Dr. John Willse, Associate Professor in ERM, has developed the computer program MGR “Measurement GUI for R”. MGR is a point-and-click interface (programmed in JAVA) that allows the user to make use of measurement and item response theory (IRT) analyses conducted by R packages without having to engage the R language. Thus, rather than having to learn the R language to make use of the IRT-related analyses possible with R, users employ the user-friendly MGR software package to run analyses. When the user selects a particular analysis from MGR’s point-and-click menu format, MGR calls the relevant analysis from R (i.e., MGR composes the R code that is then read into R by MGR) and then prints out the results in user-friendly output tables (i.e., MGR reads the output of R and reformats this output into user-friendly tables). Over time MGR will have the potential to run any measurement analysis contained in R’s ever-growing library of analyses.
MGR has several advantages over currently available IRT and measurement software programs. First, MGR is free (as is R), and thus it is available to all individuals at no charge. This allows the application of the analyses supported by MGR to be available to everyone. Second, because MGR uses a point-and-click environment, it will be very popular with a wide range of applied constituencies, such as applied measurement specialists, researchers, and students. Third, because MGR has the potential to run all analyses supported by R, it will be possible for MGR to have extensive capability over time. Given the speed at which R packages are being generated for analyses related to measurement, the capabilities of MGR will continue to expand in the future.
The possible capacity of MGR as an interface for conducting measurement analyses is extensive. Not only does MGR have the potential to become a widely used interface for conducting traditional measurement analyses (e.g., classical test theory, dichotomous and polytomous IRT model calibration), but it also can be extended to include newer modeling techniques (e.g., diagnostic classification models, mixture models, multidimensional item response theory models, etc.).