Definition
Plain language
The most basic way to fit a straight-line relationship through data.
As stated in the literature
Ordinary Least Squares regression, which estimates coefficients by minimizing summed squared residuals; the baseline model whose results opposing analyst personas interpreted as either evidence or a confound to adjust away.
Also called: ordinary least squares, pooled OLS
Why it matters: It is the default first tool for measuring how one thing relates to another, so it's the common baseline whose results different analysts interpret in opposing ways.
For example, OLS draws the single straight line that comes closest to a cloud of points showing income versus years of schooling.
Heard on the show
“The anti-immigration agent writes, quote, "Baseline OLS models show consistently negative effects — immigration, less welfare support.”Episode 196 — AI Agents Reached Opposite Conclusions From the Same Data — and Passed Review