Examining task engagement in sensor-based statistical models of human interruptibility

Abstract

The computer and communication systems that office workers currently use tend to interrupt at inappropriate times or unduly demand attention because they have no way to determine when an interruption is appropriate. Sensor-based statistical models of human interruptibility offer a potential solution to this problem. Prior work to examine such models has primarily reported results related to social engagement, but it seems that task engagement is also important. Using an approach developed in our prior work on sensor-based statistical models of human interruptibility, we examine task engagement by studying programmers working on a realistic programming task. After examining many potential sensors, we implement a system to log low-level input events in a development environment. We then automatically extract features from these low-level event logs and build a statistical model of interruptibility. By correctly identifying situations in which programmers are non-interruptible and minimizing cases where the model incorrectly estimates that a programmer is non-interruptible, we can support a reduction in costly interruptions while still allowing systems to convey notifications in a timely manner.

Interruptibility of Software Developers and its Prediction Using Psycho-Physiological Sensors

“…Fogarty et al also simulated sensors by manually encoding mouse and keyboard actions, such as highlighting a line or editing code. They measured interruptibility in terms of the interruption lag-the time between the notification and the start of an interruption-during coding tasks and achieved 72% accuracy for two state interruptibility classification (interruptible and engaged) [18]. Using a pressure sensor sheet on the desk, Tani et al were able to achieve a similar two state interruptibility classification accuracy on typing and mouse operation with easy and hard phases [43].…”

“…Such an automatic interruptibility measure can then be used to better coordinate interruptions by, for instance, providing visual cues or postponing them to a more suitable moment [37]. Prior work examined the use of context-aware sensors to gather information, such as audio and video streams, keyboard and mouse interaction, or task characteristics (e.g., [17,18,27]). …”

Leveraging Biometric Data to Boost Software Developer Productivity

“…For instance, Hudson et al were able to classify interruptibility into two states with 78% accuracy by manually coding video and audio streams based on features, such as the phone being on the hook or people speaking [90]. Fogarty et al also simulated sensors-manually coding mouse and keyboard interactionsand were able to predict two states of interruptibility based on the interruption lag with 72% accuracy [91]. More recently, researchers started to explore the use of context-aware sensors with no need for manual coding.…”

Adaptive Interfaces and Agents

“…A good deal of research has examined effective cues for the recognition of availability and interruptibility (see, e.g., Fogarty et al, 2005;Ho & Intille, 2005;Iqbal, Adamczyk, Zheng, & Bailey, 2005).…”