Launch the Standard Least Squares Personality

Standard least squares is one of several analytic techniques that you can select in the Fit Model launch window.

This section describes how you select standard least squares as your fitting methodology in the Fit Model launch window. Options that are specific to this selection are also covered.

Fit Model Launch Window

You can specify models with both fixed and random effects in the Fit Model launch window. The options differ based on the nature of the model that you specify.

Fixed Effects Only

To fit models using the standard least squares personality, select Analyze > Fit Model and then select Standard Least Squares from the Personality list. When you enter one or more continuous variables in the Y list, the Personality defaults to Standard Least Squares. Note, however, that other selections are available for continuous Y variables. When you specify only fixed effects for a Standard Least Squares fit, the Fit Model launch window appears as shown in Fit Model Launch Window for a Fixed Effects Model. This example illustrates the launch window using the Big Class.jmp sample data table.

Fit Model Launch Window for a Fixed Effects Model

When the Standard Least Squares personality is selected in the Personality list, an Emphasis option also appears. Emphasis options control the reports that are provided in the initial report window. Based on the model effects that are included, JMP infers which reports you are likely to want. However, any report not shown as part of the initial report can be shown by selecting the appropriate option from the default report’s red triangle menu.

For details about reports that are available for each Emphasis option, see Emphasis Options for Standard Least Squares.

Random Effects

If the specified model contains one or more random effects, then additional options become available in the Fit Model launch window. Consider the Machine.jmp sample data table. Each of six randomly chosen workers performs work at each of three machines and their output is rated. You are interested in estimating the variation in ratings across the workforce, rather than in determining whether these six specific workers’ ratings differ. You need to treat person and machine*person as random effects when you specify the model.

The Fit Model launch window for this model is shown in Fit Model Launch Window for a Model Containing a Random Effect. When the Random Effect attribute is applied to person, a Method option and two options relating to variance components appear in the Fit Model Launch window.

Fit Model Launch Window for a Model Containing a Random Effect

Standard Least Squares Options in the Fit Model Launch Window

Fit Model Launch Window Options for Standard Least Squares Personality describes options appearing in the Fit Model launch window that are specific to the Standard Least Squares personality.

Fit Model Launch Window Options for Standard Least Squares Personality

Emphasis

Controls the types of reports and plots that appear in the initial report window. See Emphasis Options.

Method

Only appears when random effects are specified. Estimates the model using one of these methods:

•	REML. See REML Variance Component Estimates.

•	EMS. Expected Mean Squares, also called the Method of Moments. See EMS (Traditional) Model Fit Reports.

Unbounded Variance Components

Selecting Unbounded Variance Components allows variance component estimates to be negative. This option is selected by default. This option should remain selected if you are interested in fixed effects, since bounding the variance estimates at zero leads to bias in the tests for fixed effects. See Negative Variances.

Note: Only appears when REML is selected as the Method.

Estimate Only Variance Components

Provides a report that shows only variance component estimates. See Estimate Only Variance Components.

Note: Only appears when REML is selected as the Method.

Emphasis Options for Standard Least Squares

The three options in the Emphasis list control the types of plots and reports that you see as part of the initial report for the Standard Least Squares personality (Emphasis Options). JMP chooses a default emphasis based on the number of rows in the data table, the number of effects entered in the Construct Model Effects list, and the attributes applied to effects. You can change this choice of emphasis based on your needs. For details about how JMP chooses the emphasis, see Emphasis Rules.

After the initial report opens, you can add other reports and plots from the platform’s red triangle menu.

Emphasis Options
Effect Leverage	Shows leverage and residual plots, as well as reports with details about the model fit. This option is useful when your main focus is model fitting. Effect Leverage is the most comprehensive option. This emphasis divides reports into those that relate to the Whole Model and those that relate to individual model effects. The Whole Model reports are in the left corner of the report window under the Whole Model title, with effect reports to the right.
Effect Screening	Displays a sorted or scaled parameter estimates report along with a graph (when appropriate), the Prediction Profiler, and reports with details about the model fit. This option is useful when you have many effects and your initial focus is to discover which effects are active, as in screening designs.
Minimal Report	Displays only the regression plot and reports with details about the model fit. This Emphasis is the default when the Random Effect attribute is applied to any model effect. This option is the least detailed and most concise. You can request reports of specific interest to you from the red triangle menus.

Missing Values

By default, rows that have missing values for Y or any model effects are excluded from the analysis.

Note: JMP provides an Informative Missing option in the Fit Model window under Model Specification. Informative Missing enables you to fit models using rows where model effects are missing. See Informative Missing in Model Specification for details.

When your model contains a random effect, Y values are fit separately by default. The individual reports appear in the Fit Group report.

Suppose that your model contains only fixed effects. You enter more than one Y response, and some of these Y variables have missing values. JMP prompts you to select one of the following options:

•	Fit Separately fits each Y using all rows that are nonmissing for that particular Y.

•	Fit Together fits each Y uses only those rows that are nonmissing for all of the Y variables.

When you select Fit Separately, a Fit Group report organizes the individual reports for the Y variables. You can select profilers from the Fit Group red triangle menu to view all the Y variables in the same profiler. Alternatively, you can select a profiler from an individual Y variable report to view only that variable in the profiler.

When you select Fit Together, a Fit Least Squares report organizes the individual reports for each of the Y variables. However, all Y variables appear in the prediction, contour, and surface profilers.