Publication date: 04/12/2021

Fit Model Launch Window Elements

The following Fit Model launch window elements are common to most personalities:

Model Specification

The Model Specification menu contains the following options:

Center Polynomials

Centers the effects when polynomials are included in the model.

Informative Missing

Creates a coding system that accommodates missing values in effects.

Set Alpha Level

Sets the alpha level for confidence intervals in the model reports.

Save to Data Table

Saves the model specifications to a script in the current data table.

Save to Script Window

Saves the model specifications to a script window.

Create SAS Job

Saves SAS code for the current model specification to a SAS Program window.

Submit to SAS

Submits SAS code for the current model specification to a SAS session.

Convergence Settings

Contains options for setting the maximum iterations and convergence limit used in the convergence of models in some personalities.

For more information about any of these options, see Model Specification Options.

Select Columns

Lists the unexcluded columns in the current data table.


Identifies one or more response variables (the dependent variables) for the model.

Note: Response columns with vector values are not supported in Fit Model.


Identifies a column whose values assign a weight to each row for the analysis. See Weight.


Identifies a column whose values assign a frequency to each row for the analysis. In general terms, the effect of a frequency column is to expand the data table, so that any row with integer frequency k is expanded to k rows. You are allowed to specify fractional frequencies. See Frequency.

Image shown hereValidation

In JMP Pro, for some personalities, you can enter a Validation column. See the appropriate Personality chapter for more information about how each personality handles a Validation column. If you click the Validation button with no columns selected in the Select Columns list, you can add a validation column to your data table. For more information about the Make Validation Column utility, see Make Validation Column in Predictive and Specialized Modeling. For more information about how a Validation column is used in JMP modeling platforms, see Validation in JMP Modeling in Predictive and Specialized Modeling.


Identifies a column that creates a report consisting of separate analyses for each level of the variable. If more than one By variable is assigned, a separate analysis is produced for each possible combination of the levels of the By variables.


Adds effects to the model. See Add.


Creates interaction and polynomial effects by crossing two or more variables. See Cross.


Creates nested effects. See Nest.


Generates effects for commonly used models. See Macros.


Applies the specified degree to models with factorial or polynomial effects generated using Macros. See Factorial to Degree and Polynomial to Degree in Macros.


Applies attributes to model effects. These attributes determine how the effects are treated. See Attributes.


Transforms selected continuous effects or Y columns. See Transform.

No Intercept

Excludes the intercept term from the model.


Specifies the fitting methodology. See Fit Model Launch Window Elements. Different options appear depending on the personality that you select.

Target Level

(Available only in certain personalities and when Y is binary and has a nominal modeling type.) Specifies the level whose probability you want to model. The default value is the higher of the two levels based on the order of the levels.


Takes you to Help topics for the Fit Model launch window.


Populates the launch window with the last model specification that you ran.


Removes the selected variable from the assigned role. Alternatively, double-click the effect or select the effect and press Backspace.


Generates the report window for the specified model and personality.

Keep dialog open

Keeps the launch window open after you run the analysis, enabling you to alter and re-run the analysis at any time.


Frequency variables, entered in the Freq text box, are supported in most Fit Model personalities. In general, a frequency is interpreted in the following manner. Suppose that a row has a frequency f. Then the computed results are identical to those for a data table containing f copies of that row, each having a frequency of one.

Rows with zero or missing frequency values are excluded from analyses. Rows with negative frequency values are permitted only for censored observations, otherwise they are excluded from analyses. When used with censored observations, negative frequency values can be used to fit truncated distributions.

Frequency values do not need to be integers. The technical details describing how frequency columns, including those with non-integer values, are handled are given in Frequencies.


Weight variables can be useful in situations where there are observations with different variances. For example, this can happen when one performs regression modeling on data where each row consists of pre-summarized means. Here, rows involving a larger number of observations (smaller variance) should contribute more heavily to the loss function than rows involving a smaller number of observations (larger variance). You can ensure that this occurs by using appropriately defined weights.

Weight variables are supported in many Fit Model personalities. Each personality that supports weight variables uses one of the following methods:

Variance Scaling

Frequency Symmetry

Variance Scaling

When estimation is performed using least squares or normal theory maximum likelihood, the weight w for a given row scales that row’s contribution to the loss function by w-1/2.

Weight variables have an impact on estimates and standard errors. However, unlike frequency variables, they do not affect the degrees of freedom used in hypothesis tests.

Rows with negative or zero values for Weight are excluded from analyses.

Frequency Symmetry

In the Nominal Logistic and Ordinal Logistic personalities, weight variables are handled as if they were frequency variables. If weight and frequency variables are both specified, these personalities handle each observation as if it has a frequency value equal to the product of the weight and frequency values.

Want more information? Have questions? Get answers in the JMP User Community (