Public-Source Compensation Benchmarking: A Method Note
Peer-survey design, job-matching discipline, and BLS anchoring for public-sector pay studies
How the Lab designs a defensible compensation/market study: a governed peer survey matched at the job-summary level, anchored and cross-checked against BLS OEWS and QCEW, with disclosed match quality, vintages, and missing-data behavior.
Purpose
Compensation and market studies fail in predictable ways: comparators chosen for convenience, jobs matched by title rather than content, undisclosed match quality, stale salary data presented as current, and missing values silently treated as zero. This note fixes the standard a WIL compensation study must meet so that a reader — an employer, a governing body, or another researcher — can verify how every figure was produced. It documents method only. No market finding is asserted here, and no result may cite this note as evidence for a conclusion the underlying data does not support.
Method requirements
Comparator (peer) sets are defined before data collection, with disclosed selection criteria — organization size, service mix, labor-market overlap, and governance form — and a minimum of ten participating entities per surveyed category.
Matching is performed at the job-summary level (duties, scope, supervision, and required credentials), never on title alone.
A match is reported only at 70 percent or greater duty overlap; matches below that threshold are disclosed as such, and blended matches (one benchmark assembled from parts of multiple jobs) are explicitly flagged.
Salary data carries its effective date; observations are aged to a common effective date with the aging factor disclosed, and survey-year data is used wherever available.
Geographic pay differentials are applied only where relevant and are disclosed with their source.
Suppressed, unreported, or unavailable cells remain missing in every table and figure — they are never recoded to zero and never imputed silently.
Every published result names its source, vintage, methodology version, and limitations.
The public-data spine: OEWS and QCEW
Peer surveys are small by construction — ten to twelve entities produce ranges, not distributions. The Lab therefore anchors every peer-survey read against public federal statistics. BLS Occupational Employment and Wage Statistics (OEWS) provides occupation-level wage percentiles (10th, 25th, median, 75th, 90th) by area, which bound what a plausible market rate can be for a matched occupation. The Quarterly Census of Employment and Wages (QCEW) provides establishment-based employment and aggregate wage levels by industry and area, which contextualize the size and direction of the relevant labor market. A peer-survey result that falls outside the OEWS interquartile band for its best-matched occupation is not suppressed — it is investigated, and the reconciliation is documented in the study. Public data anchors and audits the survey; it does not replace it, because government pay structures and utility trades are matched more precisely by job summary than by standard occupational classification alone.
Peer-survey design and collection
Comparator entities are agreed with the client before collection and disclosed in the study. Collection is by direct contact with each entity's human-resources function, using finalized job summaries rather than titles, with structured follow-up until each entity's submission is complete or its non-response is recorded. Compiled results report, per benchmark job: the number of matches, the match-quality distribution, base-pay comparisons at consistent percentiles, and — where the engagement includes it — total-compensation comparisons with each element identified. Discrepant observations (aged structures, mid-cycle pay-plan changes) are annotated rather than dropped.
Limitations
This method produces market comparisons, not forecasts, and describes pay positioning, not labor availability. Small survey populations mean results are presented as ranges with match quality disclosed, and a study's conclusions are limited to the entities, jobs, and effective dates surveyed. OEWS and QCEW anchors carry their own publication lags and estimation methods, which are cited in each study. Nothing in this method authorizes converting a pay comparison into a claim about shortage, difficulty, or capacity — constructs this note does not measure.
Version 1.0 · Published 2026-09-07 · Permanent ID WIL-RB-2026.3-COMP-METHOD. This record is versioned; the URL is permanent and stable for citation.
Export citation (BibTeX · RIS)
@techreport{WILRB20263COMPMETHOD,
title = {Public-Source Compensation Benchmarking: A Method Note: Peer-survey design, job-matching discipline, and BLS anchoring for public-sector pay studies},
author = {AlphaHire Workforce Intelligence Lab},
institution = {AlphaHire Workforce Intelligence Lab},
type = {Research Brief},
number = {WIL-RB-2026.3-COMP-METHOD},
year = {2026},
note = {Version 1.0; methodology WIL-CBM-1.0},
url = {https://wilinstitute.org/library/p/public-source-compensation-benchmarking-method},
}RISTY - RPRT AU - AlphaHire Workforce Intelligence Lab TI - Public-Source Compensation Benchmarking: A Method Note: Peer-survey design, job-matching discipline, and BLS anchoring for public-sector pay studies PY - 2026 PB - AlphaHire Workforce Intelligence Lab M1 - WIL-RB-2026.3-COMP-METHOD ET - Version 1.0 UR - https://wilinstitute.org/library/p/public-source-compensation-benchmarking-method AB - This method note documents the Lab's approach to public-sector compensation benchmarking: comparator selection, job-summary-level matching with a disclosed 70-percent match threshold, blended-match flagging, salary-data vintage and aging rules, geographic differentials, and anchoring against BLS Occupational Employment and Wage Statistics (OEWS) and the Quarterly Census of Employment and Wages (QCEW). It asserts no empirical findings; it defines the standard a WIL compensation study must meet before results are published. ER -