MENTAL MODEL · M7836

Jackknife Method

TechnicalHigh supportStatistics

Included

Version 1.0.0 · Updated 2026-07-30

CORE DEFINITION

Sequentially remove one observation, recompute the statistic, and use this to estimate bias and standard error.

SCAFFOLDING EFFECT

psychology

Reduce cognitive load

Use "leave-one-out" to assess the stability of the estimate. It is a resampling technique that predates the Bootstrap.

anchor

Anchor fast decisions

Proposed by Quenouille (1949) and named by Tukey (1958). From a dataset of n samples, sequentially delete the i-th observation, recompute the estimator with the remaining n-1 observations, yielding n "pseudo-values" from which bias and variance are estimated. The mechanism is: by comparing the "overall estimate" with the "leave-one-out estimates", extrapolate to reduce bias.

MINIMUM ACTION

In progress 0/3

Practice this model in one real situation:

For i=1..n, compute the statistic θ₍ᵢ₎ after removing the i-th sample.
Compute pseudo-values ψᵢ = nθ − (n−1)θ₍ᵢ₎.
Use the sample of pseudo-values to obtain a bias-corrected estimate by averaging, and estimate the standard error from their variance.

Check to track your progress (stored locally)

Learning progress0%

account_treeGenealogyexpand_more

menu_bookReferencesexpand_more

Source support: Explicit

link
zh.wikipedia.orghttps://zh.wikipedia.org/wiki/%E5%88%80%E5%88%87%E6%B3%95ZH · Explicit
verified

RELATED MODELS