Reading articles in that format drives me crazy. For example: "For example, we now know that 35% of the variation in a team’s performance can be accounted for

Reading articles in that format drives me crazy. For example: "For example, we now know that 35% of the variation in a team’s performance can be accounted for simply by the number of face-to-face exchanges among team members" Can I get a reference to a scientific paper that measures this or where this result is discussed? Off the top of my head - this depends hugely on the task at hand. Certain tasks require constant nonstop communication and updates - other tasks are more modular and the communication sweet spot is different. Even within the same task, different phases of the task require different amounts of communication - when you are whiteboarding the design of a particular library, constant feedback is exceptionally useful. When it's time to sit down and write difficult code involving pointer math, constantly discussing things will just throw you off. I am sure there was some proper science done behind that number - but that number refers to teams of a certain size, working on a specific task, in a given context (I would also be curious to know the C.I. on that 35% too). Trying to draw a nice narrative with a simple message from a narrow experiment is basically everything that I dislike about this Gladwellesque style pop science. Reading narratives like this gives a completely unwarranted sense of confidence into something which we still understand very, very little about.