Keywords

1 Introduction

In recent years, learning environment in which many learners participate is widespread. Many educational institutions introduce LMS (Learning Management System) and carry out long term operation. LMS manages online learning such as e-learning. In addition, LMS stores learning logs such as login and logout. Therefore, LMS such as Moodle or Canvas have enabled collect large scale learning logs easily. The ADL (Advanced Distributed Learning) defined xAPI (Experience API) that is a technical specification of the learning logs [1]. xAPI specifies a structure to describe learning experiences and defines how these descriptions can be store LRS (Learning Record Store). LRS is database system for standardized learning logs by xAPI. According, it is easy to analyze the learning logs to figure out the learner’s achievement level and some problems. LA (Learning Analytics) catch a great deal of attention in learning research domain. SoLAR (Society for Learning Analytics and Research) defined LA as the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimizing learning and the environments in which it occurs [2]. The latest study focuses record not only learner’s learning logs but also teacher’s teaching behavior after class hours. Analysis of teaching logs provide important teaching behavior and corroborating of learner’s behavior.

The study of Majumdar (2019) record teacher’s behavior after class hours [3]. However, learners usually learn individually during class hours especially exercise. In addition, learners question and receive advice during class hours from mentor(teacher and teaching assistant). Accordingly, mentoring (teaching behavior) affects learners not only after class hours but also during class hours. Usual LA researches dose not focus on the face-to-face mentoring.

The previous our study, we developed STELLA (Storing and Treating the Experience of Learning for Learning Analytics) [4]. STELLA standardizes learner’s detailed learning history by xAPI, store the learning history in LRS. However, we also didn’t consider face-to-face mentoring to learners. Therefore, we need to develop system which store mentoring history during class hours.

In this study, we propose that store mentoring history in our LA environment. Besides, we develop LA environment that store and analyze mentoring history. To realize the environment, we developed MLR (Mentoring Log Recorder) and MRS (Mentoring Record Store). MLR stores mentoring log during class hours as a mentoring history in MRS. In addition, MLR can share storing mentoring history between mentor. Furthermore, we report impact of mentoring for learner’s learning behavior and achievement level of exercise.

2 Our LA Environment

We developed former LA environment. Figure 1 shows our former LA environment. The previous study, we developed STELLA. STELLA is module of LMS. Teachers can upload teaching materials of PDF format in STELLA linked to the LMS. Teachers and learners use STELLA for viewing the teaching materials. STELLA outputs their viewing behavior as learning logs and records pseudonymously the learning logs. The learning logs are including such as pseudonymously userID, the teaching materials name, page number and viewing time. STELLA standardizes the learning logs by xAPI and store the learning logs in LRS. The visualization tool visualizes their page transition of the teaching materials from the stored learning logs.

Fig. 1.
figure 1

Our LA environment

Full size image

3 LA Environment Introducing Mentoring History

Figure 2 shows our LA environment introducing mentoring history. MLR has two functions. One is to store mentoring performed by mentors as a mentoring history in MRS (Meontoring Record Store). The other is to share how each learner received instruction during class hours among mentors.

Fig. 2.
figure 2

LA environment introducing mentoring history

Full size image

3.1 Mentoring History Storing Function

MLR stores the mentoring history in MRS. Details of the mentoring history are as follows; timestamp: mentoring history occurrence time, actor: mentored learners, checker: mentor who performed mentoring, course: class name, exercise name: exercise name, number of exercise: number of exercises, exercise num: exercise number, evaluation: exercise score, comment: advices and questions. By collecting mentoring histories with MLR and storing the histories, it is possible to share mentoring during class hours and analyze the relationship between the degree of achievement and the mentoring history.

3.2 Mentoring Sharing Function

MLR is improved real-time sharing and management system for exercise progress in practical and exercises. MLR has two pages. One is to confirm exercise achievement of the entire each learner. The other is to confirm the exercise achievement of individual. Mentors can check and evaluate for exercise using MLR. MLR has exercise pages for evaluating the entire each learner. Also, MLR has personal exercise pages for evaluating each learner. In addition, in order to use student information registered in LMS, the Moodle module is used to store detailed mentoring history in MRS.

Exercise Page

Exercise page have a function to check whether each student has achieved exercises. MLR is LMS module. Therefore, we can create by referring to the student number and the name of the learner who belongs to the course in exercise page. Mentors can refer to it when checking the percentage of attendees and the percentage of exercise achievements. Also, announcements can advise the entire learner when the learner is performing exercises. The history of advices displays when the entire learner receive advice. Regarding the contents of MLR table, the number of cells changes according to the number of exercises. The five choices of a four-step evaluation and question as to whether the learner has achieved exercises, can evaluate the learner’s exercise. When we perform the evaluation, the time at which we performed the evaluation is displayed at the final check time, and “done” indicating that transmission has finished is displayed in the evaluated cell. Also, when the learner has achieved exercises, the number of persons achieving exercises is counted. The background color of the learner’s cell changes yellow-green (Fig. 3).

Fig. 3.
figure 3

Exercise page (Color figure online)

Full size image

Personal Exercise Page

Personal exercise page has functions to perform students’ advice, responses to questions, and evaluations. The page can not only evaluate each exercise but also leave comments for the learner. Mentor can check the comments for each learner to understand which exercise the student has stumbled on and where they could not understand. The learner’s situation can understand by leaving a comment in the student information. For example, exercises could not be completed due to poor physical condition.

4 Experimental Result

The purpose of this experiment is to verify whether mentor behavior influence learner based on timing of mentoring, changing of learner behavior or achievement of exercises. We also visualize learning history and mentoring history. Therefore, we collected the learning history and mentoring history that occurred in the class held at Kochi University of Technology.

4.1 Experiment Outline

We collected the learning history and mentoring history for programming experiment 1. The period is six times between July 2 and July 19, 2019. The class hours are the third and fourth classes on Tuesday and Friday, and the location is PC room. In the first half of each class, lesson explanations are given, and in the second half, three to five exercises are asked. The collection of learning histories targets at students who have used STELLA among 109 registered students. So that students who do not use STELLA can view class materials, KUTLMS of Kochi University of Technology display the same class materials. Students must complete exercises during class hours and have the mentor evaluate exercises. The maximum number of mentors who evaluate exercises is about 10 mentors(2 teachers and 8 Teaching Assistants).

4.2 Stored Mentoring History

Table 1 shows the number of mentoring logs the mentor gave to each student in the six classes. We also used mentoring history and exercise achievement to analyze.

Table 1. The number of mentoring logs
Full size table
Table 2. The number of mentorings
Full size table

4.3 Result

Analysis Method

We use all student exercise achievements in each class to analyze. In addition, mentoring histories use advice to students, responses to questions from students, and evaluation of student exercises. The analysis method uses Welch’s t-test. We find the significant difference between the average of the exercise achievements of two independent groups in each exercise. This time, we conducted an analysis to find the significant difference between the average of exercise achievements of students who received advice from the mentor and students who never received advice from the mentor (Table 2).

Visualization of Learning Behavior and Mentoring

Figure 4 shows the timing of learning behavior and mentoring. it is possible to confirm the change in learning behavior and the timing of mentoring.

Fig. 4.
figure 4

Visualization of learning behavior and mentoring

Full size image

5 Conclusion

This paper described our LA environment introducing mentoring history and analysis of mentoring history. Our study, we developed MLR and MRS. MLR store mentoring during class hours as a mentoring history in MRS. MLR can share storing mentoring history between mentor. We conducted experiments of to store and share mentoring history. In addition, we visualized and analyzed impact of mentoring for learner’s learning behavior and achievement level of exercise. As the result, the visualization had suggested that mentoring impact on learning behavior. Besides, we confirmed that mentoring impact on achievement level of exercise.

In our future work, we continue to use MLR for other classes. We hypothesize that mentoring affects learning behavior. We will analyze some relationship of learning behavior and timing of mentoring. In addition, we will analyze learning behavior and content of mentoring. We will research mentoring that expand an understanding of class of learner.