domains / families / research / jop-data-analysis

jop-data-analysis

separate

Use for analysis-stage decisions on a The Journal of Politics (JOP) manuscript — uncertainty, robustness, and reporting norms — written so the work is reproducible from line one. JOP makes acceptance contingent on replicability and a JOP replication analyst re-runs the code, so every reported number must come from a script. Guides analysis; it does not fabricate results.

When to use

• Setting up the estimation/analysis pipeline
• Deciding which robustness checks belong in the main text vs the Online Appendix
• Responding to reviewer requests for additional specifications, uncertainty, or sensitivity
• Preparing numbers that must match the deposited replication package exactly
• Reporting uncertainty, CIs, SEs, and substantive effect sizes
• Ensuring specification transparency and targeted robustness
• Fitting the analysis to the page budget
• Executing analysis via StatsPAI or Stata MCP

When not to use

• Fabricating results
• Hand-editing tables and figures
• Reporting numbers that the deposited code cannot reproduce
• Using unseeded randomness or unpinned versions
• Star-gazing with no effect sizes or uncertainty
• Choosing robustness checks to inflate the count rather than rebut a threat
• Cramming every specification into the main text

Skill page

Source