startthat.work home

CK-03 Bootstrap confidence intervals for agreement and accuracy (bootstrap resampling of items)

Quantify sampling uncertainty in an agreement or accuracy figure computed from a review sample by resampling items, and produce an interval around the coefficient.

Check extracted resultsEvidence: DocumentedVerification technique, runs in step 6Strongest statement: describe

What it does

Quantifies sampling uncertainty in a reliability or accuracy figure computed from a finite review sample, by resampling items. Artefact: an interval around the coefficient. Read the lower bound against the threshold needed.

When to use it

Use when you have an agreement or accuracy figure from a small review sample and need to know how precise it is.

When it does not fit

Needs independent items. Clustered items, such as many snippets from one document, need cluster resampling (inference). Small samples give wide intervals.

Result format

The result gives, for each reliability or accuracy figure, the point value and its bootstrap interval, with the sample size beside it. It states whether the lower bound of each interval meets the threshold needed.

Requirements on the material

PropertyRequirement typeRequirement
Unit of analysisAlways appliesrequires=record; requires=document; requires=claim
Level of detail availableAlways appliesrequires=record-level
Appraises source credibilityApplies in some casesno

Not restricted on: Structure, Content type, Variable types, Time structure, How the sample was chosen, Number of sources, Whether sources can be joined, Amount of material, Known data problems, Producer interest, Source dependence, Origin, Document genre, Language, Personal information, Security classification, Licence.

Minimum materialat least 30 independent reviewed items
Additional sources needednone
Not for (rules)none
Not for (text)Wrong choice when items are clustered, such as many snippets from one document, unless clusters are resampled. Small samples give wide intervals that must not be hidden.

Choosing and ordering

Kind of techniqueVerification technique, runs in step 6
Phase tagvalidate
Other phase tagsnone
Kind of findingrisk
Question answeredhow much
Strongest statement it supportsdescribe
Needs as input
Makes as output
Must exist firstNone
Loop targetproceed with limit stated

Library checks by stakes

Each stakes level includes the checks of the level below it. Checks that need a person apply at high stakes only, so the process leaves them out at lower stakes (decision 38). Repeatability: repeatable with fixed settings.

StakesCheckPerson involvedCounts as a different method
exploratoryinput validationnoneno
exploratoryrepeat runsnoneno
decision-informinginput validationnoneno
decision-informingrepeat runsnoneno
decision-informingcode execution for arithmeticnoneno
high-stakesinput validationnoneno
high-stakesrepeat runsnoneno
high-stakescode execution for arithmeticnoneno
high-stakessensitivity analysisnoneno
high-stakesindependent reviewneededno

Failure modes and limits

Model suitabilitymodel plus code
Typical model failureReports an interval without computing it, which gives invented precision.
Error sourcesinput data; statistical assumptions
Certainty limitsCertainty drops when items are not independent (risk of bias) and when the reviewed sample is small (imprecision).

Evidence and tag basis

Documented. The source recommends bootstrapping for alpha and reporting confidence intervals. General bootstrap theory was not opened. (Artstein and Poesio 2008, Inter-coder agreement for computational linguistics)

Tag basis: input profile inference; selection inference; sequencing inference; robustness inference. All tag values are inference unless a source is named.