> ## 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.

# Comparing surveys or questions

> Comparing waves, surveys or questions with each other.

The platform allows two types of comparison:

* **Survey comparison**: compare several surveys with each other, typically the waves of a tracker.
* **Questions comparison**: compare similar questions within the same survey.

## Comparing surveys or waves

<Steps>
  <Step title="Select the surveys">
    Choose the surveys to compare. Add more via the **+** (e.g. wave 1, wave 2, wave 3).
  </Step>

  <Step title="Order the display">
    Set the order of the surveys, from left to right.
  </Step>

  <Step title="Rename the waves">
    Replace each survey's technical name with a clean label (e.g. "Wave 1").
  </Step>

  <Step title="Choose the display">
    Stacked bars, lines, vertical bars per wave or grouped horizontal bars; in individuals or in percentage.
  </Step>

  <Step title="Apply a filter">
    Restrict to a subtarget if needed (e.g. the under-35s).
  </Step>
</Steps>

### Question matching

A tab summarizes all the questions per survey, the number of variables the matching is based on, and flags the labels that differ.

The rules to know:

* **Questions** are matched on the **variable name**, whatever their position in the questionnaire.
* Within a question, **items** are matched on their **label**.
* Adding an item is not a problem: the other items keep comparing, whatever their position in the question — and whatever the position of the question.

<Warning>
  If an item's label changes (e.g. "Brand 1" → "Brand 1 B"), it is no longer recognized: it is displayed on the wave where it exists, but stays empty on the others. To change a label while keeping the comparison, use the **translation system**: it displays the new label to the respondent but keeps the old one for the matching.
</Warning>

### Filters across waves

A filter only works on a wave if it is present in that wave's preferences **and** carries the **same variable name** on all the waves.

### The comparison settings

The comparison has its own settings, independent of the flat analysis:

* Optional total column.
* Its own significance test (e.g. comparing each wave to the previous one).
* Basis for displaying results (questionnaire flow or total base).

## Comparing similar questions

Within the same survey, you can compare questions of the same nature (e.g. all the appeal questions, or the purchase intent ones).

* **Manually**: select the questions to compare.
* **Automatically**: give the **same variable name** to all the similar questions. The platform groups them and compares them in a few clicks.

<Tip>
  The shared variable name method is ideal for concept tests with several loops: all the loops are compared at once.
</Tip>

Export is possible in Excel and in PowerPoint.

<Card title="Merge the results of several surveys" icon="merge" href="/en/analyse-des-resultats/regrouper-des-etudes">
  To stack the respondents of several surveys into a single base, rather than place them side by side.
</Card>


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