User Research
in the Safety of using in-car Touchscreen
A driving game with simulator was used as our test material to explore the relationship between touchscreen distractions and the number of driving errors, in order to associate how those distractions affect the driver’s safety.
| Project Detail
Project Type: Experimental Research
Timespan: Nov.2017-Dec.2017
Team Size: 5 people
| My Role
User Researcher
Data Analyst
| Method
Literature Review
Experimental Design
Quantitative Analysis
Research Process
Finding
-
Data Analysis
-
Discussion
Experiment
-
Participants Recruitment
-
Data Collection
Study Design
-
Material preparation
-
Pre-test
Literature Review
-
Previous study review
-
Hypothesis defined
Day1
Final Paper
-
Conclusion
-
Writing
Day21
Introduction
We as researchers were curious as to why trends in car designs today are being geared towards adding more touch interfaces for the driver to interact with, when research and data have indicated they are a large contributor to distracted driving. Continuing from our previous study, we decided to focus our experimental research on determining whether our previous claims that the increased amount of capabilities in mobile devices and touch interfaces in vehicles is a large contributor to driving errors and accidents.
Our null hypothesis is: “Distractions caused by touchscreen interactions will not affect the number of driving errors”. The alternative hypothesis is: “Distractions caused by touchscreen interactions will affect the number of driving errors”.
Hypothesis
H0: Distractions caused by touchscreen interactions will not affect the number of driving errors.
H1: Distractions caused by touchscreen interactions will affect the number of driving errors.
Experiment
Material
-
The driving game"Midtown Madness" on a 17" laptop
-
Smartphone
-
Smartphone mount
-
Toy steering wheel and pedals
-
Timer
-
Counter
Participant
Participants were selected based on availability and interest. All were college students between the ages of 19 and 30. 4 women and 8 men, for a total of 12 participants.
Process
A script describing the procedure was read to each participant to minimize extraneous variables and ensure testing consistency. Participants were then guided through a tutorial where we asked them to push hard on the pedals, turn rapidly, and hold steady to familiarize themselves with the mechanics of the game.
After the tutorial was completed, participants first listened to the experiment instructions and were then told to begin driving. The actions of the participants were measured in two 150-second intervals. During one of the two intervals, the participant was instructed to perform an action including checking the message and changing the music on the mounted touchscreen every thirty seconds, each action serving as a touchscreen distraction. Participants were instructed to follow all touchscreen distraction tasks within 10 seconds of the command. The interval with the touchscreen distractions could be either the first or second interval; this was randomized to provide counterbalancing. Participants were instructed to follow all traffic laws during the timed portion. Driving errors, including any collisions or traffic law violations, were counted separately for the interval without touchscreen distractions and the interval with touchscreen distractions. These totals were recorded in a spreadsheet.
Key Finding
On average, participants made more traffic errors when touchscreen distractions were present (M = 1.75, SD = 1.60) than when they were not present (M = 0.83, SD = 1.19). This difference was significant t(11) = 2.73 , p = 0.0197, two-tailed, and represented a medium-sized effect d = 0.66.
Based on the p < .05 from the t-test result, we reject the null hypothesis: Distractions caused by touchscreen interactions will not affect the number of driving errors.
Total number of error the participants made with touchscreen distractions and without touchscreen distractions.
Chart describing Participants results under respective condition
A Pearson correlation coefficient was computed to assess the relationship between the frequency of traffic error with touchscreen and without a touchscreen. There was a positive correlation between the two variables, r = 0.6892, n = 12, p = 0.013165. Overall, there was a medium, positive correlation between using touchscreen or not.
The findings show that, on average, the participants would generally make about one more mistake under the distractions of using the touchscreen versus driving normally.
We used an older model Android cell phone with home screen widgets to simulate a touchscreen, and there were a few participants who were unfamiliar with the operating system used in context to what we were asking them to do. In a somewhat potentially counter-intuitive logical step, this actually seems to somewhat help our study design, as touch screen interfaces across the many vehicles that employ them seem to serve different needs and have sometimes completely different interface design ideologies and abilities.
Conclusion
In this research, our findings showed that participants made more mistakes under the touchscreen distractions than the no distraction driving scenarios, which indicates that the touchscreen distractions would have an impact on the drivers to let them increase their driving errors when they are driving. Our results of the outliers also revealed that participants may need a longer time to get familiar with the driving simulator. Other points that were noted included participant bias between the game and real road environment, lack of familiarity with the simulate touch screen system, noisy testing environment and so on, which suggest opportunities for better case study design for these spaces in the next related research.
Regarding to future works, we would try to use a driving game that has a more authentic road environment, add more types of distractions, or even use a real vehicle to drive the road for the research to gain greater understanding about how the usage of the touchscreens in cars can impact on the driver’s safety.