Table Of Contents

- Manual
- Getting Started
- Starting the Program
- Retrieving Data
- Manipulating Data
- The Variable List
- The Variable List Menu
- Filter Observations/Selecting
- Add New Variables
- Delete Variables
- Edit Metadata
- Set Replicate Weights
- New Variable Reserve
- Edit Value Labels
- Dummy Code Categorical Variable
- Collapse Categories of Categorical Variable
- Set Missing Values
- The Expression Evaluator

- Saving and Re-running Actions

- Sampling
- Procedures
- Measurement Models
- MML Models for Test Data
- Other Available Procedures

- Graphics
- Tools
- Estimation Methods
- Optimization Techniques
- Variance Estimation

- Post-hoc Procedures
- More user input instructions
- The User Interface
- Input Instructions
- Options
- Output Precision

- Glossary of Terms and Symbols

- Getting Started

Detailed Help Instructions - DRAFT VERSION

*Specifying Independent Variables*

Click once on the original *AM* window and click once on the tab reading Variables. Left-click on the desired variables and drag them into the Independent Variable box in the Analysis dialogue box. .

*Specifying Dependent Variables*

Dependent variables are variables that are thought to be a function of the independent variables. Different procedures take different types and numbers of dependent variables. Please see the "How To" section under the specific procedures to learn about the types of dependent variables analyzed by each type of procedure.

To specify dependent variables, first select the appropriate page of the Variable List. Some procedures take a test or a subtest as a depedent variable (e.g., MML procedures). Others take one or more "regular" variables. You can select the "tests" page by clicking on the "Tests" tab, or the variables page by clicking on the variables tab.

Choose the desired tests or variables by Left-clicking on the desired variables and dragging them into the Dependent Variable box in the Analysis dialogue box.

A future release of the software will incorporate facilities to estimate the item parameters.

*Specifying the Design Variables*

*AM* calculates Taylor-series standard errors. To do so, it must know the strata, the primary sampling units ("cluster"), and the weight. Variables containing these data can be defined using the "Design" keyword in the data dictionary. These items may be changed or eliminated as you choose.

To select a strata variable, click once on the original *AM* window. Left-click the desired variable and drag it into the Strata box in the Analysis dialogue box. (Note: This variable should be listed as "Strata" under Description column and listed as "DESIGN" under the Type column.) To remove a pre-existing strata variable, click within the Strata box and press delete. Follow similar procedures for selecting or deleting Cluster and Weight variables.

*Specifying Cut Scores*

To specify Cut Score values, double-click on the item reading "CutScore 1." A new window will appear and prompt you to enter a Cut Score value. When you are finished click OK. Repeat this process with up to seven values.

*Analysis Window Additional Options*

Three additional options may appear in the lower right hand corner of the Analysis dialogue box.

*Change Starting Values.* The software will calculate decent starting values for you, but you can change these values by clicking in the square box next to this option. Overriding starting values will generally cause performance to deteriorate and could render the model inestimable. However, advanced users who have some other knowledge (say, from other estimates) may improve performance by specifying different starting values. Also, different starting values may be needed to check that the maximum found is global in some problems.

*Use Constant.* Clicking in this box will toggle the check-mark off and remove the constant term from the regression.

*Output.* Choose either Web Browser Output or Spreadsheet Output from the drop-down menu to select the desired format for reviewing the results of the Analysis.

*Advanced Window Additional Options*

Several more options are available through the Advanced dialogue window of many analyses.

*Subtest.* You can work with a given subtest of a larger test by selecting it from the drop-down menu.

*Convergence.* This specifies the change in the likelihood function designated as essentially zero. This default value may be too conservative for very large data sets. You can change the convergence value by clicking in the Convergence box and editing the value.

*Optimization.* This allows you to choose among several methods to estimate a parameter that is only partially observed. You can choose a desired method of optimization by selecting it from the drop-down menu.

*Maximum.* This allows you to indicate the maximum number of iterations allowed for the model to converge.

*Diagnostic Log.* If selected, *AM* will create a log describing any problems encountered while performing the analysis.

*Quadrature Points.* The software program's default creates 21 quadrature points, but you can click in the Quadrature Box and edit the value if you choose.

*Minimum.* This specifies the minimum value in the range covered by the quadrature points. If the item parameters are not based on a normal 0,1 metric, you may choose to change this value by clicking in the Minimum box and editing the value.

*Range.* This specifies the range covered by the quadrature points, so specifying 8 with a minimum of -4 covers the range from -4 to +4. You can change the range by clicking in the Range box and editing the value.

*Subtest Weight.* You can change the weight given to a subtest by clicking in the Subtest Weight box and editing the value.

*Location.*

*Scale.*