Expanding the Functionality and Accessibility of Software for Diagnostic Measurement

The expansion of package functionality during the first phase of the project will feature an iterative design process, whereby expert and stakeholder feedback is sought at each stage of development. Feedback will inform both changes to existing features and the next steps of development. The package will follow tidy design principles to ensure that the package is intuitive and user-friendly. Additionally, unit tests will be included to confirm that the methods are properly implemented and safeguard against unintended changes in the future. Finally, the package will be hosted in a remote repository (GitHub) with continuous integration so that all checks and unit tests are run each time the package is updated.

User Testing: The project team will conduct multiple rounds of user testing. First, researchers and practitioners will be able to submit feedback via GitHub issues that could include bug reports and requests for additional features that would be helpful in their applied work. Second, the project advisory committee has agreed provide feedback on their own experience using the package and that of their students when applicable. Finally, training sessions will be conducted at professional conferences with a specific focus on applications of DCMs to real data sets and research questions, with measr as the tool to access these models. The training sessions will provide an opportunity for project staff to get feedback directly from users as they learn how to use the package and apply DCMs to their own research questions. A survey will also be administered at each training workshop to evaluate the usability of measr and guide future development.

Use in Applied Education Research: DCMs have many potential applications to education research. For example, by providing multiple fine-grained scores on assessments, results can become more actionable because teachers and students are given specific information on skills that need additional work. Beyond assessment, many intervention studies are interested in changing the state of individual (for example, non-proficient to proficient) rather than moving an individual along a continuum. DCMs are an ideal choice for this type of analysis given the categorical nature of the latent variables in a DCM. However, DCMs are rarely used in applied settings.