Data Analysis & Robustness (jar-data-analysis)

When to trigger

• Data are built and it is time to estimate and report
• You are unsure how to cluster standard errors for your panel
• Referees will probe endogeneity, measurement, or the channel
• You must assemble the reproducible data-and-code package JAR posts

Get the standard errors right (a JAR signature check)

Empirical-accounting referees scrutinize inference. Default to clustering by firm, and consider two-way clustering by firm and year (Petersen) when both cross-sectional and time-series dependence are present. With few clusters, use the wild-cluster bootstrap rather than asymptotic cluster-robust SEs. Match the clustering to the source of correlated shocks implied by your design, and report the choice explicitly.

Execute the identification, don't just assert it

• DiD: report pre-trends and use heterogeneity-robust estimators for staggered timing (Callaway-Sant'Anna / Sun-Abraham), not naive two-way FE.
• RD: report the optimal bandwidth, robust bias-corrected estimates, a manipulation (density) test, and covariate balance at the cutoff.
• IV/2SLS: report the first stage and instrument strength (e.g., F-statistic), and defend the exclusion restriction in words.
• Event studies: report abnormal returns with a defensible benchmark model and window, and confront confounding events.

Measure accounting constructs credibly

Use measures with precedent in prior JAR/JAE work (discretionary accruals, earnings persistence/smoothness, disclosure indices, comparability, audit-quality proxies, bid-ask spread / PIN for information asymmetry). Show the proxy behaves sensibly (validation, correlations with established measures) and test sensitivity to alternative proxies — proxy fragility is a common rejection reason.

The robustness battery referees expect

• Alternative specifications: controls in/out, alternative fixed effects, alternative measures of the key construct.
• Subsamples and falsification/placebo tests (e.g., effect absent where theory says it should be).
• Sensitivity of identification assumptions (alternative instruments, donut RD, bounds).
• Cross-sectional partitions that confirm the predicted channel (the conditional predictions from jar-theory-development).
• Sample-construction and winsorization choices documented and varied.

Reproducibility is a deliverable, not a courtesy

JAR's data-and-code sharing policy requires and hosts the materials. Keep top-to-bottom runnable scripts that regenerate every table and figure from raw extracts; document screens, vintages, and access dates; respect the terms of use (academic-research-only, acknowledgement of the JAR publication and code authors, authors retain copyright). On a Registered Report, the executed analysis must match the pre-approved Stage 1 protocol.

Execution bridge (StatsPAI / Stata MCP)

Run the battery, don't just enumerate it. Full map: execution-with-mcp. JAR is archival/empirical accounting; foreground identification around disclosure and regulation shocks, with modern DiD where adoption is staggered.

• 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

• SE clustering matches the design (firm / firm-and-year; wild bootstrap if few clusters)
• Identification executed with diagnostics (pre-trends / bandwidth / first-stage / balance)
• Modern DiD estimators used for staggered treatment timing
• Key construct validated; results robust to alternative proxies
• Robustness, falsification/placebo, and channel partitions reported
• Winsorization/sample screens documented and varied
• Reproducible data-and-code package assembled per JAR policy

Anti-patterns

• White/robust SEs on panel data ignoring within-firm correlation.
• Naive two-way-FE DiD with staggered adoption.
• One proxy, no validation for a contested accounting construct.
• Significance fishing across windows, bandwidths, or specifications.
• "Code available on request" — JAR requires the package to be posted.

Output format

【Estimator & SEs】model; clustering (firm / firm×year / wild bootstrap)
【Identification executed】diagnostics reported (pre-trends/bandwidth/first-stage)
【Construct measurement】proxy + validation + alt-proxy robustness
【Robustness/falsification】[...]
【Channel partitions】conditional predictions confirmed? [...]
【Reproducibility】data/code package status per JAR policy
【Open issues for referees】[...]
【Next step】jar-contribution-framing

Resources

• ../../resources/official-source-map.md — official JAR/Chicago Booth/Wiley URLs (accessed 2026-06-01)
• ../../resources/external_tools.md — econometric packages (reghdfe / fixest / csdid / rdrobust / boottest) and data sources