Heroic Features - Start Bootstrap Template


METRICS (Methods for Equating, Testing, Regression, Item response theory, Classification, and Scoring) is a web-based environment for statistical and psychometric computing. This application includes an array of methods useful for psychometric analysis, interactive visual displays, and regression models using observed data impacted by measurement error.

Comprehensive Set of Tools

METRICS provides a large set of tools for data analysis in one place. METRICS includes an engine for IRT test scoring, six different test equating models, powerful, interactive visual displays, error-in-variable regression models including linear, mixed linear, and direct estimation regression using marginal maximum likelihood.

Powerful and Fast

METRICS is built entirely with R functions and uses parallel processing to improve computational speed. This means that large data problems can be handled quickly and efficiently. Scoring thousands of tests can be done in minutes as opposed to hours!

Unique Statistical Methods

METRICS implements unique statistical models not available in any other software program. For example, the regression features allow users to run models that account for measurement error in the conditioning variables. Furthermore, the direct estimation methods can yield population estimates when the dependent variable is only partially measured.

Easy To Use

METRICS is a fully developed web application, so it requires no programming or installation of any software even though it uses R in the background. This means you can access METRICS from any location with internet access and simply use the point and click features available in the web browser to analyze your data.

Save Data

Export Graphic

Data Tools

Points near clicked area on plot

To produce a fit plot for items, please choose three files. Once these three files are read in, simply use the drop down menu to choose and plot different items. Choose a file with the theta scores. The theta variable can be an MLE, MAP, EAP, but it must be on the same scale as the item parameters.
Check box if datafile has variable names in first row
Choose the data file holding the item responses.
Choose the file holding the item parameters (upload shell).
Check box if datafile has variable names in first row

Export ICC Plot

Export Q3 Statistic

Fit Plot

Q1 Fit Statistic

Q3 Statistic

Item Parameters

Test Summary Plots

Read in a .csv file holding the ITS IDs of the items you wish to include

Export Test Plot

Score Conversion Table

Select the parameters for the score conversion table
Export Score Conversion Table

Test Summary Plot

Marginal Reliability

Score Conversion Table

Run IRT Equating

Impact Data Parameters
Population parameters for the Stocking-Lord and Haebara Methods
Step values, if they exists for polytomous items, can either be treated as a vector and averaged or the mean over the steps within an item can be computed and the mean can then be treated as one item
Remove each linking item iteratively and estimate linking constants Remove items based on the D2 fit statistic or Difference in b-parameters
The linking constants in the summary have been applied to the item parameters. Click export below to export them to a .csv file Export Rescaled Item Parameters

Equating Summary



Check box if datafile has variable names in first row

Make sure minimum items under item treatment has a value. Then, enter item parameters below as either a single value or as a vector separated by a comma.

Choose which modules to be run

Click to save all settings for this test

Score Summary


Scoring Diagnostics


Check box if datafile has variable names in first row

Classification Result


Check box if datafile has variable names in first row

You can choose from the following to export to Excel
Export Regression Object
You can export Empirical Bayes (BLUPs) from Level 2 or Higher
Export Empirical Bayes

Regression Summary


Regression Diagnostics

Choose the level to plot

Direct Estimation Summary