Regression (statistics)
Etymology
From Latin “regressio” (a going back), from regredi (to step back), re- plus gradi (to step). Francis Galton coined the statistical term in the 1880s from his finding of “regression toward mediocrity” in heights, and it later broadened to any fitted relationship.
noun · /rih-GRESH-uhn/
A method for estimating how one variable changes as others change, drawing a best-fit relationship through data.
What it names
It is the workhorse behind most claims that “X affects Y, controlling for Z” in economics and social science.
Example
A regression might estimate that each extra year of schooling is associated with an 8 percent rise in wages, holding experience constant.
Related idea
Statistical regression is distinct from regression to the mean, the phenomenon Galton named it after; and “controlling for” a variable in a regression narrows but never fully closes the gap to genuine causation.
noun · /rih-GRESH-uhn/
A return to an earlier, less advanced, or less desirable state.
What it names
a backward movement in development or condition
Example
Under stress he showed a regression to childish behaviour.
Related idea
Distinct from the statistical method — here regression is simply going backward, the surfacing of something latent rather than progress forward.