From Attention Span by Gloria Mark loc 1637:
In a first study, we logged computer activity of thirty-two people over five days to get a precise measure. We found that our study participants averaged checking their inboxes seventy-four times a day, replicated a year later in another study where we logged the computer activity of forty people for twelve workdays and found that they checked email an average of seventy-seven times a day. One dedicated person checked their email 374 times a day. We could also get a fairly good estimate of how often email checks were due to external or internal interruptions by whether people used email notifications. Most participants—41 percent—checked email without notifications, i.e., they self-interrupted to check it, and 31 percent of our participants primarily checked email from notifications, which would be external interruptions. The remainder—28 percent—did have notifications turned on but reported that they checked email about equally due to external or self-interruptions.
It isn’t quite what Gloria Mark is saying here, but I’m tempted to frame this as bad faith. We impute a capacity to control our work to external factors, whereas in a reality we are self-interrupting albeit in a way that loosely anticipates and responds to those external interruptions. The way in which we craft ourselves in relation to external constraints, above and beyond what is strictly speaking required by them*, makes the ensuing problem much worse than it would otherwise be. What is the problem?
As interrupted tasks pile up, you carry that tension around with you, and it becomes more and more of a drain on your resources. You certainly want to limit the stress by the end of the workday as there are carryover effects of bringing stress from the office into your personal life.[30] Those unfinished tasks churn around in your thoughts, and what you can do to reduce that churning is to externalize your memory of that unfinished task.
Attention Span by Gloria Mark, loc 1751
*See also academic over-publishing while blaming the incentives of research assessment which encourage a smaller volume, as well as sometimes explicitly calling for quality over quantity.

