I have long worried that the Web isunsuited for real learning. The basic problem is users'superficial "surfing" of information: as countless studies have shown since 1997, people tend toscan text on websitesinstead of reading it closely.
Given this, we must acknowledge that users are rushed and don't care much about our websites — they just want to satisfy their current needs and leave as quickly as possible. This leads to classicwriting-for-the-Web guidelinessuch as:
- Craft thefirst 2 wordsof headlines and othermicrocontentfor the scanning eye.
- Usenumerals instead of wordsfor numbers that represent data or information that users might scan for.
- Use bulleted lists :-)
All very fine, but there are cases in which users really do need to learn something rather than get only the highlights. Educational sites obviously fall in this category, whether targeted atteensorcollege students. But many commercial sites also need to educate users.
Let's say, for example, that you're producing a pharmaceutical site about a drug for patients with high blood pressure. One of your goals is to advise them on what to do in addition to taking their pill, but you can't just say, "eat less sodium." You must teach people what this really means; for example, you might offer strategies for ordering at a Mexican restaurant. People with a strong conceptual model are more likely to successfully reduce their blood pressure and to keep it down over the long run.
How can we help users learn more from our website content? A recent research study comes to the rescue here.
Students who completed an elaborate test after reading the textremembered 145% more contentafter a week than students who simply read the text and didn't do anything else.
Interestingly, the people who took the test认为that they had learned 15% lessthan the people who read the text but didn't take the test, even though they'd actually learned much more. This discrepancy between belief and reality adds another twig of evidence to the huge pile of existing studies telling us to observe users' actual behavior andnot go by what they say.
在这种情况下,很明显(事后,计谋rse) why the test-takers thought they'd learned less when they'd actually learned more: taking the test exposed whatever holes they had in their knowledge and thus (correctly) undermined their confidence. In contrast, people who read the text but didn't get tested were smug and predicted that they'd score better than they actually did on the following week's exam.
The test-taking condition was aretrieval practice test, which is much more elaborate than traditional tests. The full procedure was:
- Read the text.
- Recall as much of the information as you can on a free-recall test.
- Read the text again.
- Complete another free-recall test.
It's no big surprise that people remember more after this elaborate procedure, which involves working with the information 4 times. However, the researchers also studied a condition where people simply read the text 4 times. These readers also remembered more a week later than people who had read the text a single time, but only64% more.
In other words, replacing 2 rereads with 2 tests strongly boosted people's week-later performance.
There are two main reasons for this:
- During step 3 — which was the sole reread for the retrieval practice readers — people would presumably pay more attention to any information that they didn't recall in step 2. In contrast, people who simply kept reading the same stuff repeatedly simply developed a false sense of superiority without awareness of their failings.
- The recall tests lodged concepts more firmly in memory. It's long been an established finding in psychology that themoreyou remember something, thebetteryou remember it. The very act of recalling something encodes it stronger in the brain for future use. (Our full-day seminar on人类思维和可用性has more on the user interface design implications of human memory and other psych findings.)
Implications for Web Design
If you're a student who wants to pass an exam, you now know how to study: do free-recall tests after reading each chapter of your textbook.
But what does this research mean for practical design projects?
There's no way we can force Web users to endure a full retrieval practice test and write two essays on everything they remember each time they visit one of our pages.
In most cases, users shouldn't even bother with retrieval practice. Going through all 4 steps of the procedure takes about300% more timethan reading information once. And it results in only145% more informationretained.
The cost-benefit analysis definitely favors the superficial approach: you can read 4 articles and retain a bit from each, or you can study a single article thoroughly in the same amount of time.Information foraging theoryand empirical observations both tell us the same thing: users prefer broader exposure to deeper learning.
Still, in a few cases, users do want to dive deeper. It's profitable to target these motivated users, just as there is an advantage towriting insightful articlesinstead of quick-hit blog postings. When you consideruser experience as branding, it's obviously beneficial if customers remember more from your site than from the superficial sites.
Although true retrieval practice isn't realistic on the Web, you could make a game of it, encouraging users to write down as much as they remember and then click through to achecklistthat lets them count how many concepts they noted correctly. For most sites, however interactivequizzesare probably more feasible: even though it's better for people to write down stuff in a free-recall test, writing feels like work and users prefer simply clicking buttons.
In any case, the point is to get users toexercise their memoryafter reading your content, and then offer them a chance torevisit the materialafter they've seen how little they remember.
Now, go and write down what you remember from this article, and then read it again :-)