UI Design Newsletter – July, 2002

Introduction

When considering the age of users – how old is "old"?

Martina Ziefle from the Department of Psychology at the Technical University in Aachen, Germany compared user performance on two different display technologies: CRT screens and TFT screens. CRT's are used on most desktops, and flat-panel TFT screens are used on most laptops. She divided her 24 participants, who were frequent computer users, into the following age groups:

Her users searched several lines of alphabetic characters looking for specific target characters. She recorded the time to find the targets, their accuracy and whether they preferred the CRT or TFT screens.

Ziefle found that the search times were reliably (22%) shorter when viewing the TFT screens for all age groups, and that 18 of the 24 participants preferred the TFT screen. There was no reliable difference in accuracy. When compared with the other two age groups (young and middle-aged), older users benefited most from using the TFT screens.

Her findings are fascinating, even though her "older" group was relatively young when compared with the findings from many other studies. Herein lies the age problem. When do older users begin to demonstrate age-related deficiencies in performance?

Unfortunately, researchers use a variety of different age categories in their research. To help illustrate the many different definitions of "old" in research studies, consider one recent study by Charness and Dijkstra (1999). These researchers actually conducted three different studies that compared the performance of younger and older adults. In their first study, the older users were defined as those "over 58;" in their second study, the older users were "over age 40;" and in their third study, the older users were "over 50."

Timothy Nichols, Wendy Rogers, Arthur Fisk and Lacy West at the Georgia Institute of Technology attempted to see which adult age classifications were most commonly reported. They reviewed the age classifications reported over the past few years in the Human Factors Journal and the journal, Psychology & Aging.

After combining the information for many studies, the researchers in these journals classified the adults into the following age groupings:

Note that some that were classified as older in some studies were classified as middle-aged in others. And some ages were left our entirely. If we take their classification and fill in the gaps, and add an "old-old" category (now being used more in the industry), we have the following:

I like the above classification, even though I am not sure whether there are any interesting (and measurable) human performance differences occurring between the "young" and "middle-aged" groups. Even so, we definitely need to standardize the age classifications for researchers, particularly for identifying "older" users, so that designers can understand whether or not study findings apply to their older users.

Well it seems pretty clear we will go with the flat panels as prices come down. Faster, preferred, and they take less desk space. It would be very handy if the age research used common classifications. I shall not be holding my breath.