MENTAL MODEL · M7838

Quantile Regression

BusinessHigh supportEconomics

Included

Updated 2026-08-10

Loading revision record…

INTRODUCTION

English translation pending.

CORE DEFINITION

Proposed by Koenker and Bassett in 1978, quantile regression estimates the conditional quantiles of a response variable by minimizing weighted absolute residuals, rather than minimizing squared residuals as ordinary least squares does for the conditional mean. This exposes heterogeneity, since a predictor's effect can differ across the distribution, so you can ask whether a policy helps the poor and the rich equally instead of learning only the average effect.

SCAFFOLDING EFFECT

psychology

Reduce cognitive load

- Go beyond the mean: estimate effects at several quantiles of the outcome distribution. - Expose heterogeneity: compare coefficients across quantiles to see who is affected differently. - Stay robust: rely on absolute-loss estimation that resists outliers in the response.

anchor

Anchor fast decisions

Ordinary least squares describes only the conditional mean, so any change in the shape of the distribution is invisible. Minimizing asymmetric absolute loss targets a chosen quantile instead, which is why one predictor can show different slopes at different points of the outcome distribution.

MINIMUM ACTION

In progress 0/1

Practice this model in one real situation:

Choose the quantiles you care about, such as 0.25, 0.5, and 0.75. 2. Fit each quantile regression by minimizing weighted absolute loss. 3. Compare coefficients across quantiles to see whether the effect is heterogeneous. 4. Use bootstrap to compute confidence intervals for each coefficient.

Check to track your progress (stored locally)

Learning progress0%

account_treeGenealogyexpand_more

menu_bookReferencesexpand_more

Source support: Explicit

link
en.wikipedia.orghttps://en.wikipedia.org/wiki/Quantile_regressionZH · Explicit
verified

RELATED MODELS