Research projects

Topic 1: Designing adaptive feedback for personalized scaffolding based on Contingent Tutoring.

We will use as a starting point the five levels of verbal and non-verbal interventions tutors can employ on the learner’s progress (Wood, Wood, & Middleton, 1978):

1. 1. General Verbal Intervention: the tutor calls the learner to undertake some action.
   2. Specific Verbal Intervention: the tutor tells the learner what to look for.
   3. Specific Verbal Intervention with Non-verbal Indicators: the tutor becomes physically involved in selecting or indicating steps.
   4. Prepared Material: the tutor simplifies the material so that the learner can complete the task.
   5. Demonstration: the tutor demonstrates a particular step in the process.

For each level, we will study how to design feedback specifically for online courses and demonstrate our findings using everyday, real-life examples. Finally, we will evaluate the effectiveness of adaptive feedback using qualitative methods (observation, interviews) and online surveys.

Topic 2: Classification of quantitative indicators in students’ interaction with Learning Management Systems (LMS) to monitor knowledge and cognitive state (co-supervised with Fredrik Milani) (reserved).

This project will explore how we can extract meaningful quantitative metrics from the user log files of an LMS to model student’s knowledge and cognitive state. We will use existing data from students’ practice with the LMS and machine learning to identify and categorize activity indicators or patterns of user activity with respect to cognitive aspects of learning.

Topic 3: Identifying learning paths in students’ practice for scaffolding motivation and promoting self-regulation (co-supervised with Katrin Saks).

This project will study the relationship between students’ learning paths, motivation, and self-regulation. In particular, we want to explore how students’ decisions while studying in digital environments are affected by and at the same time, have an impact on motivation and self-regulation.

Topic 4: Exploring the effect of response time as a student characteristic (reserved). 

(programming knowledge and basic data analysis knowledge required)

This work builds on the hypothesis that there is no linear relationship between step duration and correctness. On the one hand, a student needs a minimum amount of time in order to process the problem, retrieve appropriate information, and to construct a correct response. On the other hand, taking too long to carry out a step could indicate a lack of background knowledge, failure to retrieve critical information, and inability to address the step. To investigate this hypothesis, we will use machine-learning to predict student performance having response time as a predictive feature. The predictive modeling approach will be tested on Intelligent Tutoring System datasets.