> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pollandroll.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Building crosstabs

> Crossing two questions, or the whole questionnaire by one variable.

A **crosstab** puts two questions side by side to see how the answers to one vary with the other. The classic example: the answers to a question by age or by gender. The dedicated tool is **Crosstabs**.

## Crossing two questions

<Steps>
  <Step title="Open Crosstabs">
    Go to Crosstabs from the **Results** tab.
  </Step>

  <Step title="Choose the two questions">
    Select the question to put in the columns (e.g. age) and the one to put in the rows.
  </Step>

  <Step title="Read the table">
    The table crosses the two questions, in **individuals** and in **percentage**, with a legend.
  </Step>
</Steps>

<Tip>
  You can add a **third dimension** to the crosstab by applying a filter (e.g. limiting to men in the Paris region). The base shrinks accordingly.
</Tip>

## Crossing the whole questionnaire by one variable

You can cross **the whole questionnaire** by a single variable, then export it to Excel.

<Steps>
  <Step title="Choose the crosstab variable">
    Only one variable at a time in the columns (e.g. age, or a concept test cell).
  </Step>

  <Step title="Export">
    Launch the Excel export: each question of the questionnaire is crossed by this variable. The file is **sent to you by e-mail**.
  </Step>
</Steps>

<Note>
  You cannot put several variables in the columns in a single export. To cross the questionnaire by **all** the socio-demographic variables at once, use a **Banner**.
</Note>

## The Banner

A **Banner** is a single variable that groups all your socio-demographic variables (gender, age, region…). By crossing the questionnaire by the Banner, you get all the socio-demographics in the columns at once.

The Banner is built from the individual data. See the **Importing a variable** page to create it.

<Warning>
  A Banner created in the Results is **tied to the survey**: it disappears if you duplicate the survey. To reuse it, create it beforehand in the **template**.
</Warning>

## Significance tests

A **significance test** indicates whether the gap between two columns is real or due to the randomness of sampling. By default, significant gaps are **calculated against the overall base** and flagged **in color** on the table (green or pink); the **in pairs** mode displays them with letters instead. The comparison mode is set in the **results preferences**.

| Mode | Comparison | When to use it |
| - | - | - |
| **Compared to Overall** (default) | Each column against the full base. | Standard reading of a crosstab. |
| **Compared in pairs** | All the columns against each other; letters indicate the reference column. | Comparing cells (concept test). |
| **Compared to the previous column** | Each column against the previous one. | Tracking change from one wave to the next — mainly in flat analysis, less relevant on a crosstab. |

<Warning>
  No test is displayed for a group of **fewer than 30 respondents**: the base is too small to draw a conclusion.
</Warning>

<Tip>
  To produce crosstabs in **PowerPoint**, go through the Report step. Crosstabs exports to Excel.
</Tip>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.