Value Attributes is a set of metadata relating to individual variables. The Value Attributes are modified by either changing Variable Set Structure, or, by clicking on a Variable or Variable Set in the Data Tree and selecting one of the following buttons in the DATA VALUES section of the Object Inspector:
- Labels. This is used to change the name displayed for each category.
- Select Categories. This is used when defining which categories are combined in a Binary - Multi or Binary - Grid Variable Set.
- Values. This is used to change the values of a variable used when numeric summaries, such as means and standard deviations, are computed.
- Missing Values. Deals with how specific categories in the data are included or excluded from calculations.
- Delete observations. Removes, or restores, rows from the Data Set.
- Revert. This resets data to the form it had when it was initially imported or created.
Editing value attributes with another program
It is also possible to edit value attributes using a spreadsheet editor such as Microsoft Excel. This is particularly helpful if you have a lot of things to change.
- In the DATA VALUES section of the Object Inspector select Labels, Select Categories, Values or Missing Values.
- Press Export and copy the value attributes into a spreadsheet.
- After editing the value attributes, press Import and paste the value attributes table from a spreadsheet into the text box. Note that the value attributes should be in a tab-separated-values format (this will typically be done automatically by the spreadsheet editor).
- Press OK twice to apply the changes
Inconsistent Value Attributes
When the value attributes of multiple Variables do not agree with one another (e.g. in a Variable Set), they are inconsistent. Inconsistent Value Attributes are displayed with a red background in the Value Attributes Dialog, or as a checkbox which is neither checked nor unchecked. When the source labels of a value are inconsistent, hovering the mouse over the label field shows a message stating "Source labels vary across variables" instead of the source label.