er-tables-figures

Tables and Figures (er-tables-figures)

When to Trigger

• Main table has 8 or more columns, and reviewers complain that "information density is too high"
• The table uses borders / shading / colors — not compliant with the Economic Research Journal standards
• Figure resolution is insufficient / color scheme is unprofessional

Table Standards (three-line table / booktabs)

Required Structure

表 3  政策冲击对企业研发投入的影响
═══════════════════════════════════════════
                  (1)      (2)      (3)
变量           研发支出  研发支出  研发支出
───────────────────────────────────────────
政策×处理后  0.045***  0.038***  0.040***
              (0.012)   (0.011)   (0.013)
              ......
───────────────────────────────────────────
控制变量          否       是       是
个体固定效应       是       是       是
年份固定效应       是       是       是
观测值          10,234   10,234   10,234
R²               0.213    0.247    0.251
═══════════════════════════════════════════
注：括号内为聚类于企业层面的稳健标准误；***、**、* 分别表示 1%、5%、10% 显著水平。

Strictly Follow Constraints

• Restrained Column Count(rule of thumb about ≤6 columns; split the table when information density is too high, subject to the official website at the time of submission)
• Three-line table: thick top rule, thin middle rule, thick bottom rule, no vertical lines
• Export tables with code (esttab/outreg2), do not copy by hand; see resources/code/stata/09_tables.do
• Significance markers *** ** * uniformly use three levels
• Standard errors in parentheses below the coefficients,do not use t-values or z-values
• Control variables row written as "Yes / No"
• Table number + table title above the table, notes below the table
• Notesmustinclude: clustering level, definition of significance symbols

Variable Definition Table

There must be a variable definition table before the main regression table; Economic Research Journal format conventions:

Variable Type	Variable Name	Variable Definition	Data Source
Dependent variable	tax_deviation	Corporate income tax expense deviation = actual tax burden - expected tax burden	CSMAR
Core explanatory variable	digital	Golden Tax Phase III implementation dummy variable	National Tax Survey
Control variable	size	Firm size (log of assets)	CSMAR
Control variable	lev	Debt-to-asset ratio	CSMAR
Instrumental variable	iv_distance	Shortest geographic distance to a Golden Tax Phase III pilot city	Calculated by the authors

Constraints:

• Each variablehas one and only one rowof definition, do not redefine it again in the main text
• The variable definition must give thecalculation formula(not a description), e.g., "= actual tax burden - expected tax burden", not "reflects tax burden deviation"
• Data sourceexact to the database name, do not write "public channels"

Data Description Paragraph Template

There must be a data description paragraph (≈ 200 words) before the variable definition table, with a fixed structure:

This paper uses [time span] of [Database A] and [Database B]; the research sample includes [N] firms / [N] observations in [industry/region]. Sample screening criteria: (1) exclude the financial industry; (2) exclude ST/*ST samples; (3) exclude samples with missing key variables. To mitigate the influence of extreme values, all continuous variables were winsorized at the 1% and 99% percentiles. Data merging method: using [firm ID / year] as the matching key.

Taboo: do not write "data come from public channels" — you mustname the database。

Figure Standards

Required Structure

• Font: SimSun / Times New Roman 9–11 pt
• Colors: black and white + 1–2 main colors (recommended dark blue / dark red)
• Resolution: ≥ 300 dpi
• Figure number + figure title belowthe figure
• Notes: must have data source + key variable definitions

Common figure types

1. Parallel trends event study plot — 95% CI, vertical dashed line marking the treatment timing
2. Policy timeline plot
3. Mechanism path diagram
4. Heterogeneity forest plot
5. Structural estimation counterfactual plot(if applicable)

Execution bridge (StatsPAI / Stata MCP)

Tables and figuresGenerated from fitting results, do not hand-copy numbers. For the complete mapping see execution-with-mcp. Economic Research Journal is a top Chinese economics journal; identification credibility is usually the binding constraint; staggered DID, weak-instrument-robust IV, RDD, and mechanism tests.

• Table:etable(multiple columns) or did_summary_to_latex directly from result_id generate.
• Figure:plot_from_result / enhanced_event_study_plot / event_study_table, axis units and standard error/clustering notes are self-contained.
• Each table notestate the estimator and clustering level, and report economic magnitudes in interpretable units.

For the complete "fitting results → tables and figures" chain see JF execution walkthrough。

Must-check list

• Number of columns in the main table kept restrained (rule of thumb about ≤6 columns, subject to the official website)
• Three-line table, no vertical lines
• Standard error format consistent
• Notes include clustering level
• Figures above 300 dpi
• Table and figure numbering consecutive
• Equations on separate lines, centered, numbered on the right

Anti-patterns

• Cramming R² + adjusted R² + pseudo R² + AIC + BIC + …… into a table — choosing 1–2 is enough
• Using a long string of explanations in the legend
• Overlaying 5 curves + 5 line types + 5 colors in one figure
• Directly screenshotting Stata / R output and pasting it into Word

Output format

【主表列数】X 列
【三线表合规】是 / 否
【标准误格式】统一 / 不统一
【图形 DPI】X
【缺漏注释】[...]
【下一步】er-policy-implication