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

# Choose chart metrics and statistics

> Choose what the bars measure, add statistical summaries, and compare compatible questions.

A chart's **Metric** chooses what its bars represent. **Statistics** adds a table beneath those bars. Use them separately: a chart of answer percentages can carry its mean below it, while a chart of means draws the mean itself.

## Choose the metric

Open a chart in **Charts**, select its question, and use **Metric** below it. The available choices depend on the answer structure and scores.

| Question | Available statistical measures |
| - | - |
| Numeric | Mean, median, standard deviation, minimum, maximum and sum. |
| Scored categorical scale | Mean, median, standard deviation, minimum and maximum; no sum of scores. |
| Numeric matrix | Measures per row; the table also reads its numeric cells in bands. |
| Allocation or Multi-Numeric | Measures per item. |
| Multiple-select | Mean or median number of mentions. |
| Categorical matrix | Mean score per row, or percentage on a scale answer or net. |

A categorical mean reads the options' numeric scores, or the supported numeric prefixes of scale labels. Inspect these scores before interpreting a mean. Use [question settings](https://support.deepfield-ai.com/building-your-study/question-types.md) to review recodes.

## Add the statistics table

Open **Statistics** and tick the required summaries. A standalone chart starts with no statistics selected. The table follows the chart's base, filters, weighting and groups, and appears in its Excel download and copied data. Mean comparisons can carry significance letters; check [the statistical rules](https://support.deepfield-ai.com/analysis/reading-significance.md) before using them.

For a question in an analysis plan, the split's statistics override the plan's choices. Without an override, the plan's statistics appear beneath its chart and in its table. A request for a statistic alone can create a chart of that measure.

## Compare several questions

Use **+ Add question** to add compatible questions to the same chart. With answer distributions, their answers appear side by side; with a statistical metric, each question supplies that measure. For example, compare a scored purchase-intent question asked at several prices.

The picker excludes questions that cannot supply the current metric. Changing the metric can remove incompatible questions and statistics. Review the selection again before saving. A percentage on one answer needs that answer to remain shown in each question; a net must exist in each question's own answer edits.

## Verify the result

Check the heading names the selected measure, inspect the base for every group or item, and compare the summary with the underlying answers. A mean is not a percentage. Review counts and weighted bases separately from percentages, then save and reopen the chart.

## Related

* [Choose a chart type](https://support.deepfield-ai.com/analysis/chart-types.md)
* [Build and organize charts](https://support.deepfield-ai.com/analysis/charts-library.md)
* [Build a crosstab](https://support.deepfield-ai.com/analysis/crosstabs.md)
* [Export your results](https://support.deepfield-ai.com/analysis/exports.md)


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