Sample distributions are used to further refine and target panelists to match quota groups within surveys and/or to balance them by specific demographics. The Distribution Manager has two main purposes:
- Downsize a sample to achieve a specific number of completions without oversampling your panel. This calculation is configured on the final distribution matrix page, and utilizes the number of completes desired and the expected response rate (automatically pulled in based on your sample, but is modifiable if you expect a different response rate).
- Weight invitations to a sample before they are sent to achieve a specific balance of demographics or panel-related attributes for your survey.
Distributions are generally not used or are not achievable within small panels (other than for downsizing samples based on a desired # of completions), and if you wish to send invitations to your entire panel you can bypass this step. Once designed, distributions can be saved and applied against multiple projects. For larger panels, it is often customary to balance invitations across demographics such as age, gender, income, education, and ethnicity. The Distribution manager is accessible from the Project functions menu.
1: Using the Distribution Manager
First give your distribution a name and, if desired, a description. Select the datapoints that you wish to base your distribution on by dragging them from the box on the left-hand side of the screen to the box on the right-hand side of the screen. Distributions can only be created from single or multi-select datapoints.
2: Nested Distributions
An example of how Nested distributions work would be as follows:
The breakdown of the sample is as follows:
|Apples and Red||30%|
|Apples and Blue||20%|
|Oranges and Red||24%|
|Oranges and Blue||16%|
|Bananas and Red||6%|
|Bananas and Blue||4%|
3: Non-Nested Distributions
An example of how Non-nested distributions work would be as follows:
There will be 60% of red and 40% of blue for the entire sample, but this breakdown will not also be reflected within each fruit category. The resulting sample will contain 50% apples, of which any amount may have red or blue as a favorite color.
Once you are satisfied with the final distribution, save it as a sample and return to the Sample Manager to confirm creation (it will appear as a sub-sample).