Philippines staffing research · Updated
Where does decision-queue time accumulate during interrupted review?
An event study of active review, interruption, reassignment, and waiting.
Research question: What proportion of documented holds receive an attributable decision in the same record?
Scope: six consecutive weeks in one defined back-office queue, including every eligible hold under a frozen taxonomy.
Methodology: Preserve the initial request, authority-matrix version, acknowledgments, reassignment, decision, return event, and final state. Register exclusions before analysis.
Unit of analysis: one hold asking one authorized owner for one bounded decision. Stable identifiers prevent messages and edits from inflating the denominator.
Classification: primary return, backup return, reassigned, returned for evidence, withdrawn, duplicate, expired, or unresolved. Observable states remain separate from causal judgment.
Measures: class counts and denominators, acknowledgment and decision-time distributions, return proportion, aged holds, and reopenings. Report counts, denominators, distributions, and missingness.
Validation: A second reviewer checks a stratified sample against source histories and the authority matrix effective at event time.
A Philippines-based specialist may collect permitted records and apply the frozen codebook. Internal owners retain privacy, security, legal, financial, employment, policy, and customer-remedy decisions.
Use named access, minimum necessary data, documented retention, and de-identified reporting where practical. Record exclusions before outcome review.
Inference limits: Results describe observed return reliability and associations; they cannot show that an owner, location, or staffing model caused delay.
Limitations: Off-system replies may be missing, rare decisions may not occur, and six weeks may not represent seasonal demand.
Conclusion: Repair record linkage where returns fail; do not convert findings into individual rankings.