Evidence for The Broken Window Theory in UX?

Question

In criminology, Broken Window Theory postulates that seemingly small defects (graffiti, broken windows) in the environment around us can lead to increase in more serious crime by the means of priming, norm-setting and signaling.

From my personal experience, there exists a similar effect between seemingly unrelated design and usability defects and user's ability to perform given task.

Example

When I go to example.com to book my flight, I notice their browser tab icon is pixelated on my retina MacBook (not updated to higher device pixel ratio even though it is 2017). Then I am needlessly prompted to save any unsaved changes in my profile page even thought I did not change any fields.

From there on I am suspicious of everything and it takes me more time to book my flight. When asked to rate my experience on the site I give it lower mark even though there was no real issue with the booking process itself.

The above example is highly anecdotal. Do others have similar experience? Am I correct to compare it the criminological theory, or is there more suitable UX concept for this? Is any research or data on this topic?

Answer 1

As Jim Ryan notes in his answer, by a strict analogy with the broken windows hypothesis, you'd be saying that if designers don't get away with small design crimes, then they won't commit bigger design crimes. For a number of reasons, I would dispute that, but I don't think that is what the question is really suggesting anyway.

The question's actual hypothesis seems more like: "small, low-stakes design flaws can undermine the overall experience, even when all the major design issues are well attended to".

It'd be hard test such a broad hypothesis quantitatively, but I think it's reasonable to assume it is somewhat true. There's no question that user expectations are a real part of the system, and a pattern of small, visible defects will affect that.

Suppose you're checking into a hotel (which happens to have the world's most reliable wake-up call service, but you don't know that), and you ask the receptionist to wake you at 6am. If everything's in order, you will forget about it, and wake up on time. But if the receptionist looks untidy, and the plant on the desk is clearly dead, you might not be so confident. You'll worry about it, and maybe call from your room to confirm the alarm. Even though the actual wake-up call works fine in both scenarios, that dead plant on the front desk has in effect cost you time and effort.

Answer 2

Broken Windows Theory is Malarkey!

(sorry, wanted to use that word :) )

Now, seriously speaking, that theory is quite controversial to say the least, and most modern researchers and scholars won't accept it at all.

Thus, it's very difficult to associate such a controversial and purely subjective theory to UX, a corpus of disciplines that aims to derive knowledge from objective data.

However, what you mention in your example has a name, and it's just cognitive load: the process itself is causing you to go slower and second guessing what you're doing because of different cues and loaded elements that shouldn't be there, or should act in another way.

In some way or another, your example refers to what is known as cognitive biases. Oddly enough, it could fit into several of the most known biases, even though the description is not very long. I think there's a mix of Attentional Bias affecting your decision making process, but also some elements of Anchoring Effect (interacting with your attentional bias). In some way, you could also say your model could be explained by Fuzzy Trace Theory, but in the end, the easy, simple and direct name for this is just cognitive load

Answer 3

Those sorts of UI cues often indicate, or seem to indicate, a site with bigger issues. However it's hard to draw parallels between crime-ridden urban neighborhoods and digital products, except in cases like that of the Dark Web, where the product is itself a crime-ridden neighborhood.

According to the Broken Windows Theory, shabby surroundings encourage bad behavior by signalling a breakdown in social norms. In the example you gave, the site's sub-optimal UI decreased your trust. That's more like going into a restaurant where the floors aren't swept or the tables cleared and assuming that the hygiene in the back kitchen won't be too great either. The only signalling that happens there is a message to the potential patron telling him/her to eat elsewhere to avoid food poisoning.

Both types of signals are valid and telling, but they address different audiences with different messages.

Answer 4

I guess you can make up your mind about the validity of the theory, even though in this case I think there isn't a definitive answer because there are possibly other factors that need to be considered.

However, assuming this is true then one might postulate that paying attention to and fixing up small usability issues should lead to design practices that addresses the deeper user experience or design problems. Well, I am sure we can all think of cases where a company offers a product or service that looks very shiny and is sprinkled with delightful micro-interactions, but doesn't actually solve the user problems particularly well (e.g. the cracked mobile phone screen theory?).

I think it is easy to dismiss a theory as being misconstrued, but often there are some elements of truth to it, even if it is because we have only considered some and not all of the factors that contribute to the overall effect. In addition, none of us are immune from cognitive biases and sometimes things can appear to us in a certain way because we are looking for it.

I would encourage you to explore the theory further and test out the idea so you can find the underlying cause for your particular example.

Answer 5

We could speculate that small, seemingly unrelated UX defects negatively affect the task completion rate and time-on-task via the means of cognitive load, user expectations, and cognitive biases.

But at the time of the bounty expiration, we are not aware of any research or data to support or reject the analogy of Broken Window Theory in UX.