IRT Calibration - Reports and Table of Contents
Item Response Theory - Calibration
Item Response Theory - Calibration reports in PARcore visualization studio include descriptive statistics reports and graphs related to the calibration analyses for Item Response Theory. The available reports for a given request is combination of the options as select in the analysis template module and the scores selected in the test settings module. For more information, please see Analysis Templates and Test Settings.
Report Links and Descriptions for Calibration:
- Summary of Data and Iterations - This report includes four different summary reports related to the analysis: a summary of the data used (including number of observations, number of items, etc.), a summary of the iteration progress, the model prediction statistics, and the reliability analysis statistics.
- Calibration Parameter Estimates - This report includes various different reports related to the parameter estimates from the calibration. Included reports are:
- Bivariate Scatterplot: The a-parameter and b-parameter statistics for all items are displayed in a scatterplot.
- Item Summary: Summary statistics for the calibration parameters for all items (dichotomous and polytomous grouped separately). Additional summaries by Form, AU, and/or LIC are available if the option was selected in the request details.
- Problem Item Report: A summary report for each flag value indicating how many items in the analysis were flagged.
- Problem Item Profile Report: This table contains only the items that were flagged in the analysis, which flag each item triggered, and the a-parameter and b-parameter statistics for the item.
- Parameter and Ability Histograms - This report includes histograms for the a-parameter, b-parameter, and ability distributions from calibration. Included reports are:
- Parameter: Histograms of the a-parameters and b-parameters for all items are displayed.
- Ability: A histogram of the ability estimates for all test takers is displayed. Additional histograms by Form are available if the option was selected in the request details.
- Conditional Expectations and Variances of Underlying Latent Vectors Given Predictors - This report includes the mean and covariance for each group (form) included in a concurrent calibration with predictor vectors. This report is not produced and will not appear on the landing page if the option to use predictor vectors for concurrent calibrations was set to "No"
Additional statistics are produced and available to users beyond what is presented in the Data Visualization Studio. Refer to the RTSSTAT API and Download Statistics help documentation for more information regarding accessing all of the available statistics produced.