Is it possible to make something too easy? Is there a point where the easier something is to do, the fewer people do it correctly? For many of us, this seems counterintuitive. For those of us working in product development, whose job it is to create simple and intuitive products, it seems even crazier. However there is a line of thinking, and even research to support the idea, that making something too easy can be detrimental to a successful outcome.
I am sure you can think of plenty of examples in your life where this is exactly the case. The homework that was so simple you didn’t bother to check your work and as a result made stupid mistakes, or how you can cook a 3 course dinner perfectly, but when you are boiling pasta you somehow always forget about it on the stove.
In this article, I will talk about the phenomenon of how making something easy does not always equal better. Specifically I will discuss what this means for the creation of Instructions For Use (IFU) and other informational text, and how product designers and UI/UX practitioners can use this to their advantage when creating content.
Isn’t UI/UX and human factors all about making things easier?
In product development, we spend a lot of time and effort ensuring a product is intuitive and easy to use. We tend to use the same principles of device development while creating informational text. We create an organization and layout that are intuitive and help the user find the information they need. We also strive to maximize ease of use and correct understanding through a variety of formatting guidelines.
The National Library of Medicine, part of the National Institute for Health (NIH), in their article titled: How to Write Easy-to-Read Health Materials, recommends using sans serif fonts such as Arial, Verdana, Frutiger, Tahoma, or Helvetica since they are easier to read. They also suggest not italicizing long strings of words and warn against placing text on top of shaded backgrounds, photos, or patterns. In their 2001 guidance on Medical Device Patient Labeling, the Food and Drug Administration (FDA) Center for Devices and Radiological Health recommends using a minimum of 12 point serif font for the text and suggest only using black type on a white background as it is easier to read. The FDA Center for Drug Evaluation and Research (CDER) published a presentation titled Patient Labeling 101 and recommends sans serif fonts at a minimum 11 point size for better visibility.
While there are slight differences between all of these recommendations, they all have one thing in common. As summed up by the guidance on patient labeling, the recommendations all boil down in some form to improving “the speed by which letters and words can be recognized”. In essence, making the content easier to read.
Why speed and ease are not always our best friend
There is an idea in human factors — one that is not often discussed but held as a guiding principle — of differentiating the goal from the means of achieving the goal. We are agnostic to the methods we utilize or the solutions we create as long as we achieve the correct result while creating an enjoyable user experience.
This is true in many fields as well. Someone in advertising shouldn’t care if they create a print ad for a magazine, a TV commercial, an Instagram post, or a Facebook video, as long as they receive the right amount of attention of the intended audience at the correct price point. Similarly, a chef should not care if they have to roast, sear, sous vide, or pan-fry a piece of meat as long as they achieve the desired texture, doneness, and taste. The list of examples can go on and on.
This may seem like common sense, but as soon as you start discussing cognition and human behavior it can get very confusing. Cognition is much more nuanced than people assume and acts in ways which may not seem logical.
Likewise, we need to separate the goal from the means of achieving the goal in terms of IFU or instructional material development. The goal is to write content that people understand and are able to follow effectively. The logical, but potentially incorrect, means currently used are what the recommendations above claim, which are about improving the ease at which people can read the content. However following the recommendations of the NIH, CDER, and the Center for Devices and Radiological Health may actually create content that is harder to understand or follow effectively. Like I said, separating the means from the goal is not always as sensical as people initially assume.
What is actually going on then?
Like I have stated in the past, HFEs designs products (read: IFUs and other instructional material) for the user in the manner which they would REALISTICALLY use them. We should also design content for how people will REALISTICALLY interpret them, and herein lies the issue.
According to a widely held psychological theory, there are two methods which the brain uses to process information. System 1 is effortless, quick thinking, and intuitive, but often prone to biases and illogical thinking; this is often the default. System 2 is slow, logical, and deliberate, but requires high mental effort and is only used when ‘necessary’. Often understanding human nature and how people came to a certain conclusions are dependent on understanding which system is being activated and used to process information.
In a series of experiments by Alter, Oppenheimer, Epley, and Eyre, they test whether experiencing difficult or disfluent information can cause a shift from System1 to System 2 processing. In the first experiment, they assess if changing the font style from black Myriad Web 12 point font (fluent condition) to 10% gray italicized Myriad Web 10 point font (disfluent condition) can increase performance on a cognitive test.
The second experiment looked at the effect of disfluency on people’s reliance on biases. Using a fictitious product review, they tested if disfluent information before the content, in the form of headline/title, can alter how people understand the content.
The third experiment tested if the unconscious thought of increased effort, through making a face of concentration (disfluent condition), was enough to alter how likely people were to commit logical fallacies. In their last experiment, using the same fonts described in experiment 1, the researchers tested participants syllogistic reasoning, a test used to assess higher-order reasoning.
In all four experiments, the researchers found the disfluent condition led to increased performance, both on the cognitive and syllogistic reasoning tests and also decreased likelihood of committing biases and logical fallacies. The researchers concluded that disfluent information, or even just the thought of the need for increased mental effort to complete a task, is enough to initiate a switch to System 2 processing.
H∅Ψ d∅ Ψe μse the$e finding$?
I think there are a lot of opportunities to use these finding in instructional documents, such as IFUs and Quick Reference Guides. However, I want to offer a caveat first. These are the findings of one research article and that does not mean they are universally true. Furthermore, because we would be utilizing the findings in an entirely different situation, we may see completely different results. As such, I think we would need to do some studies to see if this is even something that can be utilized for instructional material.
For longer documents, there are a couple of different options. The same options as above are available, but there is another option that I think could work quite well. Considering System 2 requires effort, it may not be realistic to expect people to utilize high cognitive effort for the entire time it takes to read the document. You potentially run the risk of people getting cognitively tired or distracted resulting in performance worse than if you didn’t activate System 2 at all. But, If you select key steps or sections you can selectively chose where you want System 2 processing to be utilized.
For example, you could highlight (through disfluent presentation of text) safety critical steps as determined through a risk based approach such as a Failure Mode and Effects Analysis (FMEA). Alternatively you could chose steps based on which steps have had the most use errors in a previous study, indicating that users need to pay more attention to those steps. Another option would to highlight an entire section that is most critical. For consumer medical devices, like an inhaler or syringe, this may be the steps associated with administering the medication as opposed to the sections on setup, storage, and cleaning.
One thing that has to be mentioned is that disfluent information is less familiar by definition. As such, people will probably not like it as much (thank you mere-exposure effect) and, I assume, it would decrease their perceived use experience.
However, like anything in life there is a balance at play here. One one hand we want to maximize the user experience and enjoyment of reading the informational content so people are more likely to actually do it. On the other hand, we want to ensure users are comprehending and following the instructional steps properly to facilitate correct use of the device.
Why is this necessary in the first place?
Now, one topic that I have not addressed is why this is necessary in the first place. Do people not understand instructions? Who even reads the instructions and why don’t you just write the information clearly and concisely?
These are all good questions. On one hand, people often do not look at instructions and no matter what changes you make people will continue to ignore them. However, there are people that do use the instructions. For most people, medical devices are a foreign object that they do not have much, if any, experience with. Medical devices work in ways we are not used to, require steps that we do not perform often, and for many are used temporarily or sporadically.
While writing the information in a clear and concise manner is always helpful, it is not always sufficient. When someone does look at the instructions or other information, we want to put them in the best possible position to succeed. This information is rarely interesting or exciting, so we need to use every tool in our toolbox to ensure users are able to comprehend and understand what they are reading, and if that means including disfluent information than so be it. Remember, differentiating the goal from the means is paramount, and while it may seem illogical that doesn’t mean we shouldn’t embrace it.