Introduction

Web sites are designed to provide the user with a resource of easily accessible and relevant information as well as an appealing design both in graphics and structure of the site. If both criteria are met, the chances of the user returning to the site escalates. Since design issues are often subjective, it is difficult to please all users. However, empirical studies conducted for the site in progress on prospective users may help determine which features appeal to the majority of the target audience. The main focus in the study of the E-Maryland site is on the menu design and the overall navigational efficiency of the pages.

1.1 Overview

The menu is the doorway to the website. The home page, which contains the initial entry into the deep structure of the site, should have an intuitive menu design that allow users to quickly perceive which information they would like to further inquire about. Unfortunately, correct information retrieval remains a difficult task on the Internet due to poorly designed interfaces. Studies have shown that users are unable to find the specific information they are searching for 42% of the time [10].

In order for users to make good judgments about which links they would like to pursue, the site should contain links with precise, well-defined categories. The more categories available, the more precise the users can hone in on their target information. At the same time, each page should be easily searchable and not appear cluttered. The original E-Maryland page has a dynamic menu design where the site responds to users mouse movements over the icons in the center. While this feature may appeal to some users, it would take a few moments for the users to make the connection between the menu, the icons, and their mouse actions. These few moments would render the site less efficient than a simple, static menu where everything is laid out for the users. The layout of the dynamic menus may also be counter-intuitive. Once the mouse pointer is over an icon the menu may pop to the right, left, top, or bottom of the main category label. This inconsistency may initially confuse the user.

Another design issue that may need improvement is the hierarchical structure of the site. Because of the ambiguities of the categories, the narrow, deep structure of the site may hinder search efficiency, since users have to speculate which categories their desired link may be located. To optimize navigational efficiency, a mix of good menu design and structure combined can create a more user-friendly environment for the users. Many cases have been studied in an attempt to determine the most efficient type of web site structure. The depth versus breadth debate in hierarchical menu design becomes an issue, and an appropriate balance must be determined to create a more efficient web interface. While a broad, shallow structure would easier to navigate, the search time may increase per page. Therefore, a precarious balance exists between a page with too much information and one that appears too cluttered, giving the user a migraine.

Examples

Several sites with vast amounts of information, such as search engines and news sites have provided relatively successful sites for users to easily navigate through. Internet users have praised the organization of sites such as www.yahoo.com, www.msn.com, www.amazon.com, etc. These sites provide a vast number of links, yet they are well-structured and distinctively categorized to provide ease in navigability. These and other generally more effective sites have static layouts for their index page to allow users to quickly scan for their target link.

1.2 Review of Previous Experiments

Static vs. Dynamic

There have been a vast number of studies on menu and web site navigability in the field of human-computer interface. Various factors have been manipulated in an attempt to determine the features of a site or menu that affect user performance in finding their target information. One issue is the design of the site. Static interfaces have been shown to result in faster search times for expert users of intra-nets. In a study by E. Nygren, results indicated that the fastest search times were obtained when the search space can be captured by the peripheral visual system [4]. Therefore a static interface, would create a more efficient search environment for the expert user. Research has also indicated, however, that expert and novice users use different search heuristics and therefore, may have differing results.

Breadth vs. Depth

The breadth vs. depth issue in web design has also been widely studied. Results from several studies have indicated that when there are many hyperlinks in a site, a medium condition of depth and breadth is the optimal condition for user performance. A site with excessive depth usually harmed performance, while a shallow, broad design is also unable to achieve the performance rate of the optimal condition. According to a study by Kiger and later confirmed by Zaphiris and Mtei, they found that in a site of 64 links, the 8x2, where there are 8 links per page and 2 levels, resulted in the fastest response times and lowest error rates [3].

Categorization

Precise assignment of links to categories is another significant factor in site navigability. In a study by Snowberry, Parkinson, and Sisson, they showed that accurate grouping of categories played a significant role in a broad shallow design. When all 64 links were on one page, coherent grouping of links resulted in better performance in speed and error rates [2].

1.3 Review of Related Literature

The theoretical basis of menu design lie within the human cognitive processes involved in menu selection processes. Users must undergo information acquisition and search, choice process and time, response time, and error detection [1]. A good menu design should minimize the time for each of the four processes while providing the precise information the user would like to find.

• Information acquisition involves knowing the exact target to search for and thus reduces the process to simple matching. If the target is unknown, then the user may have to speculate which links would provide the information that they would need, increasing error rate and search time. Another factor involved in acquiring the information is the distinctiveness between the categories, both visually and semantically. If areas are ambiguous and terms can be easily confused, the user is more likely to stray from the correct path.
• Choice process and time can occur after the users have inspected all their choices on a page or in conjunction with searching if they can target a specific category from all the available categories. Uncertainty between choices may delay them from making a decision or increase their frustration levels during navigation.
• The response process is the physical process that users must act on to make their selections. For websites, this usually involves clicking on links with their mouse in order to retrieve their information. Factors that affect the response process would be font size, distinctiveness, and color of the link.
• After the users have made their selection, they would evaluate their choice by viewing the results to determine if the correct information was retrieved. Evaluation is usually an immediate a process.

Depth vs. Breadth Trade-off

The trade-off between breadth vs. depth is directly related to short-term memory. Larson refers to G. Miller’s 7 +/- 2 items, which can be stored in short-term memory at any given moment in time. Therefore the number of categories existing on a level should not exceed the normal processing capabilities of short-term memory. Studies have shown that access time is proportional to depth in menu selection [3].

Another factor involved in the effectiveness of web sites is described as “scent”. The Information Foraging Theory defines scent as the measure of remote indication that the user can glean based on the sites design structure and labeling [2]. By grouping hyperlinks into their precise categories, users are able to detect the correct path to their desired information. The assignment of categories is especially important in sites with deeper structures. Under these conditions, ambiguous sites may cause the user to be lost in the site. According to Chi, information scent is the “imperfect, subjective, perception of the value, cost, or access path of information sources obtained from proximal cues such as web links, or icons representing the content sources” [12].

Effective view navigation in sites, even in sites with vast amounts of information, should have a small viewing space, navigation should take only a few steps, and an easily discoverable route to the target information [5]. Furnas defines Efficiently View Traversible (EVT) as both small views and short paths. Therefore, both long scrolling lists or a deep structured site are inefficient for navigation. Again, Furnas confirms the optimal conditions for navigation are sites which have medium breadth and depth. Long scrolling lists also increase the chance of the users overlooking a specific target since most users perform a quick scan for 2 or 3 seconds rather than carefully examining their choices [6].

Scrolling vs. paging is another factor in the choice between breadth or depth. A broader version would require more scrolling, while the deeper versions require more paging. Results have shown that reading times do not differ, but users tend to build strong mental representations in both memory and comprehension when paging through vast amounts of information [11]. Furthermore, it has been shown that it is more effective for older users to page rather than scroll.