# Make the Table 1 for your manuscript

> Upload the dataset and name the groups. Yila checks each variable's distribution, summarises it the way reviewers expect, picks the test row by row and saves an editable Word table.

URL: https://yila.ai/playbooks/table-1

## What you get

- **An editable Word table** (DOCX): Ready to paste into the manuscript, with group sizes in the header and the test named on every row.
- **The same table as data** (CSV): Every cell in a CSV, for a journal's own template or a supplementary file.
- **The reasoning per row** (CSV): Which summary and which test each variable got, with the normality check behind the choice.
- **The script** (Python script): The complete code in Python (or R if you ask), so the table can be rebuilt when the data change.

## How it works

1. **Upload the data and name the groups**: Give Yila a CSV or Excel file and say which column defines the groups. Mention how categories are coded if it is not obvious, for example 1 = male.
2. **Yila decides how to summarise each variable**: It checks each continuous variable's distribution in every group, then picks mean (SD) or median (IQR) and the matching test, and reads the coding of categorical variables.
3. **Get the table in Word**: The table is built, checked against the counts and saved as Word and CSV with the script. Ask to add a variable, report percentages by column or switch to Fisher's exact test in the same conversation.

## Made for

- **A clinical study manuscript**: The baseline table almost every journal asks for, comparing arms, exposure groups or outcomes.
- **A registry or cohort report**: Describe a large dataset by subgroup without formatting hundreds of cells by hand.
- **A thesis methods chapter**: A consistent descriptive table with the tests explained, ready for the committee's questions.

## Questions

### What is a Table 1?

The first table of most clinical and epidemiological papers. It describes the study population at baseline, usually split by group, so readers can judge who was studied and whether the groups were comparable.

### How does Yila choose mean (SD) or median (IQR)?

It checks each continuous variable for normality in every group and looks at skewness. Roughly normal variables get mean (SD) and a t-test; skewed ones get median (IQR) and a Mann–Whitney U test. The decision for each row is in the output, so you can override it.

### Should a Table 1 have p-values?

For randomised trials, CONSORT advises against testing baseline differences, and standardized mean differences are the better guide. For observational studies they are common. Yila includes both by default; say if your journal wants the p-values removed.

### What about more than two groups?

Each group gets its own column, and the tests switch to ANOVA or Kruskal–Wallis for continuous variables and chi-square for categories.

### How are missing values handled?

Yila counts them per variable and group. If a variable has missing data, the number of available cases is shown, so the denominators stay honest.

### Can it follow my journal's format?

Yes. Say how many decimals, which units, and whether categories should show n (%) or % only, and the Word table follows. The CSV lets you drop the numbers into any template.

### Where does the example data come from?

From the open Heart Failure Clinical Records dataset (Chicco and Jurman, 2020), published under CC BY 4.0 on the UCI Machine Learning Repository. Yila never uses users' data as examples.

---

- [Make the Table 1 for your manuscript](https://yila.ai/playbooks/table-1)
- [llms.txt](https://yila.ai/llms.txt)
