Introduction
Back coding , also known as up coding, refers to situations where a question contains one or more Other specify options and there is a need to allocate these responses into the original code frame (and, as appropriate, create new responses).
For example, if you ask someone their race and they answer "Other" rather than White, Black, Asian, etc. in some instances, it may be possible to backcode the variable rather than leaving them in the "other" category.
Requirements
- A text variable containing Other (Specify) responses.
- A Pick One or Pick Any question to back code into.
- If the original question is in Pick One - Multi format then right-click the question and use Insert Ready-Made Formula(s) > Pick One - Multi -> Pick Any (by Category) to create a Pick Any version of it.
- If the original question is in Pick Any - Compact format, first change the Question Type to Pick One - Multi and then take the steps above for Pick One - Multi to create a Pick Any version.
Method
Back coding using Q's coding tools
Back coding is done as follows:
- In the Variables and Questions tab select the text variable that you wish to back code (i.e., which contains the other specify responses).
- Right-click on it and select Insert Variable(s) > Code Text > New Code Frame > Manual Categorication (or Semi-Automatic Categorization) > Multiple Overlapping Categories.
- Manual categorization automatically creates an additional category on the right called New Category that we don't need. Before we begin, we will delete that category.
- On the right, right-click New Category and select Delete
Now we will proceed with backcoding the Race - other variable. - At the bottom of the dialog, click the Inputs and Back Coding... button
- In the Categorization Input dialog, from the Corresponding back coding variables menu, select the variable containing the codes you want to use, which in this example is Race and click OK.
You will see the following message: - Click OK.
The original codes are on the left, and the codes you want to reassign them to are on the right.
- Select a category on the left you want to back code, click the category on the right that you want to move it to, and then click the Categorize as button. In this example, we will move Caucasian into the White category.
- Repeat the process until you are reassigned as many values as you can.
- Click Save Categories to save the results to a new variable. In this example, the new variable is named Race - other - Coded
- Switch to the Outputs tab to take a look at the new variable.
Notice that the number of Whites has increased from 1,547 to 1,551.
Back coding using externally coded data
Single response data
- Select the pre-coded variable on the Variables and Questions tab and press
.
- Select Missing Data for the other category and press OK.
- Select the two variables (i.e., the original variable and the variable containing the coded data) in the Variables and Questions tab.
- Right-click and select Insert Ready Made Formula(s) > Merge Questions.
Multiple response data
- If necessary, merge the coded data (see Merge Data Add Variables) and import the revised data.
- Ensure that the original question (i.e., which contains the Other option) has been correctly set up as a Pick Any question.
- If the coded data is in Pick Any - Compact format:
- Set the Question Type to Pick One - Multi.
- Select the variables and select Insert Ready Made Formula(s) > Pick One – Multi - > Pick Any (By Category).
- Select the original question and the coded question (both of which should be Pick Any) questions, right-click and select Insert Ready Made Formula(s) > Merge Questions.
- If the Base n for the categories in the new question is different for some of the options:
- Right-click on a category, select Values and uncheck the Missing Data selection.
- If necessary, filter the question to rebase it.
Next
How to Automatically Code Simple Text Variables
How to Re-Use a Code Frame When Coding Text Data
How to Automatically Code Unstructured Text Data in Q
How to Code Multiple Other Specify Options
Comments
0 comments
Article is closed for comments.