Comparisons and Stats Testing

Comparisons:

The Comparisons tool allows the user to create single-variable or multi-variable banners to “cut” the survey results.  To access the Comparison tool, select the “Filters & Comparisons” button located on the top-right of the screen, and then selecting “Compare.”

Comparison variables can be created in two ways:

1. Compare by question: Use questions and response options within the activity to compare results.First, select the question of interest.

A banner will be created for each response option that is selected.  To generate the banner, select “Compare results.”

QUICK TIP: The above option will include responses from those who may not have completed the entire activity.  To create a filter for only those who completed the activity, use the advanced filter option.

2. Compare by advanced filter: Use already created advanced filters to compare results. For more information on how to create an advanced filter, see the “Filtering Results” article. First select, “Compare with advanced filters” under the “Compare” tab of the “Filters & Comparisons” button.

Next, select “Create a new advanced filter comparison” where you can select the advanced filters of interested (that have already been created) and then select “Compare.” You will have an opportunity to “Name” your newly built comparison.

QUICK TIP: Advanced filters used in comparisons should be mutually exclusive and NOT include any overlapping variables. If one or more variables are included in more than one filter in the same comparison, then the stats testing tool will not work properly, i.e. a respondent cannot fall under more than 1 segment that is being used in the comparison. For example, a respondent cannot fall under both “Architect” and “Engineering” when creating a comparison of Architect v. Engineering v. Construction.

The tables will update to include a banner for each of the identified variables. For more information see the “Stats Testing” section below. 

Stats Testing:

After comparisons are applied, each variable (or column) will be assigned a capital letter.  In the example below, the “Male” column is assigned with “A” and the “Female” columns is assigned with “B.”  The statistical differences between the Male and Female segment will be noted using these letters. 

Looking again at the below example, the “B” means that Males are significantly more likely (at a 95% confidence level) than Females to have visited this store in the “past 4 weeks.”

QUICK TIP: Upper-case letters note a 95% confidence level and lower-case letters note a 90% confidence level