A key piece of consumer research analysis is understanding when responses are statistically significant over other options. KnowledgeHound makes the application and reading of this information quick and simple.

First, a few notes about the testing that KnowledgeHound is doing.

  1. The test we run is a Z-test. The following is the official definition of a Z-test from the Insights Association.

  2. Stat testing is conducted at the 95% confidence level. For more information on confidence levels see this article by Jim Frost.

  3. Stat testing only runs when you have a breakout applied. For example, you are looking at the results from a question and you select to break those results out by Age. Once that breakout by Age is applied KnowledgeHound is able to run stat testing.

There are two ways to view stat testing on an analysis page. 

  1. Automatically on the spreadsheet view.

  2. Select "View Data Labels" when you have chosen a visual chart type.

To read the stat testing, use the letters to understand the comparisons. 

Spreadsheet View - Hover over the cell to view where significance exists. In this example 50% of the General Population responded "10-14" and that is significant over both the Total Population (represented by a) and the Millennial population (represented by b) that responded the same way. 

Chart View(s) - To include stat testing information on charts, select View Options and View Data Labels to apply. To read the stat testing, use the letters next to each bar and the legend at the bottom. In this example, we continue to see that 50% of the General Population responded "10-14" and that is significant over both the Total Population (represented by a) and the Millennial population (represented by b) that responded the same way. 

Did this answer your question?