Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› revedres-tables-figures
revedres-tables-figures
GitHub为教育研究综述和荟萃分析设计PRISMA流程图、森林图及证据表。确保数据一致性、可信度加权及APA格式规范，根据综述类型匹配图表组合，以清晰呈现综合结果。
Trigger Scenarios
需要展示系统综述或荟萃分析的PRISMA流程图
需要生成森林图或漏斗图以展示荟萃分析结果
需要构建编码/证据映射表以支持叙事或批判性综合
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill revedres-tables-figures -g -y
SKILL.md
Frontmatter
{
"name": "revedres-tables-figures",
"description": "Use when designing the PRISMA flow diagram, forest\/funnel plots, and coding\/evidence tables for a Review of Educational Research (RER) review or meta-analysis. Builds exhibits that carry the synthesis; it does not run the analyses (revedres-comprehensiveness-and-balance) or settle reproducibility of the underlying data (revedres-transparency-and-reproducibility)."
}
Tables & Figures (revedres-tables-figures)
When to trigger
- The synthesis is settled and you need exhibits that prove and carry it
- A draft describes the search in prose with no PRISMA diagram
- A meta-analysis has results but no forest plot or bias figure
- Tables list studies without making them comparable to a reader
Exhibits an RER review is expected to carry
A review's exhibits are not decoration — they are how a reader audits your search and reads your synthesis at a glance. Build the right ones for the review type.
| Exhibit | Required for | What it must show |
|---|---|---|
| PRISMA flow diagram | systematic reviews & meta-analyses | identification → screening → eligibility → included, with counts and exclusion reasons that reconcile |
| Study characteristics table | all review types | per-study: sample/context, design, measures, key result, risk-of-bias rating |
| Forest plot | meta-analysis | each effect with CI, weight, and the pooled estimate; subgroup panels for moderators |
| Funnel plot (+ Egger/trim-fill) | meta-analysis | publication-bias diagnostics, not asserted in prose |
| Coding / evidence-map table | narrative & critical syntheses | studies arranged by the framework's cells, so the spine is visible |
| Conceptual figure | critical/integrative syntheses | the organizing model or taxonomy as a diagram |
Design rules so exhibits do real work
- The PRISMA numbers must reconcile. Identified − duplicates − screened-out − full-text-excluded = included. Reviewers add them up; a mismatch undermines the whole review.
- Compare only commensurable estimates. A study-characteristics table that mixes incompatible outcomes invites false "the literature finds…" claims; group by construct and note the metric.
- Carry credibility in the table. Every finding/effect column needs a risk-of-bias rating beside it so emphasis is visibly credibility-weighted, not count-weighted.
- The forest plot is the result. Order it by the moderator that matters (so heterogeneity is visible), show weights, and place the pooled diamond with its CI and prediction interval where useful.
- The conceptual figure should be re-buildable from memory. If a reader cannot redraw your framework after one look, it is too busy.
- APA 7 formatting. Number, title, and note each table/figure per APA 7th; figures legible in greyscale; alt-text/notes where required.
Matching the exhibit set to the review type
Do not build a forest plot for a conceptual synthesis or skip the PRISMA flow for a systematic review — the expected set differs:
- Meta-analysis: PRISMA flow + study-characteristics table (with effect sizes and risk-of-bias) + forest plot + publication-bias figure are all effectively mandatory. A subgroup forest or meta-regression figure carries the moderator story.
- Systematic review (narrative): PRISMA flow + a study-characteristics table + an evidence-map table organized by the framework's cells; a forest plot is inappropriate if effects are not pooled.
- Critical / conceptual synthesis: a conceptual figure of the framework is the centerpiece; an evidence-map or comparison table shows how the literature populates it. No PRISMA flow is required if the synthesis is explicitly non-systematic, but the search logic must still be described.
The exhibits should let a skim-reader reconstruct the search, the corpus, and the synthesis without reading the prose — that is the test of a good RER exhibit set.
Checklist
- PRISMA flow diagram present (systematic/meta) and its counts reconcile exactly
- Study-characteristics table: sample, design, measure, result, risk-of-bias per study
- Forest plot ordered by a meaningful moderator; weights + pooled estimate + CI shown
- Publication-bias figure (funnel/Egger/trim-fill) included for meta-analyses
- Narrative/critical review has an evidence-map table arranged by the framework's cells
- A conceptual figure renders the organizing model (for integrative syntheses)
- Only commensurable outcomes are tabled together; metric noted per group
- Tables/figures formatted to APA 7; legible in greyscale; embedded in the Word manuscript
Anti-patterns
- No PRISMA diagram, or one whose numbers do not add up
- A study table that mixes incompatible outcomes into one comparison
- A forest plot with no weights, no pooled estimate, or arbitrary study order
- Asserting "no publication bias" in prose with no funnel/Egger/trim-fill figure
- A conceptual figure so cluttered it cannot be redrawn from memory
- Finding columns with no adjacent quality rating, so weak and strong studies look equal
- Pasting original-regression output of the author's own data — RER reviews synthesize, not report new estimates
Output format
【Review type】systematic | meta-analysis | critical-integrative
【PRISMA】flow diagram present; counts reconcile? Y/N
【Study table】sample/design/measure/result/RoB per study? Y/N
【Forest plot】ordered by moderator; weights + pooled CI? Y/N (meta)
【Bias figure】funnel/Egger/trim-fill included? Y/N (meta)
【Evidence map】arranged by framework cells? Y/N (narrative/critical)
【Conceptual figure】model/taxonomy diagram present + re-drawable? Y/N
【APA 7】numbering, titles, notes, greyscale legibility checked? Y/N
【Next step】→ revedres-writing-style (write the synthesizing RER voice; abstract + intro last)
Version History
- 1839142 Current 2026-07-05 14:22
