Use when executing and reporting the analysis for a Communication Research (CR) manuscript so it survives expert, double-anonymized review — ANOVA/regression/SEM, mediation/moderation with honest uncertainty, reliability, and APA statistical reporting. Guides analysis norms; it does not fabricate results.
NewMaintained
Automatic
◆Research pack
Last verified 9/1/2026
What it changes
Passed 15 of 23 test cases with the skill, and 15 of 23 without it.
Without skill65%
With skill65%
Checks it passed that the model alone missed
Outputs 【Next】 commres-tables-figures under the final bracketed tag.
Reports reproducibility verification using an explicit [Y/N] indicator under 【Reproducible】.
1 fixed14 passed either way7 still failing1 broke
Pass-rate lift
0 pp
Tokens
+47%
Agent turns
0%
Response time
-7%
Model gemini-3.6-flashMeasured 9/1/2026v115 of 23 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
A test case the skill fixed. Same model, same prompt, with and without the skill.
Prompt
I have prepared a draft combining automated text classification with a two-wave survey to study political communication dynamics. Before submitting to Communication Research, I want a comprehensive methodological and reporting check on my statistical estimates, scale omega values, inter-coder metrics, and script setup. Provide your review structured exactly in the bracketed output format:
【Main estimate】
【Model】
【Mediation/moderation】
【Robustness】
【Reliability】
【Confirmatory vs exploratory】
【Reproducible】
【Next】
What was checked, without and with the skill
Checked
Without skill
With skill
Evaluates human-labeled validation metrics for the text classification pipeline.
Passed
Passed
Reports reproducibility verification using an explicit [Y/N] indicator under 【Reproducible】.
Missed
Passed
Outputs 【Next】 commres-tables-figures under the final bracketed tag.
Missed
Passed
case-03 · gemini-3.6-flash · 9/1/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md104 linessingle file
Data Analysis (commres-data-analysis)
CR reviewers are quantitatively sophisticated, and the journal expects APA-style statistical
reporting (effect sizes and standard deviations, not stars alone). Analyze as if your numbers will
be scrutinized — because they will. This skill covers execution and reporting norms; design decisions
live in commres-research-design, and deposit live in commres-transparency-and-data.
When to trigger
Running main and supporting analyses; building the Results section
A reviewer asked for robustness, an alternative specification, or a mediation re-analysis
Reconciling preregistered vs. exploratory analyses
Making the analysis reproducible before deposit
Analysis norms CR expects
APA statistical reporting. Report effect sizes (d, η²ₚ, R², standardized β) and dispersion
(SDs, CIs), test statistics with df, and exact p where feasible — not significance stars alone.
Right model for the design. ANOVA/ANCOVA for factorial experiments; OLS/logistic regression
with proper controls; SEM/CFA for latent constructs; multilevel models for nested data
(e.g., messages within participants, students within classrooms).
Mediation/moderation done right. For PROCESS/SEM models, justify the causal ordering; report
indirect effects with bootstrap CIs; for moderated mediation report the index and conditional
indirect effects; acknowledge cross-sectional limits on process claims.
Report uncertainty honestly. Confidence intervals and effect magnitudes; interpret the
substantive meaning of the estimate, not just whether it crossed .05.
Robustness that probes, not decorates. Show specifications that could break the result
(alternative measures, covariate sets, estimators, exclusions), and say what you learn.
Measurement and reliability. Report scale reliability (alpha/omega) and, for content analysis,
intercoder reliability; show results are not an artifact of a coding/scaling choice.
Preregistration discipline. Clearly separate confirmatory (registered) from exploratory
analyses; reconcile and justify deviations; correct for multiple comparisons where you test many.
Computational / text-as-data specifics
Document model/version, hyperparameters, seeds, and validation against human-labeled samples.
For topic models/embeddings/LLM pipelines, report stability and a validation step; don't treat
outputs as ground truth.
Reproducibility while you work (not at the end)
One master script regenerates every table and figure from the (raw or constructed) data.
Use when executing and reporting the analysis for a British Journal of Political Science (BJPS) manuscript so it survives expert, double-blind review — honest uncertainty, robustness, and triangulation appropriate to quantitative, experimental, or computational work. Guides analysis norms; it does not fabricate results.