Proceedings of HF 2002, Nov. 25-27, 2002, Melbourne, Australia

Designing usable applications for older users - an example

Dan Hawthorn

School of Information Systems and Computing,

UNITEC, Auckland.winburne University of Technology

E-mail: dhawthorn@uitec.ac.nz

Keywords:  Aging, interface design.

Abstract

There is now a body of general interface design principles for older users. The paper reports work in progress on a prototype email system aimed at providing proof of concept for these principles. The user interface is described together with indications of the aspects of older users that prompted various design decisions. Based on testing with a small group of older novices the system appears to be highly suitable for older users and it is suggested that this project is on track towards its goal of proof of concept.

1.      Introduction

This paper reports on redesigning a user interface in order  to make an application more accessible to older users. Aging leads to a number of reductions in ability that impact on computer use. Morrell et. al. (2001) provide a key review of relevant research, especially in relation to web design for older users. Hawthorn (2000a) examines the implications for interface design of a wide body of research on aging, it is these papers that provides the theoretical underpinning for the current paper. Earlier work by Hawthorn (2002) has successfully applied the design implications of these finding to interactive tutorial design for older users and Morrel and others have applied the principles to successful Web design. The challenge now being addressed is to see if the same design recommendations can lead to a successful application design. Sharit and Czaja (1999) note that there is little difference in performance between older and younger people using computer applications until the complexity of the application and/or task increases. Sharit and Czaja suggest that this implies that older users in particular would benefit from altering the interface design so as to reformulate the problem posed by the task in a way that simplifies it. The question they pose is how can one carry out such a redesign in ways that meet older users' needs. The slightly different approach used in this paper has been to apply the design recommendations from Hawthorn (2000a, 2002) to redesigning a system that some older users find too complex. If the design recommendations are useful such a re-design should provide the desired functionality in a form that such older users find accessible.

The next section of the paper briefly outlines the study design and results. The bulk of the paper then discusses the solutions to interface design issues for older users that are used in this study. The paper concludes that on the limited evidence accumulated so far in this study, redesigning user interfaces with the needs of older users in mind can markedly increase the usability of applications for this group.

2.      Study Design

The application chosen for redesign was a Windows email system. It was felt that more realistic responses would be obtained from older people involved in usability testing and in making design suggestions if they were dealing with an application area that was important to them. Before design commenced four focus group discussions were carried out with two groups of older users and two groups of people who supported older email users. On the basis on the initial picture of older emailers' needs and difficulties and of the existing interface design recommendations for older users, a series of mockups of potential email systems were constructed and older users responses to these were obtained. A prototype email system nicknamed SeniorMail was then constructed based on the responses to the mockups. A small group of people was recruited to take part in limited usability testing. This group consisted of 6 older users in the 75-80 age range with very limited experience and 2 additional users in the 50 - 60 age range. The users were given five minutes instruction and were asked to carry out a set of tasks from a printed list using a test version of SeniorMail primed with dummy emails.

3.      Results

This paper describes work in progress and further testing with older users is planned. However the design approaches described in the remainder of the paper have been sufficiently successful that a group of inexperienced older users were able to carry out a set of reasonably demanding email tasks using a prototype system. The task list provided to the participants described what was to be achieved but not how to achieve it. The tasks included opening and sending attachments, finding recently deleted mail, forwarding saved mail and using the Address Book. These tasks were well beyond the scope of anything these users had previously attempted.

The group of users chosen provided a reasonably severe challenge to the interface design. They included a person who after a year of using Microsoft Outlook Express (MSOE) did not use the Reply or Forward features - this person sent all emails from their address book. Another two people had left all emailing to their spouses because MSOE was "too complex". Three of the group had only a few hours of computer use. Two had poor vision including one person with the start of macular degeneration and the person with the longest computer experience was one who had been observed to be very slow in acquiring competence with new applications.

All users completed all the tasks with enthusiasm and only minor prompting. They were slow, carrying out 12 tasks took over an hour. This consisted of linear progress towards goals rather than trial, error and frustration. Features in the slow completion included very deliberate reading of screen information and very slow typing. All participants were surprised and pleased by their success rate. The universal question at the end of the session was, "When can I get a copy?" Another telling response was the frequent and heartfelt question, "Why can't other computer programs be this easy to use?"

4.      Discussion of Interface Design Issues

SeniorMail allowed a challenging group of older users to succeed on email tasks beyond what they had previously achieved. The experience of carrying out these tasks was seen positively and the users ended by wanting to own copies of the application. These are preliminary results but they do indicate that the design issues tackled in SeniorMail are worth describing. What will be presented in the following sections are the specific design features developed in this system to respond to typical difficulties that older users find with learning, vision, manipulation, navigation and complexity of the system model.

5.      Simplicity

One of the central issues in the system design was to reduce complexity. The starting point for doing this was to remove features that were not needed by older users. Focus groups with older people were used to examine the use patterns of older email users and identify the features that were most important for an email system tailored to their needs. The key points that emerged were that these older users had fairly simple emailing needs and sent relatively few emails per week to a very small group of people. However the ability to send and to receive emails was important to them.

One could design an email system stripped right down to a [Get new email] button, a viewer and [Reply] and [Send to…] buttons. Such a system would use predefined addresses and would not store email once read. However such a system would be too simple for many older users and would be unlikely to be accepted by the people who advise older users on their software choices. Further from the standpoint of research on designing software for older users it would avoid rather than address most interface issues. Therefore the challenge that has been taken up is to design an email system that achieves easy use by older users with restricted needs and at the same time can be used as the regular email system for users who make more extensive use of email.

6.      Learning

Older users are typically slow learners of computer applications, reasonable effectiveness with a low level of mistakes can take weeks or months to achieve, Bosman and Charness (1996). This is a demanding period in which older users can experience considerable levels of frustration and may abandon attempts to use the application. A central principle of SeniorMail was to design a system that required almost no learning to use effectively. Typical systems ask users to find and remember the purpose, name and location of features buried in menus, to experiment and to remember a plethora of small details and to transfer knowledge gained from other Windows packages. As younger or as experienced users we take a large number of simple skills for granted and perform them at an automatic level. We manipulate and remember menus, tab to new fields, Alt-Tab to different applications, resize and manipulate windows, drag and drop, select, copy and paste and use a host of other skills with little conscious thought. It is sobering to watch older users who lack such skills and who find difficulty in acquiring them. Competent computer use (and typical interface design) depends on a large number of automated user skills, each of which is individually simple but needs initial learning. Few computer actions are really intuitive for absolute beginners. If such learning (and automation) does not occur, users are simply unable to act in ways that designers take for granted. Particularly for older users, actions that we would assume to be standard and straightforward screen manipulation become an exercise in frustration and in trying (and failing) to remember and co-ordinate a multitude of sub-tasks.

Figure 1. Main menu of the current version of SeniorMail. Note how options are grouped vertically to facilitate one dimensional searching.

What was done in SeniorMail to reduce learning was to make it easy for the user to search for the feature that they wanted if they failed to remember where it was located. The search was restricted to a linear scan over a few standard buttons. In the main menu shown in Figure 1 the buttons were grouped and the search was within a vertical group. All the other SeniorMail windows followed the pattern shown in Figures 2.and 3. where the buttons that gave access to features were always located on a toolbar at the top of each window. The basic learning required was, "Look along the toolbar at the top for a useful button". Novice users were found to be able to perform tasks successfully with almost no errors, although they were slow.

It was observed that the older users had begun to remember the location of the most commonly used buttons after completing only two or three tasks. The expectation is that the older users will, over time, remember the button locations reliably, increase their speed and build an effective mental model of the system. Crucially, the design let users to perform relatively complex tasks without previous experience.

An important point here is that this design protects older users from the frequent experience of forgetting parts of previously learnt skills. Termed "fragile knowledge" this phenomenon is common among older learners and a significant barrier to competence, but with this design there is relatively little penalty for forgetting exactly how to carry out a particular task, one simply resorts to searching the top toolbar again.

7.      Visual design

The ideas behind the screen designs used in SeniorMail came from previous work by Morrell et. al. (2001) and Hawthorn (2000) on designing for older users. Usability testing of prototypes was then done with a group of older users to fine tune the designs.

SeniorMail uses a full screen design rather than using overlapping Windows. This increases visual simplicity and allows the use of larger fonts and larger components. Further reasons for using full screens are covered in the sections on manipulation and on navigation below.

A number of older users suffer from a degree of impaired vision. SeniorMail used several simple approaches to counter this. The font used is Arial, this had been found to be easily readable for older users in earlier work, Hawthorn (2000b). The basic font size for message headers and text was reasonably large (12 point) and could be further increased up to 16 point. A larger than normal font was also used on buttons and hint windows. Standard Windows error message dialogs were replaced with message boxes using a large Arial font on a light background.

Figure 2 The screen layout used for lists of email headers.

With macular degeneration in particular, sufferers struggle to distinguish text against colored or patterned backgrounds. It was observed that this included reading the captions of standard Windows buttons (black text on a mid gray background) even with an increased font size. The solution was to use a non-standard button background color, a very light gray ($00DFDFDF) with a slightly darker window or toolbar background ($00DBDBDB). The result was much improved readability for older users with poor eyesight. In some ways it would have been desirable to move away from the standard Windows button color scheme altogether but it was considered that doing so would break the users' expectations of what a button looked like. Large blocks of text as in the Viewer, Editor and Help screens were displayed on a very pale yellow ($00EAFFFE) background to reduce glare, another area that older users find difficulty with. A couple of users with deteriorating vision including one with the onset of macular degeneration noted that the system was exceptionally easy to read.

Older people are less able to filter out extraneous information so that the presence of information in the display that is irrelevant to the immediate task increases their cognitive load. This was part of the reasoning behind the decision to use a full screen rather than a tiled design. Again in order to reduce irrelevant detail no decoration was used apart from icons on the buttons. Older users were asked if they desired icons on buttons and the balance of opinion was that they did. The icons were not necessarily interpreted accurately but they served to make it easier to distinguish between buttons when searching. Icons also directed attention to buttons that were off the main toolbar. For example when updating details in the Address Book the user needs to click [ OK ] or [ Cancel ] buttons, these were missed by some users until the buttons were provided with icons. "Flat" toolbar buttons should, in theory, be desirable (they provide an additional cue for successful target acquisition and offer a reduction in extraneous detail by removing the borders of irrelevant buttons) however when given a choice older users rejected flat buttons, they wanted buttons that are recognizably buttons at all times.

Older users are poorer at visual searches unless there are preparatory positioning cues and / or restrictions on the search area. At the same time older users have a reduced effective visual field and may miss tool or navigation features that are widely separated from the area of the screen they are working with. Watching older people working with MSOE indicated that they had problems in scanning a two dimensional space for appropriate choices.  Older people are also slower and less effective in scanning lists, older participants in the SeniorMail project have been observed missing MSOE menu items they were searching for, apparently because of the sheer volume of choices. Therefore as far as possible all SeniorMail features were accessed from a few, consistent, large toolbar buttons at the top of each screen. The number of buttons was restricted to eight at the most even where this meant restricting the available features. For example [Reply] and [Forward] buttons would have been convenient on the toolbars for the email header lists but, because extra buttons would overcrowd the toolbars for these screens, these options are only available from the Viewer. There is a deliberate tradeoff of achieving visual simplicity even if access to features then requires more steps. It is worth noting that where the pattern of relying purely on toolbar buttons was broken, as in the tabbed pages used in the Viewer and Editor, novice users did not see or understand the tabs until these were pointed out to them. A later version of SeniorMail using icons on the tabs so that they are more akin to the toolbar buttons appears to have increased the ease with which older users find the tabs.

Figure 3. The Viewer with the HTML view showing and a plain text view available. Note the oversized cursors added to the screen dump, the default cursor is to the right of the word "Viewer" at the top, the text cursor is to the right of the word "Cheers" at the bottom.

Another example of the problems of older users with visual search is the observation that some users had difficulty in locating the standard cursors against a background of email header lists or message text. Hence SeniorMail uses a very large default arrow cursor with a double width black outline and a new, large, text cursor that is particularly easy to find, see Figure 3. These cursors were enthusiastically supported by the older users in the study.

The Viewer is intended to display emails that include HTML or enriched text formatting instructions. However, since the format chosen by the sender may be poorly designed from an older reader's perspective, SeniorMail provides a plain text view as the default with a web browser view as an alternative if required.

Older users are likely to find it difficult to resize overlapped or tiled windows by dragging the edges, window edges are too small a target for many older users to acquire. Older users are also likely to unknowingly click outside the boundaries of foreground windows and be mystified when the foreground window vanishes. These arguments contribute to the decision to use full screen windows in SeniorMail. In addition all pop-up windows were displayed modally so that clicks on the background had no effect.

Scrolling is another area of difficulty where attempts to scroll through a long document can frustrate older users who typically are unsure of the differing consequences of clicking different parts of the scrollbar and may have problems dragging the slider along the track. The use of full screen windows for email header lists and viewing emails in SeniorMail reduces the need to scroll but does not eliminate it. One could provide buttons that allow the user to control scrolling without using the scrollbar but a useful alternative is already available in the use of the Page Up and  Page Down keys and Up and Down arrow keys. The issue is how to make older users aware of these keys. A possible approach that has not yet been implemented would be to monitor mouse movement and provide suggestions about using the keyboard alternative if trouble was evident with mouse use while scrolling.

8.      Navigation 

Older users have been shown to perform more poorly in navigating complex web sites, searching or working with complex programs, (Czaja and Lee 2001, Sharit and Czaja 1999). Unfortunately moderate levels of complexity are usual in the design of popular email systems. Typical email systems provide a flexible navigation model where the user can jump to nearly any folder or tool from any starting point usually in one or two steps. There is also a mixed pattern of tiled and pop-up overlapping windows, intended to present as much of the system as possible. This is combined with menus, toolbars and a forbidding level of jargon.

For older users there are drawbacks to having multiple visible windows. The overall screen display is more complex because it presents more navigation choices and this requires the user to have a reasonable mental model of the system and to remember more about the location of features that allow access to parts of the system. Older users often have significant difficulty with both remembering navigation features and with searching for them in a visually complex space.

Complex screens also imply a need to reduce the font used in order to give space to all the competing features. Another implication is the reduction in target size as well as the proliferation of targets. There is the implied expectation that users will be comfortable with relatively frequent navigational jumps and windows management acts such as resizing, maximizing or minimizing. For older users with poor vision, reduced manipulative accuracy and poor short term recall of immediately preceding actions, this is likely to result in misplaced clicks that can mean unexplained jumps to unknown parts of the system with no self evident way of getting back. Again the argument supports the use of full screen window displays.

The main SeniorMail task windows are accessed from a simple main menu screen, see Figure 1. Options on the main menu screen are grouped so that visual searching is through a linear space within one of three main group. A choice on a main menu button typically leads to a window displaying a list of email headers. These list displays are standardized with some large buttons on a toolbar at the top and a list displayed below, see Figures 2 and 3. To open an email the user clicks a line in the list of email headers and then clicks the [ Open ] button.

Navigation is usually linear. Users move outwards and then return step by step to the main menu using a consistently located [ Back ] or [ Menu ] button. Typical patterns are:

Viewing:           menu -> list -> viewer -> list -> menu

Replying:           menu -> list -> viewer -> editor -> list -> menu

Write email:      menu -> editor [-> popup address list] -> menu

Elderly novices were observed to succeed very rapidly with this constrained but simple model.

Another problem encountered by older users was the terminology used in many email systems. This does not lead to the sort of simple understanding of what features are available that leads to effective navigation. For example in MSOE the button labeled [New Mail] does not indicate if this is inwards or outwards mail, a window captioned "Composer" does not clearly express that this is the window is to be used for writing a new email. The design approach used was to discuss button and list captions with older users and select those that gave greatest clarity in indicating the purpose of the button or list.

It is assumed that the typical SeniorMail user will have some support from a  more competent Windows user and so some features such as the details for connecting to ones internet provider, setting up a list of commonly needed folders to search for attachments or setting up email categories are available from the Options window rather than directly from the main menu.

9.      Complexity

There is a general issue when designing a simplified system of how far one goes to accommodate features that are expected in more fully fledged systems. In the area of design for older  users this opens the issue of the extent to which software for older people is distinct from software for the general public or how far one can go to improve the usability of general purpose software by designing in features that suit older users. In a recent paper, Hawthorn (2002), suggests that older users will in fact require specialized software, at least for the current cohort with limited computer experience, but probably for longer. There is insufficient space to explore this issue here but a brief indication can be given of some specific ways in which SeniorMail attempts to make some more complex features available to those users who want them. One of the key difficulties faced by older users is dealing with a hierarchical file system. Observations from previous studies by this author suggest that many older users do not achieve a good understanding of files and directories and continue to have difficulties using them. With email this becomes an issue when attachments are to be saved or, more particularly, sent. The approach used is to try and sidestep the complexities of the file system, a new folder called My Attachments is automatically created under My Documents and is the default area to which incoming attachments can be saved if desired. For sending attachments the approach is for the senior user's support person to nominate a few folders that are most likely to contain files that the user will want to attach and SeniorMail provides a simplified step by step guide to finding files in these folders and attaching them.

The next area of significant complexity lies in organizing one's emails. The standard solution is to use folders and allow the user to create extra folders. There are problems here for older users who may find creating folders daunting, who may forget to use them, forget that an email has been stored in a particular folder or who may find the drag and drop manipulation that goes with folders is too hard to carry out. Yet at the same time some older users and a number of their supporters and family may see folders as a pre-requisite of a respectable email system. SeniorMail in fact uses a number of fixed folders; the In and Out boxes, Saved mail, Recent Copies and Recently Deleted. These folders are referred to as lists and displayed as such in SeniorMail. Emails are transferred between these lists by selecting email header lines in the list and clicking [Delete] or [Save] buttons, thus avoiding drag and drop. The user cannot add to these folders thus preserving a simple and understandable structure. However there is a way of adding further structure for those who desire it. An option is available to let the user classify the email headers into categories when transferring them from the inbox to the saved emails list. A window for maintaining a list of categories is available from the Options screen. If the user is chooses to use categories the behavior of SeniorMail alters in two places. The display resulting from the Sort button is extended to allow filtering of the saved emails by category. In addition the Save button now displays a dialog box where the user chooses a category to save the email under. The point here is that the system only adopts this more complex behavior on demand, users who do not need to further classify their mail do not have to consider the features that work with classification thus preserving the design simplicity.

10.  Conclusions

This is a report of work in progress. Looking at small numbers of older users there has been a very positive response to the email system. All the older participants were able to perform to levels that they found acceptable in their first session with the system.  For the participants who had experience with MSOE the redesigned email system was seen as a major advance in usability. Preliminary observations indicate that speed will increase considerably as users move from searching for the correct toolbar button to remembering its location. This level of success is all the more interesting given that the users had minimal background in computing or limited success with MSOE. The success comes from adopting a range of interface design principles for older users developed in detail in previous papers by this author and others. These design principles are in turn derived from consideration of the considerable body of work on the changes in perceptual, cognitive and physical ability that accompany aging. The SeniorMail project appears to be on the way to offering proof of concept for the value of the following ideas when designing for older users: simplifying learning and demands on memory, simplifying the visual layout and improving text readability, reducing choices and features, simplifying manipulation and constraining navigation. At the same time some features such as the minimal learning and penalty free forgetting afforded by the toolbar approach are new and particularly suited to older users. However the experience of the author, in working on this project, is that in themselves the design principles only provide a useful starting point. What translates the initial work into successful design for older users is the participation of older people in the development of the design. Work is now proceeding on carrying out further cycles of usability testing and refinement. One aim is to produce a system that is robust enough to install in older users machines so that reports on long term use can be obtained. A second aim is provide wider ranging investigation of the usability issues in the design so that explicit recommendations to designers can be offered with some confidence.

11.  References

·         Bosman E. A. and Charness N. (1996) "Age-related differences in skilled performance and skill aquisition" in Blanchard-Fields & Hess T. (eds) "Perspectives on cognition in adulthood and aging" McGraw-Hill, New York.

·         Czaja, S. J., & Lee, C. C.  (2001).  The Internet and older adults: Design challenges and opportunities. In N. Charness, D. C.  Park, and B. A. Sabel, (Eds.), Communication, Technology, and Aging:  Opportunitites and Challenges for the Future.  (pp. 60 - 78).   New York:  Springer Publishing Company.

·         Hawthorn D. (2000a) Possible implications of aging for interface designers, Interacting with Computers, 12, 507-528

·         Hawthorn D. (2000b) Designing for older users: Display quality versus task complexity In Proceedings, Australasian Computer Human Interface Conference. Sydney.

·         Hawthorn D. (in preparation) Design factors in a successful interactive tutorial for older users, a usability study

·         Morrell, R. W., Holt, B J., Dailey, S. R., Feldman, C., & Mayhorn, C. B.  (2001). Older Adults and Information Technology: A Compendium of Scientific Research and Web Site Accessibility Guidelines.  Washington, D. C.: National Institute on Aging.

·         Sharit, J, & Czaja, S. J. (1999).  Performance of a computer-based troubleshooting task in the banking industry:  Examining the effects of age, task experience, and cognitive abilities.  International Journal of Cognitive Ergonomics, 3, 1 - 22.

·         Smith, M. W., Sharit, J., & Czaja, S. J. (1999) Aging, motor control, and the performance of computer mouse tasks.  Human Factors, 41, 389 - 396.

·         Sit, R. A.  (1998).  Online library catalog search performance by older adult users.  Library & Information Science Research, 20, 115 - 131.