DropDownMenus

Part of an email I wrote (when I was a UsabilityTsar) defending Nielsen to a client.

5) You quote Nielsen :

Another example is the drop-down menu. Users always love the idea: finally a standard user interface widget that they understand and that stays the same on every page. However, while they offer users a sense of power over the design, drop-down menus often have low usability and either confuse users or lead them to unintended parts of the site.

and sceptically comment : I find lots of assertions but little in the way of evidence ... Could we have some evidence that this is true please?

In fact the paragraph you quote contains a link to a previous article (which http://www.useit.com/alertbox/20001112.html) which) refers to an actual experiment : "At the recent Internet World conference in New York, Kara Pernice Coyne and I gave a talk on Web usability methods. As part of our presentation, we ran a small user test for the audience ... The test user typed her full street address in the text entry field, because that's what she'd always done in the past. The drop-down menu then came as a complete surprise and she had to go back to the text field and erase part of her already-typed address information.

''This small study, conducted in front of a crowd of hundreds, shows that sometimes it is enough to run tests with a single user to clearly illustrate a point. Once you see such confusion in action, you realize that using a "helpful" drop-down

menu to save users a few keystrokes can hurt more than it helps. ''"

OK, this reveals the empirical method. It isn't enough to establish a scientific law. But even so, a collection of heuristics, a bag of patterns of the kind of mistakes people make, is useful. Sure you can find people who don't make the same mistakes; and sure there will be mistakes people make that you never pick up by usability testing. But neither of these facts detracts from the point that there is something wrong with the things that you can observe confusing people. (And Nielsen does have experimental evidence that a few users will find most relevant errors : If http://www.useit.com/alertbox/20000319.html If users were all differently confused by different things, then the curve wouldn't have this shape. Instead each new user would add as many new problems as any other. The fact that they don't suggests that users converge on being confused by the same things.)

This article also reveals that Nielsen is not against all drop-down menus. Simply those used inappropriately or confusingly.