Reminder: Assigning questions as weights is done in the Import Edit Mode.
Working with weights is allowed after a dataset has been processed through Data Import.
After a dataset completes the import process, a summary will be displayed.
The data import process in Analyze attempts to automatically detect any variables that should be used as weights.
If any weights are detected automatically, they will be noted in the summary message.
In this example, 0 weights were automatically detected.
Note:
If a weight variable is included in the original data file, but they were not detected automatically when imported into Analyze,
it could be the result of the weight variable not having enough documentation or metadata associated with it.
Assigning any question in the dataset as a weight can be done in Edit Mode.
After the dataset is opened in Analyze, expand the 3-dot menu.
Click on the Edit Questions & Sections option and select the dataset name to edit.
Locate the questions in the dataset that should be assigned as weights.
Depending on how the original data file was setup, the weight variables might have been placed next to each other or they might be scattered throughout the dataset.
The order of the questions in the imported initial dataset in Analyze reflects the order of the variables from the original data file.
There can be more than one weight variable included in a data file.
Best Practice Tip:
Create a new section that only includes the questions that will be used as weights.
This will help with organizing and navigating the list of questions when editing the dataset.
After all the questions to be used as weights have been organized into the same section, confirm that the questions are numeric.
The example above shows the 3 questions to be used as weights (Sales Weight, Post Weight 1, and Post Weight 2) as category-type questions, as indicated by the icon with the 3 circles.
Before assigning these questions as weights, their question type will need to be changed to numeric.
Select a question intended to be used as a weight.
Under the Question Summary section, change the setting for Question Type from category to Numeric.
To rename the question, enter a different name in the Question Label field.
In this example, 'Sales Weight' is used as the question label.
After a question is changed from category to numeric, an additional option, Use As Weight will appear at the bottom menu.
Click the Use As Weight button.
When a weight is the first one to be assigned in a dataset, there will be an option to assign that weight as the primary weight or an alternative weight.
The primary weight is the default weight.
Only one primary weight is allowed in a dataset.
The primary weight will be applied to all data runs in Analyze, unless the user specifies the use of an alternative weight.
Any alternative weights will be available to use when running analysis on the dataset.
There can be multiple alternative weights in a dataset.
When running data in Analyze with an alternative weight, the user will have to specify which alternative weight to use instead of the default, primary weight.
Click the Use As Primary Weight option to assign the weight as the default, primary weight.
After assigning a weight as the primary, the question label will change to 'Primary Weight'.
A scale icon will identify all weight questions. (Post Weight 1 and Post Weight 2 still show the category icon because they have not been converted to weights yet.)
A scale icon with a star identifies that weight question as the primary weight.
In this example, 'Sales Weight' is assigned as the primary, so the question label was changed from 'Sales Weight' to 'Primary Weight' and a star was added to the scale icon.
Proceed with defining the other questions as weights.
Change the Question Type from category to numeric.
The icon will change to indicate a numeric question.
Optional step: Edit the question label by updating text in the Question Label field.
In this example, 'Post Weight' will be changed to 'Shop Weight'.
Click the Use As Weight Button.
Since there is already a primary weight assigned to the dataset, that option is now disabled.
Assign any remaining weights as alternative.
Unlike a primary weight, the name for an alternative weight is used from the text entered in the Question Label field.
(example of assigned primary and alternative weights)
Additional Notes
Changing a Weight Back to a Question
To undo a question assigned as a weight, first select the weight.
At the bottom of the Question Summary, a Use As Question option will be available.
Clicking this button removes the weight assignment and changes it back to a question.
No Assigned Primary Weight
If a dataset includes weights, but they are all alternative and no primary weight is assigned, a default weight of "1" will be used as the primary weight.
This is equivalent to using un-weighted data.
The alternative weights will still be selectable as options in the dataset, but by default any data runs will be run un-weighted.
