Do subjective item difficulty estimates correlate with the objective item estimates of an adaptive learning algorithm ?

Kalinko, Nicola (2022) Do subjective item difficulty estimates correlate with the objective item estimates of an adaptive learning algorithm ? Master thesis, Psychology.

Text Master Thesis Nicola Kalinko.pdf Restricted to Repository staff only Download (1MB)

Abstract

To engage in successful learning, it is important to rehearse items in an effective way. The adaptive fact learning system Slimstampen schedules the repetition of items tailored to a learner based on the response time and accuracy history through a single parameter called the rate of forgetting. At the beginning of the learning session, no repetition data of a learner is known which leads to an initial mismatch at the beginning of a learning session. If the system could start with personalised starting values with subjective item difficulty predictions made by a learner, the scheduling of items could be optimised and lead to higher response accuracy. In this study, we asked forty participants to rate the likelihood with which they think they will remember Swahili-english word pairs on an upcoming cued recall test after a study session. The participants were either in a group with the adaptive fact learning system or a less adaptive flashcard alternative. The subjective difficulty estimates made by participants before and after the learning session correlate with the rate of forgetting and can predict test performance on a final test in both conditions. Future research is needed to estimate whether judgements before a learning session could therefore be an useful addition for Slimstampen before a learning session and help to personalise starting values. Subjective item difficulty estimates made after a learning session in Slimstampen do predict test performance on a final test, though are less predictive compared to a less adaptive flashcard alternative, possibly through Slimstampens compensation in item difficulty.

Item Type:	Thesis (Master)
Supervisor name:	Rijn, D.H. van and Velde, M.A. van der
Degree programme:	Psychology
Differentiation route:	Cognitive Psychology and Psychophysiology (CPP) [Master Psychology]
Date Deposited:	22 Jul 2022 08:33
Last Modified:	22 Jul 2022 08:33
URI:	http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/1093

Actions (login required)

View Item