inst/app/www/info.md

In rbcavanaugh/pnt: Computer adaptive forms of the Philadelphia Naming Test

Shiny Computer Adapative Philadelphia Naming Test

AS, RC…GF,WH?

last updated: 2021-07-09

Ideas we want to convey here:

• what is CAT and why do you want to use it
• what are the advantages of CAT + PNT
  • why naming assessments? (brief, high correlations to overall severity, discourse…etc)
• what you don’t get from CAT + PNT (error stuff)
• other details about where the current CAT + PNT researach stands
• how to leave feedback, report issues or ask questions (ideally github issues but also contact info)
• privacy/security (nothing stored)
• how to cite the app?
• license info/credits?
  • MRRI
  • MAPPD
  • Roach 1996
  • Mirman 2010
  • Walker 2012
• link to pre-print of publication

Some initial text from Alex:

This shiny application provides a simple user interface that provides various options for delivering the Philadelphia Naming Test (PNT; Roach et al., 1996) in an item response theory framework (IRT). The PNT is a confrontation naming test utilized for assessing anomia severity from people with aphasia. The IRT parameters of this model were estimated with an archival picture naming tests set of people with aphasia (Mirman et al., 2010). Please note, that the intended population for this assessment are individuals with acquired language impairments and the testing instrument may be unreliable for other populations.

This application provides clinicians the ability to assign the number of items to be delivered to a participant. Options for the number of items include 30, 175, and a variable length version. All tests deliver items per a computer adaptive testing algorithm (CAT) that which is described in detail below. Picture naming severity, as measured by the number of items responded to correctly, on the 30 item PNT CAT (PNT-CAT30) correlate highly (r = 0.95) with the number of incorrect responses on the full PNT (Fergadiotis et al., 2019). The variable length PNT-CAT (PNT-CATVL) was designed to be administered after the PNT-CAT30 to assess change in anomia. Importantly, he PNT-CATVL does not administer any overlapping items with the PNT-CAT30. Instead, the items delivered to the participant are chosen based on the participants previous naming ability estimate. The item selection algorithm for the PNT-CATVL stops administering items when the modeled precision of the ability is greater than or equal to the value obtained by the PNT-CAT30. Estimates of anomia severity estimated between the PNT-CAT30 and the PNT-CATVL correlate highly (r = 0.89)

Upon completion of the PNT-CAT30, PNT-CATVL, or the full PNT the application provides a naming ability and its standard error of measurement for the participant and an option to save this information desktop. Participant information is not saved by the application. If you wish to utilize the PNT-CATVL to assess change in naming severity, you must have a saved and unmodified, version of the output to provide the application prior to administering the PNT-CATVL.

rbcavanaugh/pnt documentation built on Oct. 20, 2024, 7:21 p.m.