RDEL #75: How do interruptions impact different software engineering activities?

Certain tasks, and their complexity, change how impactful an interruption is. Research also discovers a disconnect between perception and physiological data.

Welcome back to Research-Driven Engineering Leadership. Each week, we pose an interesting topic in engineering leadership and apply the latest research in the field to drive to an answer.

Interruptions are part of every engineer's workday, but how do different types of interruptions affect productivity and stress? This week we ask: How do interruptions impact software engineers’ productivity and stress, and what can leaders do to mitigate their effects?

The context

Software engineers juggle numerous complex cognitive tasks like writing code, understanding complex algorithms, and reviewing others’ work. These activities demand deep focus, while interruptions—ranging from a coworker's quick question to a pop-up notification—break that flow. However, not all interruptions are created equal. Factors like the urgency of the interruption or the authority of the person interrupting can affect how disruptive the interruption feels.

Remote work has introduced additional layers of complexity. With more on-screen notifications and fewer in-person interactions, engineers are experiencing interruptions in new ways. Understanding how different types of interruptions impact productivity and stress is critical to fostering better work environments.

The research

Researchers at Duke and Vanderbilt analyzed how interruptions influence three common engineering tasks: code writing, code comprehension, and code review. Twenty participants performed these tasks while experiencing six types of interruptions—both in-person and on-screen. Researchers collected physiological data (via wristbands) and self-reported stress ratings.

A description of each of the six interruptions used in the study

Key findings include:

1. Performance and productivity are impacted by interruptions, in nuanced ways

   1. The importance of the requester increased the impact of interruption.

   2. In-person and on-screen interruptions together significantly extended the time needed for code review.

   3. Simpler code comprehension problems were more impacted by interruptions than complex problems, meaning they took longer.

2. The type of task, and type of interruption, changes the stress associated with an interruption

   1. Interruptions in code writing led to higher stress measures than interruptions in code comprehension and code review.

   2. Interestingly, in-person interruptions had a significant positive impact on physiological measures (using heart rate variability), meaning they were less stressful.

3. Perception and physiological data don’t always align

   1. Even though physiological data showed that stress was lower with in-person interruptions, participants perceived in-person interruptions as substantially more stressful than on-screen interruptions.

   2. Code review showed lower physiological stress scores, but 45% of participants perceived the work as more stressful.

The application

This research highlights the nuanced impact of interruptions. There’s no doubt that they impact the productivity of software engineers, but the impact can differ. Though physiological and perception data often show discrepancies, our recommendation is to rely more frequently on perception data. Perception is based on what an engineer believes is true, which is much more likely to impact their actions and behaviors. When making improvements, addressing perception will be more impactful towards improving the productivity of the team. For more research on interruptions and focus, we recommend a few additional studies - here and here.

Here’s how to apply these findings:

• Minimize high-dominance interruptions. Reduce urgent notifications from authoritative figures during critical focus periods, unless urgent. Use scheduled check-ins instead.

• Find ways to reduce interruptions during code writing tasks. Avoid interruptions during code writing, which has the highest stress correlation. This can be achieved using “Do Not Disturb” blocks within calendar and chat, and a culture that respects that time.

• Measure, so you can improve. Use a combination of survey instruments and telemetry data to understand the impact of interruptions on the engineering team in a more comprehensive way. This can be done manually, or with tools. The goal is to establish a baseline, introduce improvements, and measure their impact over a regular (ie quarterly) cadence how this increased the team’s productivity.

—

Wishing you a week of few distractions and high collaboration. Happy Research Tuesday!

Lizzie