Tables & Figures for a Review (arecon-tables-figures)

Tables & Figures for a Review (arecon-tables-figures)

When to trigger

• The synthesis is done and an adjacent-field reader needs to see the field at a glance
• A controversy or a body of estimates would be clearer as a table than as prose
• The organizing framework would land better as a diagram
• You are tempted to paste a regression table — but this is a review, not a primary paper

The three exhibit types an ARE review actually uses

ARE exhibits summarize across the literature; they do not present the author's own estimation. The workhorses:

Building a credible summary table

• Compare like with like. Group studies that estimate the same object; never put non-comparable estimands in one magnitude column (the cross-study pooling error from arecon-evidence-standards).
• Carry uncertainty and credibility. A finding column without a credibility/identification note invites the question "but how good is that study?" — answer it in the table.
• Self-contained captions. An ARE exhibit is read on its own by an outsider; the caption states what the table shows, the unit, how to read a row, and the source studies.
• Source every cell. Each entry traces to a paper in the evidence matrix; an unsourced number in a review of record is indefensible.

The conceptual figure is ARE's signature

Because the readership is cross-field, the framework figure carries disproportionate weight: it is what an adjacent economist remembers and reuses. Invest in one diagram that renders the spine cleanly — a taxonomy tree, a mechanism flow, or the simple model — so a reader who recalls nothing else can reconstruct the field's structure from it. Annual Reviews production supports full-color figures; design for legibility in the published two-column format and confirm current figure specs on the author pages (检索于 2026-06；以官网为准).

Meta-analysis caution

If you assemble effect sizes into a quantitative synthesis (forest plot, meta-regression), you are doing a meta-analysis, with its assumptions — comparable estimands, publication-bias diagnostics, weighting. Do this only where the literature genuinely supports it; otherwise a qualitative who-found-what table is more honest than a spurious pooled number. ARE readers include the methodologists who would catch invalid pooling. If you do run a meta-analysis, its data/code should be reproducible (see arecon-submission).

Reproduced vs. re-drawn figures

When a figure from a reviewed paper is central, prefer a re-drawn synthesis figure (your own panel placing several studies on common axes) over copying one paper's exhibit — it serves the review's argument and avoids consensus-by-accident. If you reproduce an original figure, attribute it and secure any permission Annual Reviews requires (检索于 2026-06；以官网为准).

Checklist