From jais-skills
Runs and reports empirical analysis for JAIS manuscripts: SEM measurement/structural models, causal identification, artifact evaluation, or qualitative data structure.
How this skill is triggered — by the user, by Claude, or both
Slash command
/jais-skills:jais-data-analysisThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- Data are collected (or the artifact built) and it is time to estimate, evaluate, and report
JAIS's pluralism means there is no single mandated estimator; the standard is the rigor norm of your tradition, reported transparently enough for a developmental Senior Editor to interrogate. Pick the row.
| Tradition | What to report |
|---|---|
| Behavioral | reliability (alpha/CR), CFA or PLS measurement model, AVE, discriminant validity (Fornell-Larcker / HTMT); structural paths with effect sizes; mediation via bootstrap CIs; moderation via simple slopes |
| Economics of IS | the identifying variation, parallel-trends/exogeneity evidence, clustered SEs, and a robustness battery (alternative specs, placebo/event-time tests, sensitivity to the key assumption) |
| Design science | artifact performance against credible baselines on held-out data; ablations; field/A-B or expert evaluation tied to design propositions; cost/utility discussion |
| Qualitative / interpretive | a transparent data structure (codes → themes → dimensions), an audit trail, and representative quotations tracing raw data to constructs |
Report the measurement model first: reliabilities, AVE, and discriminant validity. PLS-SEM suits predictive/formative models; covariance-based SEM suits theory-testing with reflective constructs — justify the choice. Address common-method bias beyond a single-factor (Harman) test — a marker variable, an unmeasured method factor, or showing interactions survive. Then report structural paths with effect sizes, not just significance. Crucially, JAIS requires you to "provide a full correlation matrix or covariation matrix as a part of articles (appendix)" for SEM studies, plus descriptives — prepare this now, not at proof stage (检索于 2026-06;以官网为准).
Lead with the identification logic, then stress-test it: alternative specifications, placebo and event-study plots, sensitivity to the key assumption, and clustering matched to the data structure. With staggered timing, use a modern estimator and show flat pre-trends. Report magnitudes and their economic meaning, not just stars.
A robustness battery at JAIS should be legible as protecting the theoretical claim, not as a checklist. For each check, state the threat it neutralizes: a placebo test guards against spurious timing, an alternative specification guards against functional-form dependence, a sensitivity analysis bounds the unobserved-confounding concern. Listing checks without naming the threat each addresses is a recurring pushback — and at a theory-forward journal, an unmotivated robustness section signals that the author is not sure which threat actually endangers the contribution.
Benchmark against the baselines a skeptic would name, run ablations to show which design principles matter, and connect each result back to a design proposition. Where feasible, evaluate in a realistic field setting. Utility for a real problem is the contribution.
Show the coding structure and an audit trail so a reader can follow how raw material became constructs. Representative quotations and negative cases build trustworthiness; the analytic narrative, not a coefficient, carries the claim.
JAIS is theory-forward, so analysis that floats free of the theoretical argument reads as dredging. After each estimate, evaluation, or coded theme, state explicitly which hypothesis, proposition, or construct it bears on and what it implies for the mechanism. A results section that marches through coefficients without re-connecting to the theory invites the "strong finding, thin contribution" critique — the most common JAIS pushback. The discipline is to report the number and immediately say what the field now knows because of it.
JAIS policy requires authors to make datasets "available on request for checking by senior editors or reviewers after care has been taken to anonymize the data," and for quantitative studies to provide "the co-variance or correlation matrix plus descriptives." If you reuse a dataset, you must justify it for an alternative theoretical purpose or a new methodological approach. Assemble these now and keep them anonymized for double-blind review.
A behavioral paper reports a clean PLS-SEM with all paths significant, strong reliabilities, and HTMT discriminant validity — but no correlation matrix and only a Harman test for common-method bias. At JAIS this stalls twice: the missing matrix violates an explicit submission requirement (datasets and the covariance/correlation matrix must be available for SE/reviewer checking), and the single-factor CMB defense is the field's textbook example of an insufficient remedy. The fix is concrete: add the correlation/covariance matrix and descriptives as an appendix, add a marker-variable or unmeasured-method-factor analysis, report effect sizes alongside the path coefficients, and prepare the anonymized dataset for on-request checking. None of this changes the model; all of it changes whether a developmental SE can defend the paper.
【Tradition & analysis】SEM / DiD-IV-RD / artifact eval / qualitative
【Validity or identification】measurement + CMB / identification + robustness / baselines + ablations
【Effect sizes / utility】magnitudes and meaning
【JAIS data materials】correlation/covariance matrix + descriptives + anonymized dataset on request: ready/gaps
【Source status】verified URL / 待核实
【Next skill】jais-contribution-framing
npx claudepluginhub brycewang-stanford/awesome-journal-skills --plugin jais-skillsExecutes and reports empirical analysis for MIS Quarterly manuscripts across behavioral IS, economics-of-IS, design science, and qualitative traditions, including transparency materials.
Guides selection and defense of research design for JAIS manuscripts, matching method to claim under methodological pluralism. Covers behavioral, economics-of-IS, design-science, qualitative, and review traditions.
Executes and stress-tests econometric, SEM/PLS, analytical-model, or ML analyses for JMIS manuscripts. Handles identification, endogeneity, construct validity, and robustness checks.