Table 1 Baseline Characteristics Generator
Generate the "baseline characteristics table" — the most common table in clinical papers — in one click: paste raw data with a header row and pick the grouping column, and it automatically gives per-group descriptive statistics and between-group comparison P-values by variable type — continuous variables as mean±SD (t-test/ANOVA) or median (IQR) (non-parametric tests), categorical variables as n(%) (χ²/Fisher). Computed locally in your browser; data are not uploaded.
① Paste data
The first row is column names; each subsequent row is one subject; columns are separated by Tab or comma (you can paste directly from Excel). Space-separation is unreliable — use commas or Tab.
How to use & methodology
How should I prepare the data?
Arrange it as a table with 'one subject per row, one variable per column', including a grouping column (e.g. treatment/control). Select the area with headers in Excel, copy, and paste into the box (Tab-separated); comma-separated CSV is also supported.
Should I choose mean±SD or median (IQR) for continuous variables?
Use mean±SD (with t-test/ANOVA) when the data are approximately normal and symmetric; use median (IQR) (with non-parametric tests) for skewed data, outliers, or small samples. The tool provides a global toggle — choose based on your data's distribution.
What if a variable type is judged wrong?
The tool judges automatically: columns that are all numeric with many distinct values are treated as continuous, the rest as categorical. If a coded categorical variable (e.g. stage 1/2/3) is misread as continuous, change it to categorical in 'Settings'.
How are the P values computed?
For continuous variables, two groups use the t-test (or non-parametric Mann-Whitney) and three or more groups use ANOVA (or Kruskal-Wallis); for categorical variables, the χ² test is used, switching to Fisher's exact test when a 2×2 table has small expected counts. Complex designs (paired, covariate-adjusted) require dedicated models.