Candidate Source Integrity: A Traceability Study

October 8, 2026

Research question

This evidence review examines how recruiting teams can preserve raw acquisition events while applying a documented attribution rule. It treats candidate source integrity as an operating-control problem rather than a vendor ranking or a universal performance benchmark.

For this stage, test traceability explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Data Sources and Methodology

The review compares public control guidance with a proposed record design for StopHighTurnover. It maps recurring concepts to observable fields, then defines a bounded sample containing ordinary work, open exceptions, and recently closed cases. No private candidate records or production credentials were used.

For this stage, test reviewer agreement explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Unit of analysis

The unit is one process item from trigger through accepted closure. The minimum evidence set is raw event, tracking value, import path, merge history, override reason, rule version, and confidence. Each field needs a definition, permitted source, accountable owner, and rule for missing information.

For this stage, test authority explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Evidence-quality model

Reviewers score relevance, authority, freshness, completeness, and traceability separately. A current timestamp does not make an unofficial source authoritative, and a complete form does not prove that another reviewer can reproduce the observation.

For this stage, test freshness explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Sampling plan

Build a sample across teams, schedules, age bands, and exception reasons. Include the oldest open item and a reopened item. Random selection alone can miss rare high-impact failures, while judgment-only selection can exaggerate memorable problems.

For this stage, test completeness explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Reviewer agreement

Two reviewers should classify the same records independently before discussion. Compare field-level disagreement rather than only an overall pass rate. If reviewers disagree about the source, owner, or closure state, revise definitions and repeat the sample.

For this stage, test traceability explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Interpretation

A long cycle may reflect missing input, deliberate review, access failure, or unclear authority. A short cycle may still lack evidence. Pair elapsed time with waiting reason, completeness, and candidate impact before drawing an operational conclusion.

For this stage, test reviewer agreement explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Governance

Assign who owns the process, who changes definitions, who grants access, who reviews samples, and who accepts exceptions. Every change needs a reason, approver, effective date, and communication plan, with prior versions archived.

For this stage, test authority explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Privacy and security

Collect only what the review needs. Keep identity, accommodation, screening, medical, and financial data in authorized systems. The analysis record should contain a secure reference and status rather than a duplicate of restricted material.

For this stage, test freshness explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Limitations

This documentary design does not establish causation, productivity, legal compliance, or financial return. Public frameworks are broad. Local law, contracts, technology, staffing, and professional duties may require additional controls and qualified review.

For this stage, test completeness explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Operational application

Pilot the design on one bounded queue. Train reviewers on definitions, collect the sample, discuss disagreements, revise the record, and retest. Expand only after the team can reproduce classifications and show closure evidence.

For this stage, test traceability explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Conclusion

Candidate Source Integrity: A Traceability Study supports a conservative conclusion: candidate source integrity becomes more reviewable when source facts, administrative actions, management decisions, and outcomes are recorded separately. The next step is a documented pilot, not a claim that one metric proves success.

For this stage, test reviewer agreement explicitly. Record the definition used, the observed evidence, the reviewer conclusion, and the reason another conclusion was rejected. That discipline makes the finding auditable without pretending the evidence answers questions outside the study.

Replication checklist

A second team should be able to repeat the review without private instructions from the first analyst. Preserve the sampling frame, selection rule, field definitions, source locations, extraction date, reviewer worksheet, disagreement log, and approved conclusion. Record excluded items and the reason for exclusion. If a system export changes during the study, retain the original extract and describe the change rather than silently combining incompatible versions.

Before publication, ask an independent reviewer to trace one ordinary item and one exception from the reported finding back to the approved source. The reviewer should be able to reproduce timestamps, ownership, waiting reason, and closure state. Differences should be resolved in the method record, not hidden by averaging incompatible labels. Report missing evidence as missing; do not infer a favorable state from an empty field.

Replication does not prove that the control caused an outcome. It shows that the documentary finding can be followed and challenged. Future comparisons should keep the same population and definitions or disclose the change. This protects the study from false precision and gives managers a practical basis for deciding whether to repair instructions, access, capacity, or escalation authority.

Suggested evidence table

Sources

1. NIST Cybersecurity Framework 2.0

2. NIST Privacy Framework

3. CISA Identity and Access Management

4. GAO Standards for Internal Control

5. EEOC Recordkeeping Requirements

6. OPM Structured Interviews

7. FTC Protecting Personal Information

8. U.S. Department of Labor Hiring Resources

9. BLS Employee Tenure Summary

10. O*NET OnLine

Related reading

See candidate pipeline health metrics and recruiting operations dashboard for adjacent operating context.