How to Create Top Category Variable(s) (Top 2 Box, Top 3 Box, Top K Box)

This article describes how to create new variables that show the proportion of respondents who selected the top category (top 2, top 3, or top K) options in Nominal, Nominal - Multi, Ordinal, or Ordinal - Multi variable sets. This could include things like combining Agree and Strongly Agree, Like and Love, or Somewhat Agree, Agree, and Strongly Agree.

The table on the left shows a Summary table of a Nominal - Multi question. The table on the right shows a Summary of the Binary - Multi after running the top 2 box feature.

Requirements

• A data set loaded into Displayr that contains at least one Nominal, Nominal - Multi, Ordinal, or Ordinal - Multi variable set.

Displayr identifies the top categories by looking at the numeric Value for each category not selected as Missing Data in the Value Attributes of the variable set and choosing the highest values. To determine the numeric values, select the variable set in the Data Sources tree, then in the object inspector , click Data > Properties > Values.

Your variable set's values will need to be in ascending order to count the "top" attributes:

If your value scale is reversed and the highest value is assigned to the "lowest" attribute (i.e., Hate is assigned a value of 5 instead of 1), then consider using How to Create Bottom Category Variable(s) (Bottom 2 Box, Bottom 3 Box, Bottom K Box) instead.

If, instead, you want to recode your variable, see How to Reverse Scales in Variable Sets.

Method

Top 2 Category Variable (Top 2 Box)

1. Select the variable set(s) in the Data Sources tree. If multiple variable sets are selected, a new top 2 box variable will be created for each.
2. Click + > Convert To > Binary - Multi > Top 2 Categories.
3. When appropriate, you will be asked if you want to remove categories that look like Don't know options. This is equivalent to setting the category as Missing Data.

Displayr will then create a new variable for the top 2 categories for the selected variable set(s). The Value Attributes may look something similar to the following:

Top 3 Category Variable (Top 3 Box)

1. Select the variable set(s) in the Data Sources tree. If multiple variable sets are selected, a new top 3 box variable will be created for each.
2. Click + > Convert To > Binary - Multi > Top 3 Categories.
3. When appropriate, you will be asked if you want to remove categories that look like Don't know options. This is equivalent to setting the category as Missing Data.

Top K Category Variable (Top K Box)

1. Select the variable set(s) in the Data Sources tree. If multiple variable sets are selected, a new top 3 box variable will be created for each.
2. Click + > Convert To > Binary - Multi > Top K Categories.
3. Displayr will ask you how many categories you wish to combine. For example, to combine the top two categories, type '4' and press OK.
4. When appropriate, you will be asked if you want to remove categories that look like Don't know options. This is equivalent to setting the category as Missing Data.
5. Displayr will then create new variables for the top K categories for the selected variable set(s).

Notes

If you want to create a top box NET category on a table, in addition to retaining the original scale, as seen in the table below, see the "Adding a new category" section of How to Recode into Existing or New Variables.

Some users might prefer to restructure a nominal - multi or ordinal - multi variable set into a binary - multi and manually indicate which values to count. Use this method if you are only interested in aggregating one aspect of the question, such as one specific code, one combination of codes, etc. See How to Merge or NET Categories on a Nominal - Multi for more information.

Next

How to Create Bottom Category Variable(s) (Bottom 2 Box, Bottom 3 Box, Bottom K Box)

How to Reverse Scales in Variable Sets

How to Recode into Existing or New Variables

How to Merge or NET Categories on a Nominal - Multi

Manipulating Data video