I first came across the concept of "attentional" gambling while I was researching an article on change blindness, a phenomenon where we fail to notice even fairly obvious changes between images if they are separated briefly by a blank field. (That is, in precisely the situation where we would most like the opposite to be truesuch as when reloading a Web page with feedback or changes.) 
Attentional gambling aptly describes the investment we make each time we decide where and how long to focus our attention on a particular part of a page. In the case of change blindness, the odds are controlled almost entirely by the subtlety of the changes from one screen to the next. The use of color and simple animation (flashing text, for example) significantly improves our chances of noticing the new content.
I think attentional gambling also serves as a useful model for a variety of interaction design issues. Take the navigation approach shown in Figures 1 and 2, for example.
The speculation here is on where I should look after I click one of the left-hand menu items. In most Web navigation, the safe bet would be on the content area, labeled `A’ in Figure 2. But in this case, there is no sure win since both the content and the navigation change simultaneously (see `B’ in Figure 2).
In this particular example, the actual content of the second screen (the text on the right) is fairly simple. Users could conclude that it was not really what they were looking for after only a cursory inspection. If they happen to notice that the navigation changed, they might then examine `B’ further. However, in usability testing of a similar design but with more complex content, most users did not notice the change in the navigation at all, effectively placing all their confidence in `A’ and losing.
There are two ways we could make this particular design a win-win situation: either expand the menu without changing the content (that is, the navigation from Figure 2 with the content of Figure 1) or ensure that the new links shown are replicated in the content. (In this case one of the links does appear in the content, but not the other.)
Consistency plays a big part in the attentional gambling model. In the betting scenario illustrated in Figure 2, inconsistent navigation means that users do not know where to look on the page because, for many of the links, only the content area changes. But before users can bet on the outcome of clicking a link, they must first find one. If, when first exploring a site, users find that a heading is not a link, you can be fairly certain they will treat similar text the same way. Likewise, if users find what they were looking for in a particular part of a page, they will bet on finding it in the same place on other pages.
So far we have considered how users gamble on where to look for information on a page, but there is also the question of how long they should give their attention to a particular part of a page. Frequently, if their confidence wanes, they will change horses. (Attentional gambling has far fewer rules than betting at the track.)
Here are a couple of tips for success:
- Categorized lists Make sure items are grouped according to users’ expectations. The size of the attentional gamble will be proportional to users’ conviction in finding what they are looking for. And that will be based on the heading or the relevance of first few items.
- Content Structure content so that it can be easily skimmed and use the "inverted pyramid" style of writing with a broad description of the content in the first paragraph. Headings, indentation, and lists all help improve the odds.
Attentional gambling might also provide an interesting way of presenting the results of usability testing. The number of "bets" made plus how many losses and wins would provide useful insights. Good designs will not only maximize users’ wins but also minimize the number of bets needed. Producing gambling profiles for users would also tell us something about their interaction styles. Some users will be cautious, placing few attentional bets but hoping for high reward, while others may be risk-takers. Knowing how these different punters (er, users!) behave could tell us a lot about how to improve our designs.
©2004 ACM 1072-5220/04/1100 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2004 ACM, Inc.