Study: Making Automated Technology Easier for Older Adults to Use

How do you make technology easier to use for older adults who struggle with diminishing cognitive ability? One way is through high set-level compatibility.

Set-level compatibility refers to the relationship between stimulus and response set and the relationship’s influence on response time. The more commonality shared between a stimulus and response set, the faster the response time (i.e., high set-level compatibility). The less commonality shared, the slower the response time (i.e., low set-level compatibility).

An example of high set-level compatibility is an IVR asking a question audibly and a user responding verbally, as shown in the picture above. The response is in the same modality as the stimulus. An example of low set-level compatibility is a user having to respond to an audible question via manual text input.

My thesis explored how set-level compatibility affects task performance of older adults while using an automated voice system.

Cognitive Decline

Previous research has shown that older adults may experience greater difficulty using automated systems because of cognitive limitations. Older adults often experience cognitive decline in areas such as memory, focus, spatial ability and speed of information processing, which are required for successful technology usage.

Memory is important because users are required to process, store and recall information given by the IVR in a short period of time. Focus enables a user to filter out distractions in the background and hear prompts for navigating the system.

Spatial ability enables a user to make a mental model of the IVR system, which helps navigation. Speed of processing enables a user to process and respond to prompts in the limited time available while using an automated system.

Thesis

In my study I created banking and healthcare IVR applications using Plum Fuse to investigate the effects of high (auditory prompt – vocal response) and low (auditory prompt – manual response) set-level compatibility on younger (18 – 39 years old) and older (65 years and older) adults.

As expected, both age groups performed better in the high set-level compatibility conditions. Of the older subjects, 78% rated the IVR systems in this study as easy to use.

Some Separate Tips for IVR Design

Here are a few suggestions for designing IVR applications to accommodate the needs of older adults (and younger adults, for that matter):

• Use a broad menu structure for your application, as opposed to a deep one. Previous research has shown that broad-based menus are better for people with lower working memory capacity, which is common in older adults.
• Provide good error recovery. Even with the best possible IVR design, it is inevitable some users will make errors.  Therefore, it is essential to incorporate good error recovery with your IVR design.

  Good error recovery involves three steps:

  1. Informing the user when an error is made and the type of error
     e.g., “you entered an invalid account number”
  2. Allowing for recovery
     e.g., “please re-enter your account number”
  3. Repeating until the error has been corrected
• Use an excellent synthetic voice or voice actor. A common complaint about IVR systems is difficulty understanding the voice prompts.Synthetic voices have continued to improve over time and now sound more realistic than they did years ago. (The Fuse platform used to create the IVR had over 50 voices from different genders and regions to choose from.)Although more expensive than a synthetic voice, a voice actor is nice to use because their voice can capture the specific tone of a business (e.g., bubbly and cheerful versus serious and reserved).Whatever the voice type chosen, all prompts should be clear and concise.
• Design prompts that are short and to the point. Another common complaint about IVR systems is that the prompts are too long.It is imperative that IVR prompts are concise and to the point. Longer prompts can make it difficult for subjects to navigate the system successfully because users have to hold many items in their memory.The prompts used in the study were short with approximately the same number of words.