MarketSight - Variable Basics
The Variables screen allows you to work with variables in your dataset. You can edit uploaded variables, organize variables with categories, and create new User Defined variables. Learn more about user defined variables, importing user defined variables, and weighting. You can also export a pre-filled copy of the dataset labeling workbook from this page.
Watch this video to learn more.
Customize Your List Page
- To change the number of variables shown on the page, adjust the page display size (B)
- To hide or show columns, select the columns to be displayed (D)
- Sort and filter the list by clicking on column headers and arrows (A) or by entering search terms (E)
- Sort columns by clicking on column headers (D)
Export a Pre-Filled Copy of the Dataset Labeling Workbook Using Your Dataset Information
- On the Datasets or Variables page, click Export in the gray toolbar
- Choose "Labeling File"
- Download the completed workbook from the Tasks list on the right side of the application
Hide MRV Predecessors (Prevent MRV Component Variables from Showing on the Variable List Page)
- Go to the Datasets page
- Right click your dataset in the left panel
- Select Properties
- Check the box to Hide MRV Predecent Variables
- Click OK to confirm
Click here to see more information on hiding variable categories.
Variable Types in MarketSight
There are several kinds of variables in a MarketSight dataset. Variables that are uploaded with your dataset can also have different settings. Uploaded variables can be:
Discrete (Nominal and Ordinal)
A discrete variable has values with value labels. A variable such as "Gender" might have the values of 1 and 2 that correspond to the labels Male and Female. A nominal variable has values which have no inherent order to them (for example, a variety of magazine titles). An ordinal variable has values with a specific order (such as a rating scale).
A variable is continuous if its values do not have any value labels. For example, respondents may be asked to estimate the number of minutes they spend watching TV.
Text (string) variables uploaded with your dataset cannot be statistically analyzed and will not be displayed when creating a new user defined variable or a new crosstab. You can see the content of text-based variables by placing them into a data view.
In addition, there are five types of variables that can be created after you upload your dataset to MarketSight:
|User Defined||Variables created within MarketSight|
|Multiple Response||Also known as multi-punch, there variables allow for double counting.|
|Filter||Created within MarketSight to allow for easier exclusion of respondents.|
|Wave||Added during the append process so data from different times can be compared.|
|Weight||Adjusts the numbers based on specified factors. Can be uploaded or created within MarketSight.|
Create a New User Defined Variable
- Click New, or
- Right-click on the original uploaded variable you would like to base your user defined variable on
- Select the type of variable you would like to make
Edit a Variable
- Left click on the variable name
- Modify as needed
- To change the nominal/ordinal setting, click (A)
- Click OK to save your changes
Find and Replace Words in Variable and Value Labels
- Click on Find and Replace (C)
- In the dialog that opens, enter the term to be found and the term to replace it with
- Choose whether or not to search Variable Values, Value Labels, or if the search should be Case Sensitive
- In the list page behind the dialog, only variables containing the desired term will be shown
- Choose Replace All, or Replace Selected
- Select the items to be categorized by clicking the check box to the left of their name
- Click "Categorize" in the menu bar
- If necessary, click New and name the category
- Otherwise, select the desired Category
Click here for more information about variable categorization.
Warning: This step cannot be undone!
- Select the checkbox next to the items to be deleted
- Click Delete
Note that variables in use cannot be deleted.