For the latest version of JMP Help, visit JMP.com/help.


Basic Analysis > Distributions > Options for Categorical Variables
Publication date: 04/21/2023

Options for Categorical Variables

The red triangle menus next to each variable in the Distributions report contain additional options that apply to the variable. This section describes the options that are available for categorical (nominal or ordinal) variables.

Display Options

Contains the following options to change the display of the report.

Frequencies

Shows or hides the Frequencies report. See The Frequencies Report.

Horizontal Layout

Changes the orientation of the histogram and the reports to vertical or horizontal.

Axes on Left

(Available only when the Horizontal Layout option is selected.) Moves the Count, Prob, and Density axes to the left side of the histogram instead of the right side.

Histogram Options

See Histogram Options for Categorical Variables.

Mosaic Plot

Shows or hides a mosaic bar chart for each nominal or ordinal response variable. A mosaic plot is a stacked bar chart where each segment is proportional to its group’s frequency count.

Order By

Reorders the histogram, mosaic plot, and Frequencies report in ascending or descending order, by count. To save the new order as a column property, use the Save > Value Ordering option.

Test Probabilities

Launches a report that enables you to test hypothesized probabilities. The null hypothesis is that the true probabilities in each level of the response are equivalent to the hypothesized probabilities. The alternative hypothesis is that at least two of the true probabilities differ from the hypothesized probabilities. The options that appear in the Test Probabilities report depend on the number of levels of the response variable.

For response variables with two levels, you have the option to choose between a two-sided chi-square test or a one-sided exact binomial test. For either test, you can specify hypothesized probabilities for one or both levels of the response. If you enter both hypothesized probabilities and they do not sum to 1, they are rescaled to sum to 1 for the test. For the one-sided test, you can choose between a test that the observed probability is greater than the hypothesized probability or that the observed probability is less than the hypothesized probability. If you specify both hypothesized probabilities for a one-sided test, the test is performed for the second level of the response. See Example of Testing Probabilities for Two Levels.

For response variables with more than two levels, you can enter one or more of the hypothesized probabilities. If you enter hypothesized probabilities for all of the levels of the response and they do not sum to 1, they are rescaled to sum to 1 for the test. If you do not specify hypothesized probabilities for all of the levels of the response, there are two rescaling methods from which to choose. You can either fix the unspecified probabilities at their observed values and rescale the specified hypothesized probabilities, or you can fix the specified hypothesized probabilities and rescale the unspecified probabilities. See Example of Testing Probabilities for More Than Two Levels.

Confidence Interval

Computes score confidence intervals for the probabilities in each level of the response. You can choose one of the confidence levels provided or select Other to specify your own confidence level. See Statistical Details for Score Confidence Intervals.

Save

See Save Options for Categorical Variables.

Remove

Permanently removes the variable and all its reports from the Distribution report.

Histogram Options for Categorical Variables

Histogram

Shows or hides the histogram. See Histograms.

Vertical

Changes the orientation of the histogram from a vertical to a horizontal orientation.

Std Error Bars

Draws the standard error bar on each level of the histogram.

Separate Bars

Separates the histogram bars.

Histogram Color

Changes the color of the histogram bars.

Count Axis

Adds an axis that shows the frequency of column values represented by the histogram bars.

Prob Axis

Adds an axis that shows the proportion of column values represented by histogram bars.

Density Axis

Adds an axis that shows the length of the bars in the histogram.

The count and probability axes are based on the following calculations:

prob = (bar width)*density

count = (bar width)*density*(total count)

Show Percents

Labels the percent of column values represented by each histogram bar.

Note: To specify the number of decimal places, right-click the histogram and select Customize > Histogram.

Show Counts

Labels the frequency of column values represented by each histogram bar.

Save Options for Categorical Variables

Level Numbers

Creates a new column in the data table called Level <colname>. The level number of each observation corresponds to the histogram bar that contains the observation.

Value Ordering

(Use with the Order By option) Creates a new value ordering column property in the data table, reflecting the new order.

Script to Log

Displays the script commands to generate the current report in the log window. Select View > Log to see the log window.

Want more information? Have questions? Get answers in the JMP User Community (community.jmp.com).