Data Capturing Mechanisms for Learning Analytics

One step of the use of Learning Analytics is capturing events when users interact with a resource. Learning Management Systems provide limited data about this process, or requires complex procedures to be accessed. Here are some ideas of simple content items and the way to observe student behavior.

The ideas explained in this page have been implemented in the Data2U Toolkit.

Videos + Multiple Choice Questions

What: Videos are now suggested as a useful way to have students prepare the lecture. But simply making a video available has no impact on students. There must be a scaffolding around it to promote engagement. One of the most used forms is to include multiple choice questions. I use these type of activities with videos recording myself writing in one single sheet of paper explaining one or two concepts. The final paper is then available for download (see link below the video in the figure). Next to the video there are a set of interactive embedded multiple choice questions. Upon providing an answer, students may verify if it is correct, try again, or see the solutions.

Analytics: The activity records the interaction with the video (load, play, pause, end), and the events for each embedded question. These events can be used as a proxy to measure the level of student engagement with the activity.

Tools: The videos are recorded with a conventional Microsoft HD web camera and the application VLC. I found Youtube to be the most flexible option to host and manage videos. Google offers an API (Application Programming Interface) that allows you to capture the events occurring while the videos are being used. The questions are created in RestructuredText and automatically translated into HTML.

Multiple Choice Questions Embedded in Notes

What: In a digital environment notes can be entirely produced in HTML. In this format, pages may contain interactive elements without complicating the authoring process. The notes used in my courses contain simple multiple choice questions to recall factual knowledge about the material just covered. The questions allow students to answer, grade the answer, and then either answer again or see the solutions.

Analytics: The outcome of the questions (correct/incorrect) and the choice of seeing the answers can be used to detect anomalies in the questions, learning strategies in the students (see all the solutions immediately), provide additional resources for the topics with a high incorrect rate, etc.

Tools: The questions are created as plain text in ReStructuredText and then processed using Sphinx to translate them to HTML with javascript to capture the interaction.

Multiple Choice Problem Sequences

What: A set of multiple choice questions that are shown to the user in a random sequence. If a question is answered correctly it disappears from the sequence. Otherwise, it is kept in the sequence and another is shown. A score is provide as the percentage of questions answered correctly.

Analytics: The outcome of the questions (correct/incorrect) and the time it takes to answer.

Tools: The questions are created as plain text in ReStructuredText and then processed using two tools: Sphinx to translate it into static HTML (surrounded by borders), and a Django application to insert the form, buttons, keep the score, and selet the next problem.