Chapter 5: Exploiting vulnerabilities in comprehension

Literacy, numeracy and problem solving

In 2013, a huge worldwide study called the Program for the International Assessment of Adult Competencies (PIAAC) was published, involving over 165,000 working-age adults in 23 different countries.¹ It looked at literacy, numeracy and problem-solving proficiencies across the world. The following summary is just for the United States, though the picture is fairly similar in many countries. According to the 2013 PIAAC findings:²

• 30% of adults in the US are likely to have difficulty sorting through emails and organising them in folders provided for them.
• 20% of adults in the US are unlikely to find the name of a congressperson with a summary information sheet that lists the district, name, year and place of birth.
• 30% of adults in the US are unlikely to be able to calculate the total cost of daily car rental when provided with miles driven that day, cost per day and cost per mile.
• 16% of adults in the US are digitally illiterate, and cannot use a computer to find a recipe, make a retail purchase or file taxes online.

As you can see, low literacy and numeracy is very common. With an exploitative mindset, this presents an exploitable vulnerability. If a business wants to hide unfair or unappealing aspects of a transaction, it can do so through the use of complex language or complex numerical content. With this in mind, it’s interesting to consider the writing style used on public service websites – plain language, short sentences, and enormous efforts taken for comprehension for all citizens – versus the writing style used in more exploitative products like crypto trading apps, where impenetrable technical terms are used extensively, very little is explained, and the user is enabled to make all kinds of risky trades with minimal safeguards or education.

How scan reading can be used to manipulate people

When we read, we don’t usually read every word on every page. Not unless we’re studying really hard or working our way through something we’re enjoying, like a novel for example. Take a look at this:³

On the left, you can see we tend to let visual hierarchy determine the order in which we read things. We’ve learned it’s a good idea to read the big, prominent things first and smaller things later. On the right, you can see how we glance at content and make educated guesses about individual words to save time. This isn’t something we’re born with. It’s a technique called ‘scan reading’, which we pick up naturally as we get better at reading. Similarly, good writers and page designers learn how to design for scan reading, to help people to do it more efficiently.

Steve Krug published Don’t Make Me Think in 2000. It’s now in its third edition, with over 350,000 copies in print. This book is highly regarded in the UX design industry, as it puts forward a clear explanation for the concept of scan reading by people who are using screens.

Let me show you two more images. The first image explains what we naively might assume is a natural way to read information. In theory, we’d expect readers to take in each successive word, thoroughly paying attention to every element of the design. This naive view of human information-seeking behaviour is similar to the concept of homo economicus from traditional economics – the idea that humans have a limitless supply of attention, energy and critical thinking skills, so we can brute-force our way through any body of content by reading every word on a page in sequence.

However, Krug argues that while authors might hope that people read every word on every page in a highly attentive and rational manner, the reality is rather different, as shown in the second image (above). In real life, most of us take a ‘billboard going by at 60 miles an hour’ approach⁴ when there’s this much information presented to us.⁵

Krug argues that users tend to ‘glance at each new page, scan some of the text, and click on the first link that catches their interest or vaguely resembles the thing they’re looking for. There are usually large parts of the page that they don’t even look at.’ He goes on to explain that we’ve been trained to scan-read from an early age, flicking through newspapers and magazines, for example, or reference books, as we try to narrow down many choices and find just the parts we’re interested in.

Other researchers found more evidence. Back in 1997, Morkes and Nielsen did a quantitative empirical study in which 51 participants tested five variations of a website, each one with a different style of writing:⁶

1. A promotional writing style – full of ‘marketese’
2. A scannable writing style – intended to encourage scan reading
3. A concise writing style – succinct content
4. An objective writing style – not using promotional language
5. A combined concise, scannable and objective writing style

Each person was given a series of tasks, generally involving looking for the answer to a simple question. The amount of time they took was recorded, as were any errors they made. The findings showed that people performed worse on the promotional style pages, but significantly better with the scannable and concise styles.

This research demonstrated what we might have divined naturally: a writing style has an impact on users’ ability to read and understand information. If users read every word on every page in a systematic way, these differences wouldn’t have been seen. In a subsequent article, Nielsen (1997) addressed the question, ‘How do users read on the web?’ with a two-word answer: ‘They don’t.’ He went on: ‘People rarely read web pages word by word; instead, they scan the page, picking out individual words and sentences.’⁷

Understanding how people read is vital if you want to design web pages or app screens that work effectively, or – conversely – if you want to create deceptive patterns.

Eye-tracking research is another useful source of insights about reading behaviour. In 2014, Pernice, Whitenton and Nielsen ran an eye tracking study with over 300 participants.⁸ In one exercise, people were asked to use a search engine and find some specific information. Eye-tracking technology followed their progress, monitoring how they fixated on the page: 17% of the time, people looked at only one result before clicking onto the next page. They didn’t fix their gaze anywhere else. Or, in other words, they picked the first result that seemed ‘good enough’ to save effort, rather than systematically reading every result on the page. This is a demonstration of an information foraging strategy, a technique that was first defined in 1999 by Pirolli and Card, who noticed similarities between animal food foraging strategies and the way in which humans search for information online.⁹ When an animal forages for food, it cannot search everywhere or it may die from starvation, so it must use a ‘good enough’ strategy that provides the most benefit for the lowest cost. Broadly speaking, information foraging can be considered a kind of goal-directed scan-reading strategy.

Generally, scan reading and information foraging can be a pretty effective way of saving ourselves time and energy. But it is only effective in a predictable, trusted environment in which the designer has your best interests in mind. If a designer wants to deceive you, they can take advantage of scan-reading behaviour by hiding pertinent information where you don’t expect it, or by using misleading headings and visual hierarchy, among other things.

Misleading information and false beliefs

If a business publishes misleading information, this can lead users to make decisions that are not in their best interests. This can range from outright lies (fraudulent claims) to ambiguous or manipulative language and design that encourages the user towards a false belief. For example, a business might exploit scan reading by making the headings, links and buttons on a page appear to say one thing, while the body text, if read word-for-word, says another. Similarly, offers can be priced in a manner that requires considerable mental arithmetic and short-term memory to compare properly. If the user is not capable of this task, they might end up with a bad deal that harms them financially. The FTC lists ‘false beliefs’ as one of the top harms posed to consumers by deceptive patterns¹⁰. In a 2021 study involving 3,777 participants, researchers Luguri and Strahilevitz found that ‘hidden information’ doubled the acceptance rates for a product offer, as compared to a neutral design. In other words, participants formed false beliefs about an offer because facts were hidden away from view, and this had a substantial effect on their decisions.¹¹