Hello, everyone.

I am Li, a research student. In this article, I would like to introduce the content and my thoughts regarding a paper I read in our most recent English Literature Seminar.

• Paper Title: Comparing Recommendation Materials and Timings in a Learning Context: A Case of Wikipedia Article Recommendations

• Journal: Computers & Education

• Volume: 203

• Pages: 104854

• Year of Publication: 2023

• Authors: Souneil Park, Sanjeev Sarkar, et al.

The following is an overview of the content of the paper.

Overview With the development of digital learning environments, recommender systems that provide learning materials tailored to individual students have attracted attention. However, in the context of learning, there is still no unified conclusion regarding “what kind of materials” should be recommended and “at what timing” they should be provided to be most effective. This study investigated how different recommendation materials and timings affect students’ learning performance and search behavior, using Wikipedia articles as the learning task.

Research Method An experiment was conducted with 148 participants. The task was to learn about a specific topic using Wikipedia. The study compared two types of recommendation materials: “Related” (content directly related to the current topic) and “Exploratory” (content that expands the scope of the topic). Furthermore, two types of timings were set: “Immediate” (recommended as soon as a page is opened) and “Delayed” (recommended after a certain amount of time has passed). Participants were randomly assigned to groups combining these conditions, and their knowledge acquisition was measured through pre- and post-tests, while their search behavior was analyzed through operation logs.

Results and Findings The analysis results revealed that the effects of recommendations vary significantly depending on the combination of material and timing. Specifically, the “Immediate Recommendation of Related Materials” was effective for short-term knowledge acquisition, helping students efficiently understand the core of the topic. On the other hand, the “Delayed Recommendation of Exploratory Materials” was found to promote broader search behavior and deeper engagement with the content. It was suggested that providing new, expansionary information after the student has reached a certain level of understanding of the current topic encourages curiosity and exploratory learning.

Personal Thoughts This study provides very interesting insights into the design of learning support. In the development of Learning Analytics Dashboards (LAD) and individualized optimization software, the focus is often on “what to recommend,” but this paper strongly suggests that “when to recommend” is an equally critical factor.

In my own research on Knowledge Maps, I am considering how to recommend the next knowledge point. Simply showing the “next related item” may not necessarily be the best for the learner’s long-term growth. By incorporating the perspective of “timing”—such as providing exploratory information at the moment a learner completes a certain knowledge node—it may be possible to realize a more effective learning navigation. I would like to consider how to implement such dynamic recommendation strategies into a system in the future.

By: Li