Pom Data Analysis

Analysis & Results (pom-data-analysis)

When to trigger

• The model is built or the data are collected and it is time to produce results
• You are unsure your numerics, identification, or validation will satisfy reviewers
• A reviewer says "the analysis does not support the inference" or "magnitude is unclear"

Analytical / modeling papers (POM's anchor track)

For optimization, stochastic, and game-theoretic work, the "analysis" is proof plus numerical illustration:

• Proofs. State each result as a numbered proposition/theorem; give clean, complete proofs. Per POM's format, push full proofs and supporting lemmas to the unlimited online e-companion, leaving intuition and the key steps in the main text.
• Structural insight. Report the structure of the optimal policy (base-stock, threshold, (s, S)) and comparative statics — how the decision moves with cost, lead time, or competition.
• Numerical study. Calibrate to realistic operational parameters; report sensitivity across plausible ranges; show the managerial magnitude of the effect, not just its sign.
• Game-theoretic checks. Confirm equilibrium existence/uniqueness; report off-equilibrium robustness where relevant.

Empirical, behavioral, and data-science papers

• Identification (empirical OM). Make the causal logic explicit; report the design (DiD/IV/RD/matching), parallel-trends or instrument validity, placebo tests, and clustered/robust standard errors matched to the operational sampling.
• Experiments (behavioral OM). Report randomization checks, power, manipulation and attention checks, and effect sizes; tie the result to the operational decision (e.g., order quantity, not just a rating).
• Operations data science. Report validation design, guard against leakage, and — decisively — the operational value: does the prediction improve a feasible policy or reduce a real operating cost (predict-then-optimize)?
• Simulation. Document parameter sources, seeds, warm-up, replications with confidence intervals, and sensitivity.

POM-specific reporting risks

• Operational variables named but measured in units a manager cannot act on.
• Statistical significance reported in place of managerial magnitude.
• ML accuracy reported with no link to an operations policy or cost.
• Same data used in prior work without the required cover-letter disclosure.

Execution bridge (StatsPAI / Stata MCP)

Run the battery, don't just enumerate it. Full map: execution-with-mcp. POM spans analytical and empirical OM; apply the chain below to its empirical-OM papers, and note when a contribution is analytical / optimization.

• Many outcomes / specifications: romano_wolf (step-down FWER) or benjamini_hochberg — report the adjusted threshold.
• OVB sensitivity: oster_delta / sensemakr.
• Inference: wild_cluster_bootstrap (few clusters), twoway_cluster / conley; multilevel data → cluster at the right level.
• Re-fit off one handle: audit_result(result_id) lists the missing checks and the exact suggest_function for each.
• Exhibits: etable / did_summary_to_latex from the handle — no retyped numbers.

Keep the decisive checks in the body and the exhaustive battery in the appendix. See the executed chain in the JF execution walkthrough.

Checklist

• Analytical: proofs complete (in e-companion), structural results + calibrated numerics + sensitivity
• Empirical: identification stated; placebo/robustness; SE clustering matches sampling
• Experiment: randomization, power, manipulation checks, effect sizes
• Data science: validation, leakage checks, operational value demonstrated
• Results expressed in decision-relevant operational magnitude
• Same-data disclosure prepared for the cover letter

Evidence pass for Production and Operations Management

Treat this skill as an executable review pass, not a prose hint. First lock the operational decision, the performance metric, and the implementable lever; then judge whether the current manuscript answers the venue's real reader: POM reviewers who want operational insight tied to production, service, supply-chain, or platform decisions.

• Do the pass: Audit the research design before polishing prose: unit of analysis, comparison set, uncertainty, sensitivity, missingness, and reproducibility must be visible.
• Return a ledger: give claim / evidence / risk / manuscript location rows, so the next agent can edit rather than rediscover the issue.
• Sibling guard: compare against Management Science for broader OR/MS theory, Operations Research for method-first optimization, MSOM for manufacturing/service operations depth; if a sibling owns the contribution, recommend re-routing before polishing format.
• Stop condition: do not give submission-ready advice until the pack's resources/official-source-map.md has been checked for volatile rules and the manuscript has one concrete fix for the largest venue-specific risk.

Output format

【Analysis type】analytical-proof / causal / experiment / simulation / predictive
【Core result】policy structure / estimate / treatment effect / decision gain
【Main threat】proof gap / identification / leakage / measurement / power
【Managerial magnitude】effect in operational units (cost, fill rate, wait time)
【e-companion】proofs / extra analyses moved online
【Next step】pom-contribution-framing

Source: brycewang-stanford/Awesome-Journal-Skills → Production-and-Operations-Management-Skills/skills/pom-data-analysis/SKILL.md