Classification and Analysis of Learners’ Proficiency Level in Marker Use Based on Learning Logs

Abstract

As the use of Learning Analytics has become widespread, real data has started to accumulate, and evaluation using such data has been practiced and shown to be effective globally. When data are used for evaluation, each teacher is required to combine and evaluate appropriate data according to the situation of each learner in their charge. However, teachers are not yet familiar with analyzing data; therefore, effective evaluation methods must be proposed and shared from Learning Analytics to encourage them to take the initiative in using data for evaluation. In this study, as an example of an evaluation method, we defined an evaluation index, “proficiency level,” for the use of markers on e-books using Zimmerman’s Multilevel model. The data used were three months of study log data of 112 Japanese secondyear junior high school students stored in a learning analysis platform called Learning and Evidence Analytics Framework. Subsequently, we focus on Active Reading (AR) English language learning which was held in the last week of that period. By comparing the reading speed of English text, Word Per Minutes, measured before and after Active Reading, we were able to clarify the differences in learning effects according to “proficiency level”. This allowed us to confirm the validity of “proficiency level” as an evaluation index. In the future, we will work on the creation of a higher-order evaluation index from both data and theoretical perspectives, and it will enable teachers to support evaluation that responds to each learner.