I had already read many papers related to this project, but it was time to do a more intensive literature review. For each of the papers that I had already read, as well as for more papers that I found, I answered each of the following questions in an all encompassing literature review document:
- What intelligent tutoring system was used?
- What was the main goal and significance of the paper?
- What age group was tested? How many? For how long?
- What skills were they learning?
- What were the learning-related states or characteristics trying to be recognized?
- How were the metrics defined for identifying these states? What student actions mapped a particular state?
- What was the ground truth for these characteristics? How were the metrics evaluated/validated? How did they 'know' in the experiment what the student did?
- Were any issues with the metrics discussed? Places where the metrics fell short?
I ended up doing a literature review for 17 papers, but I am still constantly looking for more related papers to read and review. The goal of the literature was to look more closely at the metrics used for similar studies. We want to look at these metrics to determine what metrics we want to use for our study.