Parker Williams and Jeff Jenkins, Information Systems
Hand tremors are experienced by millions of individuals on a daily basis, the main cause being neurological disorders and neurodegenerative diseases. Essential Tremor, “a progressive neurologic condition that causes involuntary rhythmic trembling of the hands,” affects 10 million Americans of all ages1. It is estimated that 87% of individuals ages 18-29, 86% of individuals ages 30-49, 76% of individuals ages 50-64, and 70% of individuals ages 65 and older go online daily2. With an increasing number of individuals who have access to and use of the internet, creating a real-time adaptable webpage environment for those with hand tremors is necessary.
This research project had two objectives: (1) to detect users who have hand tremors through analyzing their mouse movements on webpages and (2) to determine how to adapt webpages in real-time to increase accessibility when hand tremors are detected. To date we have not yet completed the project but do have preliminary results to examine.
To date we have designed and carried out one experiment with two different sub groups. The experiment consisted of two Qualtrics surveys and a website that participants navigated.
The designed website resembled that of an e-commerce site selling sandals and socks for children, women, and men. Figure 1 shows an example of what this website looked like.
Figure 1: (A) Display of the website’s product page, (B) Display of the Kauai, Women’s Sock, product
We initially had trouble finding individuals who experienced hand tremors. This was an unforeseen obstacle that we had to navigate. After a few months we made contact with the International Essential Tremor Foundation based out of Overland Park, KS. With their help we were able to distribute our study to 1,700 individuals. This yielded roughly 200 respondents who had hand tremors.
In order to have a control group and increase the size of our respondent pool, we administered the study on Amazon Mechanical Turk. We again received roughly 200 respondents. In total we had 466 respondents begin the study with 268 completing it. These participants produced five million distinct data points for us to analyze.
We specified a general linear model that regressed whether or not someone had tremors (a binary factor) on the twelve specific mousing statistics. As displayed in Table 1, we found that tremors statistically influenced five different variables. We also found that tremors do not statistically influence the remaining seven variables as seen in Table 2.
Table 1: Variables Statistically Influenced by Tremors
Table 2: Variables Not Statistically Influenced by Tremors
Based on the preliminary results detailed above we successfully determined that we can detect hand tremors by analyzing mouse movements on webpages. This was one of two objectives we set out to accomplish with this research project. As we continue this research we will address the second objective of determining how to adapt webpages in real-time to increase accessibility when hand tremors are detected.
In conclusion, we have determined that we can in fact detect hand tremors by analyzing mouse movements. This discovery will help us in creating a real time solution to adapt web pages to individuals who experience hand tremors. We look forward to continuing this research and learning more about hand tremors through mouse movements and creating adaptive systems due to this discovery.
- Stephens, S. (2011). Essential Facts about Essential Tremor: This “quiet” disease, which affects 10 million Americans, is anything but benign. Neurology Now, 7(1), 21-23, 27. Retrieved from American Academy of Neurology.
- Zickuhr, K., & Madden, M. (2012). Older adults and internet use. Washington, D.C.: Pew Research Center’s Internet & American Life Project.