The Declarative to Procedural (D2P) tutoring system [large figure] is designed to support tutoring procedural skills that can be and need to be described to learners initially with declarative knowledge (Ritter et al., 2013). This approach is inspired by the learning and retention theory in Jong Kim's thesis (Kim, 2008; Kim, Ritter, & Koubek, 2013, Kim & Ritter, 2015). D2P is a page-based system. Pages are created in XML (in a page editor that can escape into HTML) and the D2P Ruby-based engine displays them. Pages can have tutorial information including videos, images, sound files, text, questions, and simulations. Pages can have questions based on videos, images, or text. Answers are entered with a mouse or with keystrokes. The results are logged in an SQL database, and feedback (as text, pictures, videos, or links) can be given or not given (e.g., for pre-tests).
A previous version could also pass questions into the simulation to be asked (this approach has also been tested in a project with James Niehaus at CRA for ARL/Orlando).
We are working to extend the architecture to use a knowledge representation from Herbal to annotate pages and for learners to know more about the structure of what they are learning. A conference paper provides an overview (Ritter et al., 2013), and a draft journal article summarising a more complex test of the tutor has been submitted, gotten feedback, and is being revised. It is available upon request.
We have created several tutors. The first, D2P/MTT teaches how much to lead moving targets. It has approximately 2-4 hours of self-paced material. We have tested this tutor three times (Yeh & Ritter, 2012; Yeh, Voller, Ritter, 2013; Yeh, Ritter et al., being revised). In our initial test the tutor had an effect size of 1.48; the revised tutor had an effect size of 3.7 using the most conservative measure). Our third test was with previously deployed Marines that are in the Enlisted Commissioning Program (ECP) and NROTC instructors at Penn State—Marines who have been trained numerous times how to shoot moving targets, and ROTC students. We believe—as do experts at MCWL—that current training based on powerpoint slides without practice leads to performance near chance. A test of a revised tutor shows improved performance with an effect size of betwen 2.56 (Marines) and 4.87 (students)—the number also varies based on which equation is used to compute it because the final SD is relatively low and the final performance mean is near ceiling.
D2P/CLS (Hobbs et al., 2012) teaches combat lifesaving skills, battlefield first aid for and to Marines and other trauma casualties. It covers the topics in tactical combat casualty care (TCCC), both as an introduction and with practice applying both general and specific skills. It appears to provide 2 to 4 hours of self-paced material. A technical review for MCWL found plenty of strengths and also numerous types and places where it could be improved.
We have created a tutor for AFRL to teach Air Force nurses concepts related to primary and secondary trauma assesment (STAT). A short conference paper (Garrison et al., 2019) and a journal article on this is are available (Garrison et al., in press). A movie summarizing the STAT tutor is available, created by Sydney Montgomery.
We are working on a further tutor in this area for the Defense Health Agency (VITAMMINS). The Vitammins tutor was used by Penn State Nursing students in a class this spring (2020) to replace instruction that was not available due to the pandemic that closed the hospitals to students. [PSU Press release] A study to test the efficacy of the Vitammins tutor is planned for Fall 2020.
Both the STAT and Vitammins projects are done with Charles River Analytics.
We have created a tutor (since April 2020) to teach Skills To Obstruct Pandemics (STOP) with funding from the ACS Lab, the College of IST, the PSU Applied Research Lab, and ONR. This work is ongoing. [PSU Press release]
We have created small tutors to help test and demonstrate the architecture on (a) Navy ribbons, (b) Navy rate and ratings, and (c) Chess pieces. Specific tutors have been created and used in a masters thesis on flow (Metaxas, 2018, MS thesis), eye-tracking of learners (Tehranchi, accepted, 2020), and maintence (ongoing, Ritter et al., 2019). James Niehaus at Charles River Analytics created a tutor for an aortic junctional bandage/device.
D2P is novel in that it appears to be easy to use (the two first two instructional desigers had only BS degrees and no experience in the domains) and inexpensive to use (D2P/CLS was created for $80k, including the technical evaluation). D2P and its approach is focused on skills that often have not been put into a tutoring system, skills that start out declaratively but end up as procedural skills. The tutors have been tested, albeit not completely or fully. But, the best test so far on the D2P/MTT suggests that it tutors in a single session well enough to train users to shoot at moving targets far better than existing instructional materials.
Also available (links below), are the tutoring engine/architecture, a manual, the page editor system and question builder, and an example tutor designed to show basic page use. We also provide a short video of D2P/CLS running. The existing D2P tutors, D2P/MTT and D2P/CLS are distributed by request; they are slightly large and installation can be slightly fussy (although it should not be, we are working on this). If you would like to use one of these tutors, please contact frank.ritter @ psu.edu for how to acquire them.
D2P Manual for the D2P engine and also the page builder and the question builder [videos demoing the Java version]
D2P Architecture 32, and 64 which lets you run tutors built with the page editor or other XML editing tool
Example Tutor designed to show basic page use by Ritter [you will have to control click to download it, or view source, because if you just open it in a browser, the XML will be stripped off! and it will look like not much. You will also have to use the manual to see how to load and run it.]
Example video of D2P/CLS
References
Garrison, C. M., Ritter, F. E., Weyhrauch, P., Niehaus, J., & Bauchwitz, B. (2019). A computer-based tutor to teach nursing trauma care that works as an adjunct to high fidelity simulation. Simulation in Healthcare [journal supplement] 14(4), Abstract #173, e128-e129.
Garrison, C. M., Ritter, F. E., Weyhrauch, P., Niehaus, J., & Bauchwitz, B. (2020, in press). A computer-based tutor to teach nursing trauma care that works as an adjunct to high fidelity simulation. Simulation in Healthcare.
Garrison, C. M., Ritter, F. E., Bauchwitz, B., Niehaus, J., & Weyhrauch, P. (2020, in press). A computer-based tutor to teach nursing trauma care that works as an adjunct to high fidelity simulation. Computers, Informatics, Nursing.
Hobbs, J. N., Ritter, F. E., & Morgan, J. H. (2012). D2P/CLS: A tutor for Combat Lifesavers. In Proceedings of the 21st Conference on Behavior Representation in Modeling and Simulation, 12-BRIMS-043, 226-227. Amelia Island, FL: BRIMS Society.
Kim, J. W. (2008). Procedural skills: From learning to forgetting. Unpublished PhD thesis, Department of Industrial and Manufacturing Engineering, The Pennsylvania State University, University Park, PA.
Kim, J. W., & Ritter, F. E. (2015). Learning, forgetting, and relearning for keystroke- and mouse-driven tasks: Relearning is important. Human-Computer Interaction, 30(1), 1-33.
Kim, J. W., Ritter, F. E., & Koubek, R. J. (2013). An integrated theory for improved skill acquisition and retention in the three stages of learning. Theoretical Issues in Ergonomics Science, 14(1), 22-37.
Metaxas, L. R. (2018). Impacts of user sentiment on information recall, intrinsic motivation, and engagement in the context of intelligent tutoring systems. Unpublished MS thesis, College of Information Sciences and Technology, The Pennsylvania State University.
Ritter, F. E., Tehranchi, F., Brener, M., & Wang, S. (2019). Testing a complex training task. In Proceedings of the 17th International Conference on Cognitive Modeling (ICCM 2019), 184-185.
Ritter, F. E., Yeh, K.-C., Cohen, M. A., Weyhrauch, P., Kim, J. W., & Hobbs, J. N. (2013). Declarative to procedural tutors: A family of cognitive architecture-based tutors. In Proceedings of the 22nd Conference on Behavior Representation in Modeling and Simulation, 13-BRIMS-127. 108-113. Centerville, OH: BRIMS Society.
Yeh, K.-C., & Ritter, F. E. (2012). An initial evaluation of the D2P/MTT, a computer-based, Declarative to Procedural (D2P) theory driven moving target tutor (Tech. Report 2012-1): Applied Cognitive Science Lab, College of Information Sciences and Technology, Penn State.
Yeh, K.-C., Ritter, F. E., & Voller, K. (2013). Notes on second test of the MTT (review of D2P/CLS Tutor-DEMO 12/17/2012). Technical Report No. ACS 2013-1.
Acknowledgements
D2P was created by Frank Ritter, Josh Irwin, Chad Chae, Korey MacDougall, Dan Guzek, Myeongchul Hong, Josh Irwin, Jeremiah Hiam, Mark Cohen, & Martin Yeh, with some base software written by Tatum Software. Chris Garrison and Ysabelle Coutu, and Allison Hughes, created the trauma nursing tutor, with help from Sue VanVactor and substative feedback from Teresa Millwater. Jon Morgan and J. Nick Hobbs created several initial tutors. Mark Cohen created the CLS simulation and helped create Herbal along with Steve Haynes. D2P is based on theories of learning and retention arising from Jong Kim's thesis work. ONR supported creating the tutoring language; MCWL has supported creating the tutors, Chris Carolan, Ray Pursel and Jack Sparks. Peter Weyhrauch at Charles River Analytics has given useful comments and helped develop the approach, as well as serving as the prime on some contracts. Jim Niehaus was PI on a project for ARL/Orlando to discuss how to tie tutors to simulations.
This project draws upon numerous ONR contracts, for the theory, for the application into an architecture, for a modeling language, and for building the tutors.