# Statistical Analysis

### What is significance testing?

The purpose of performing a significance test can vary but the most common purpose is to identify differences between subgroups or subgroup’s results compared to the totals in the survey. 

There are many different types of significance testing. PPTX Builder uses `Z-Test` on row/column proportions. 

It can also for instance be performed between brands, target groups and time periods. To enable significance test in a chart, select Statistical Analysis on the right hand pane and enable significance test, as shown in the image below

![](https://135336476-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-MHVX6v0D91V0QshNt1F%2F-MJRk7v--OE_Sci5QlOO%2F-MJRm1-LA7vNwPBObCF9%2Fimage.png?alt=media\&token=5a731e5d-9690-49e0-bbd5-38bb41970ae8)

{% hint style="info" %}
If you want to perform significance testing on your data you'll need to purchase the `sigtesting` feature.
{% endhint %}

### Base limit

Set the base size limit – no test will be performed if the base size is too low.

### Test Between

Decide if you want to perform the test on Columns or Rows.

### Analysis Types

Choose from 4 significance analysis types:

1. **All vs All:** Tests every group against the other groups
2. **Control Group:** Tests each of the test groups against the control group
3. **Next Data Point:** Tests each group against the next group
4. **Previous Data Point:** Tests each group against the previous group

### Significance Confidence

As with everything else, we can never be 100% sure of anything. Confidence levels are a way of determining just how “sure” we are that we are making a difference. The higher the confidence level we set, the more positive we are that these differences are worth paying attention to. 

Choose from 3 level of confidence:

* 90% confidence level, there is a 10% chance that the difference we are seeing is just the result of “noise” in the numbers.
* 95% confidence level, there is a 5% chance that the difference we are seeing is just the result of “noise” in the numbers.
* 99% confidence level, there is a 1% chance that the difference we are seeing is just the result of “noise” in the numbers.