Concepts (page 68)

Source-class inflation is this publication's name for the failure in which sources of wildly different evidentiary standing are presented inside one undifferentiated apparatus, so that a reader who…

Specification curve analysis runs every defensible version of the same analysis on the same data and reports the whole distribution of results, so that no single estimate can be chosen after the…

The split-attention effect is the finding that learning suffers when related information, such as a diagram and the text that explains it, is placed apart so that the learner has to combine the two…

The spotlight effect is the tendency to overestimate how much other people notice your actions and appearance. Thomas Gilovich, Victoria Husted Medvec and Kenneth Savitsky named it in a 2000 paper…

A stand-up is the morning ritual of United States Air Force pilot training, described by Hasard Lee in The Art of Clear Thinking. A randomly chosen student stands in the middle of the classroom and…

Standardisation versus adaptation is the choice between keeping a brand identical across markets and adjusting parts of it for each. The Why We Project uses it as a working split, not a finding: a…

Power is the probability that a study would detect an effect if the effect is there. A study with low power that returns a null tells you almost nothing, and a study with low power that returns a…

The arrangement already in place is preferred to the alternatives beyond what its merits justify. It is why a default, once set, is so difficult to shift.

Status signaling is behaviour whose main function is communicating social standing, taste or group membership, rather than producing a direct material benefit. Thorstein Veblen named the consumer…

Steiner's task taxonomy is Ivan Steiner's 1972 classification of group tasks by how individual contributions combine into a group product. It is a framework rather than a finding, and it is the…

Berger's six-part answer to why some things get talked about and others don't: Social Currency, Triggers, Emotion, Public, Practical Value, Stories.

Stereotypical dataset bias is a group being under-represented in one specific class of a training set rather than across all of them. Iris Dominguez-Catena and colleagues engineered both kinds of…