Abstract:

The University is nowadays a complicated educational environment, where, on the one hand, dozens of thousands of students study and get the maximum opportunities for self-improvement, and on the other hand, the high-qualified staff makes the important data-driven decisions.

Individualization is the fundamental trend of contemporary higher education. Usually, the intensity of individualization faces the problem of human resources needed to implement it on a high-quality level and increasing the cost of education. But it’s just a common target of using any digital resources - to make your staff more effective, increase productivity. This is the aim of efforts made by the research group of Ural State University. For the year “Digital tutor” project as a part of a big government program “Digital University” was worked out.

Primarily, the assessment tools, used in the courses were evaluated - it’s the basis of further data analysis. If assessment tools are weak - the teacher just doesn’t know what is the real effect of his course, due to the lack of information on students’ outcomes. In this case, Digital Tutor can show the flaws of the existing test, and recommend to improve them in order to get data, usable for the following steps. Basing on the relevant results of tests and some sorts of other students’ learning activities the system can predict the results of a given student or even cluster of students during the term, not just after the final tests. And by comparing results - historical and expected - Digital Tutor can provide teachers and deans with the information on course quality.

This information consists of several parts:
- applicability of this course for the specific audience. For example, one math course is good for students with a strong focus on learning, but another course is much more applicable for the groups with a lower level of self-discipline.
- index of virtual attendance - that means the share of lectures attended by students (or by a group of students) or materials read or download from the e-learning platform or whatever metrics of “attendance” we have.
- index for the solvability of tasks - which tasks in tests are especially hard for solving. This can indicate the part of the course which is the hardest for understanding by the students. This part requires special efforts of the lecturer or some additional materials or any other help to the students.
- expectation of successful course completion.

And for sure all the common metrics like students’ activity funnel, the exact amount of enrolled students, students who took the test, etc are shown.
All this allows Digital Tutor to provide tutors and lecturers and any other users with recommendations focused on making the course more interesting for the audience, adopted for successful results of the study for the maximum number of students, who started it.

These recommendations, we hope, make tutors and lecturers more productive by decreasing the time needed for monitoring the current state of the course’ audience and by allowing to focus on issues needed to be improved and students which have to be directed and motivated by university’s staff.

All this makes it possible for the limited number of university employees to monitor the path of education of a wide range of students and individually correct it, not spending working days in efforts to gather information, analyze it, and search the ways of improvement of the students’ study.

Keywords:

Individual learning path, online course, assessment tools, digital trail, learning analytics.