for example from gold to bronze.

As a teacher I encourage students to look for more recent articles. Users could earn a “gold badge” by citing more recent material.

As a teacher I encourage students to look to meta-analyses or review articles – which often overcome the limitations of any particular study. These too could help people earn the “gold.”

As a teacher I encourage students to go beyond opinion pieces and try to uncover the original research that the authors of these pieces cite. People who make this effort, again, could be rewarded with a higher-level badge.

If the badge system was itself inter-subjectively defensible, and flexible enough to be accepted by the community of users, it could motivate better habits and, at the very least, disrupt the current “niche” that malicious parties occupy in the information ecosystem.


Intervention Path 1:

#exemplification: by making sure that content provides a nuanced perspective on public opinion, we can shape both the salience of different social groups and prceptions of public opinion within these groups. This intervention can have downstream effects. Beyond mere distribution of opinion, we can shape perceptions of the people within these groups. Are the likeable? Are they moral? Will they cooperate? This in turn can shape our users’s communication in the future, which can shape the open-minded and closed-mindedness with which they approach new information. Think of the spread of disinformation on social media. If it favors their side, people who want to cooperate with their in-group may accept it or defend it (Sude, Garrett, & Riva, 2020). If it attacks their side, they may instead reject it (reflecting a competitive motivation).

Intervention Path 3:

#self-expression: By shaping the ways that users can express themselves, we can shape perceptions of themselves and others. Is the emphasis on expressing their own views? Justifying their views? How about evaluating the views of others? By shaping the way users can express themselves we can promote more cooperative interactions directly by encouraging more thoughtful responses in the moment and indirectly by fostering high quality conversations (which may then be observed by others).

#citation As with self-expression, citation helps foster higher quality interactions.