Searching for information needed to complete a desired action is a common everyday experience. Examples-such as finding the telephone number of a local department store, searching for a book title on Amazon, or locating the required troubleshooting steps to fix a computer glitchare both familiar and numerous. The relevant information is generally available in a directory, in a manual, or on the Internet and needs to be searched for, often by category.Consider an individual using a relatively unfamiliar word-processing program at his or her workplace. If a situation arises that requires the user to perform a novel action, the first step that the individual might take is to search through the online help manual for that action, which requires searching through menus of categories and subcategories for the required information, as well as verifying that a given item solves the problem at hand. Similar situations arise when searching through Web sites or even through an unfamiliar grocery store for certain products. A difficult search is likely to be inefficient and might result in failure with consequent stress and frustration on the part of the user.Previous research has revealed that users' ability to locate information in a menu-based hierarchically organized system is variable and, in some instances, is highly inefficient. Whereas some studies have found that users can optimally locate (i.e., select the correct categories and subcategories without having to return to previous menus) target information as often as 95% of the time (e.g., Allen, 1983;Miller, 1981), others have found much lower rates. Frankhuizen and Vrins (1980) reported optimal location of targets only 56% of the time, with participants accessing approximately twice the minimum number of menus required, and in a study by Whalen and Latremouille (1981), users completely failed to locate the target information 28% of the time. In general, better performance is associated with smaller set sizes and more distinct categories (Dumais & Landauer, 1984), transfer to new category exemplars is better with a larger number of trained exemplars (e.g., Homa & Vosburgh, 1976), and performance improves with increased practice on the task or with expertise in the search domain (e.g., Salmeron, Canas, & Fajardo, 2005).A complication that might impact performance, particularly on a menu-based category search task, is that category sets and category contents may be dynamic. That is, the set of categories used to partition the search space, as well as the items that constitute the search space, may change over time. For example, with regard to software programs, entirely new categories of functions may be introduced over time, or functions might be added to or removed from the program under an existing category. Even when the content of the categories and subcategories is static, users may be required to search for novel target information or through categories previously available but never before containing target information. Interesting questions arise from the c...