Don’t Bother Me. I’m Socializing!

IT DOES BOTHER US when we see our friends checking their smartphones while having a conversation with us. Although people want to focus on a conversation, it is hard to ignore a series of notification alarms coming from their smartphones. It is reported that smartphone users receive an average of tens to hundreds of push notifications a day [1,2]. Despite its usefulness in immediate delivery of information, an untimely smartphone notification is considered a source of distraction and annoyance during social interactions.

deferred_notifications
(Left) Notifications interrupt an ongoing social interaction. (Right) Notifications are deferred to a breakpoint, in-between two activities, so that people are less interrupted by notifications.

 

TO ADDRESS THIS PROBLEM, we have proposed a novel notification management scheme, in which the smartphone defers notifications until an opportune moment during social interactions. A breakpoint [3] is a term originated from psychology that describes a unit of time in between two adjacent actions. The intuition is that there exist breakpoints in which notifications do not, if so minimally, interrupt a social interaction.

video_survey_screenshot
A screenshot of the video survey. Participants are asked to respond whether this moment is appropriate to receive a notification.

TO DISCOVER SUCH BREAKPOINTS, we devised a video survey in which participants watch a typical social interaction scenario and respond whether prompted moments in the video are appropriate moments to receive smartphone notifications. People responded that the following four types of breakpoints are appropriate breakpoints in a social interaction; (1) a long silence, (2) a user leaving the table, (3) others using smartphones, and (4) a user left alone.

SCAN_social_context
Types of social context detected by SCAN.

BASED ON THE INSIGHTS FROM THE VIDEO SURVEY, we designed and implemented a Social Context-Aware smartphone Notification system, SCAN, that defers smartphone notifications until a breakpoint. SCAN is a mobile application that detects social context using only built-in sensors. It also works collaboratively with the rest of the group members’ smartphones to sense collocated members, conversation, and others’ smartphone use. SCAN then classifies a breakpoint based on the social context and decides whether to deliver or defer notifications.

SCAN HAS BEEN EVALUATED on ten groups of friends in a controlled setting. SCAN detects four target breakpoint types with high accuracy (precision= 92.0%, recall= 82.5%). Most participants appreciated the value of deferred notifications and found the selected breakpoints appropriate. Overall, we demonstrated that breakpoint-based smartphone notification management is a promising approach to reducing interruptions during social interactions.

WE ARE CURRENTLY EXTENDING SCAN to apply it to various types of social interactions. We also aim to add personalized notification management and to address technical challenges such as system robustness and energy efficiency. Our ultimate goal is to release SCAN as an Android application in Google Play Store and help users to be less distracted by smartphone notifications during social interactions.

You can check out our CSCW 2017 paper to read about this work in more detail.  

“Don’t Bother Me. I’m Socializing!: A Breakpoint-Based Smartphone Notification System”. Proceedings of CSCW 2017. Chunjong Park, Junsung Lim, Juho Kim, Sung-Ju Lee, and Dongman Lee (KAIST)


[1]”An In-situ Study of Mobile Phone Notifications”. Proceedings of MobileHCI 2014. Martin Pielot, Karen Church, and Rodrigo de Oliveira.
[2] “Hooked on Smartphones: An Exploratory Study on Smartphone Overuse Among College Students”. Proceedings of CHI 2014. Uichin Lee, Joonwon Lee, Minsam Ko, Changhun Lee, Yuhwan Kim, Subin Yang, Koji Yatani, Gahgene Gweon, Kyong-Mee Chung, and Junehwa Song.
[3] “The perceptual organization of ongoing behavior”. Journal of Experimental Social Psychology 12, 5 (1976), 436–450. Darren Newtson and Gretchen Engquist.

About the author

Chunjong Park

Chunjong Park received his B.S. and M.S. degree in computer science from the Korea Advanced Institute of Science and Technology (KAIST) in 2015 and 2017, respectively. During his Master's studies, he worked with Prof. Sung-Ju. Lee, Prof. Juho Kim, Prof. Dongman Lee, and Prof. Junehwa Song. He will join the University of Washington in 2017 fall as a Ph.D. student. His research interests primarily lie in mobile computing and human-computer interaction.

View all posts