A forest plot puts study-level effect estimates and confidence intervals into one aligned figure. SciDraw can generate the first plot from a structured results table, while SciVis provides the follow-up workspace for labels, spacing, colors, annotations, and layout.
What data does a forest plot need?
Upload one CSV, TSV, TXT, XLSX, or XLS file. Excel imports the first worksheet. A useful table includes:
| Column | Purpose | Example |
|---|---|---|
study | Study label | Adams et al. 2015 |
effect | Point estimate | 0.944 |
ci_lower | Lower confidence limit | 0.715 |
ci_upper | Upper confidence limit | 1.174 |
weight_percent | Optional study weight | 15.1 |
subgroup | Optional subgroup label | Overall |
metric | Effect measure | Odds ratio |
Each row should represent one estimate. Keep confidence limits numeric and make sure the lower limit is below the point estimate and the upper limit is above it.
Step 1: Choose an example or upload your table
Open the forest plot generator. Six datasets demonstrate odds ratios, risk ratios, hazard ratios, mean differences, subgroup analysis, and prevalence. Choose the closest example to your analysis or upload your own file.
One file is normally sufficient because study labels, estimates, intervals, weights, and subgroups can be stored as columns in one table.
Step 2: Generate the forest plot
Map the effect and confidence-interval columns, choose the effect measure, and generate the chart. The effect measure controls the reference line:
- use a null value of 1 for odds ratios, risk ratios, and hazard ratios;
- use a null value of 0 for mean differences and standardized mean differences;
- use a log x-axis for ratio measures when appropriate;
- size study markers by weight only when valid weights are supplied.
Step 3: Continue editing in SciVis
After generation, select Continue editing in SciVis. Use the editor to align study labels and numeric columns, adjust row spacing, move legends or annotations, refine the axis, and keep subgroup headings readable.
If you include a pooled diamond, heterogeneity statistics, or subgroup totals, use values produced by your meta-analysis software. The visualization tool should present the results, not replace the statistical model.
Pooled effects, weights, and subgroups
A pooled diamond is centered on the summary estimate and spans its confidence interval. Study marker size may represent meta-analysis weight. Subgroup plots can add a subtotal diamond for each group and an overall estimate at the bottom.
For transparent reporting, name the model used to calculate the pooled result, such as fixed effect or random effects, and report relevant heterogeneity statistics in the caption or figure when required.
Common forest plot mistakes
- Wrong null line: Use 1 for ratios and 0 for differences.
- Mixing effect measures: Do not combine odds ratios and risk ratios in one pooled scale without a justified conversion.
- Invalid confidence intervals: Check lower, estimate, and upper values before plotting.
- Weights that do not match the model: Import weights from the actual meta-analysis.
- Missing scale information: State whether a ratio axis is logarithmic.
- Unverified pooled statistics: Confirm diamonds and heterogeneity values against the analysis output.
- Crowded columns: Keep study labels, estimates, intervals, and weights in aligned lanes.
Frequently asked questions
Can I make a forest plot without RevMan or R?
Yes. Upload a structured effect-size table to the forest plot generator, generate the figure, and continue editing it in SciVis.
Can I upload Excel data?
Yes. XLSX and XLS are supported, with the first worksheet imported. CSV, TSV, and TXT are also accepted.
Is one file enough for subgroups?
Yes. Add a subgroup column to the same table and assign each study row to its group.
Does the tool calculate a meta-analysis?
Treat it as a visualization workflow. Supply and verify effect estimates, confidence intervals, weights, pooled results, and heterogeneity statistics from your statistical analysis.
Create your forest plot
Start from one of six sample datasets or upload your own results table in the forest plot generator. Generate the aligned plot, verify every estimate, and finish the publication layout in SciVis.
