Especially in higher and further education, today's learning more and more allows and also requires individual learning strategies. On the one hand, this development is a step towards lifelong learning and a higher personal engagement. On the other hand, many learners struggle with organizing learning processes on their own. Learning diaries were proven to be supportive for self-regulated learning. In this paper, we show the concept and design of a mobile learning diary application, which extends standard functionality like planning, documenting and re ecting learning behavior. Through a distributed architecture with a central learning record store, the application has access to data of all users and all learning applications within the architecture. We also report rst insights of collected data during a small eld study and some feedback obtained through a questionnaire.
The rise of E-Learning and the requirement of lifelong and exible education have
paved the way for individually organized learning processes. Although there is a
common goal, like a nal course exam, learners are able to organize the process of
reaching that goal on their own. They can nd a plethora of services and
materials on the web, which they can use in addition to provided course contents. They
can also choose when, where and with whom they want to learn and how they
want to cooperate with others. Studies proved that this self-regulated learning
can improve the learning outcome as learners experience a higher personal
engagement with the learning contents [
In Section 2, we motivate our work by introducing related work in the eld of supporting self-regulated learning. Section 3 describes the concept and design of our learning diary application and the architecture which allows for sharing learning activity data. After that, in Section 4, we will discuss the results of an initial evaluation comprising an analysis of data collected during a small eld study and feedback obtained through a questionnaire. Finally, we will draw some conclusions from our work and give some insights to our plans for the future. 2
Pedagogical psychologists distinguish between three phases of self-regulated
learning [
The provision of means for re ection and comparison are also part of the main
goals in the research area Learning Analytics [
The proposed mobile learning diary application includes basic learning diary
functionality, which comprises planning learning activities, marking activities as
done and visualizing planned and done learning activities for the re ection of
own learning behavior. The application is innovative over traditional learning
diaries in three additional functionalities:
1) Partially adaptive noti cations for better maintenance of the learning diary.
In order to reduce the daily e ort for planning learning activities, the learning
diary application reminds its users to plan learning activities for the upcoming
week on Sunday afternoon before the respective week. Because we do not expect
learners to know the exact time for every learning activity in the upcoming week,
only the module the activity is belonging to, the type of the learning activity, the
day of the week and the rough duration of the activity3 has to be provided. On
days of the upcoming week, for which learning activities are planned, learners
are reminded in the morning to plan their activities of that day with a concrete
time. In the evening they are further reminded of marking activities as done or
modifying them if necessary. This step motivates regular re ection and further
ensures the validity of the data in the learning diary. The app reminds users in
form of noti cations. Those noti cations are adaptive in that they only occur
if learning activities are planned and have not been marked as done for the
current day. The noti cation in the evening also occurs when learning activities
from other applications were recognized and no temporally matching learning
activity was entered in the learning diary. The learner can then add potentially
missing activities, which increases the completeness of the collected data.
2) Integration of learning activities from di erent learning applications. For the
integration of learning activities from di erent learning applications, there is
a need of an architecture, see Figure 1, connecting applications and allowing
to share data between them. We make use of the standardized concepts xAPI
and a central LRS. To connect learners using di erent identities in di erent
learning applications, we also use a component for identity management [
We applied our approach during April 24th and June 17th, 2017 in a seminar o ered by the department of psychology at a German university. The seminar, which was on self-regulated learning for prospective teachers, started with a four week phase of autonomously working on the course contents. Participants were asked to document their learning behavior using the app or a traditional learning diary. Due to the limitation that the app was only available on Android phones, and due to some technical problems in the beginning, most of the students decided to use the traditional approach. However, there were some students showing high interest in using the app despite of the technical issues. Unfortunately, only two users used our app over a longer time: Learner A used the app for 44 days and produced 169 statements describing 112 learning activities. Learner B used the app for 50 days, producing 120 statements for 84 learning activities. We analyzed their usage behavior of the app and present the results in the following. We will also give insights to feedback we obtained using a questionnaire.
Evaluation of partially adaptive noti cations for better maintenance of the learning diary. Ideally, we would expect a number of three statements per learning activity (one per state: roughly planned, planned with time, done). However, for both learners we only observe roughly 1.5 statements per activity. For Learner A, 46% of activities have been both planned (either roughly or with time) and marked as done, 21% have been only planned, and 32% have been directly entered as done. For Learner B, only 28% of activities have been both planned and marked as done, while 11% have been only planned and 61% have been directly entered as done. We conclude that it is di cult for learners to adhere to the three phases of self-regulated learning even with reminders. We further suggested learners to plan their learning activities for the upcoming week on Sundays. Learner B seems to follow this suggestion, as 42% of planned activities were created on Sundays. Learner A planned the majority (29 %) of activities on Sundays, too. However, he or she also often created planned activities on Tuesdays (26%) and Fridays (24%). We further analyzed how long learners plan their activities in advance (lead time, i.e., the time the activity was rst planned until the time the activity was planned for) as well as how long after a learning activity learners mark it as done (follow-up time, i.e., the time the activity was planned for until the time the activity was marked as done). Figure 3 visualizes the results for Learner A. We can observe a pattern in the lead time indicating a step by step planning of learning activities for upcoming days of the week. In contrast, the follow-up time remains relatively constant, meaning that performed learning activities are marked as such soon after execution. In the feedback questionnaire, learners state they wish to con gure if and when reminders are shown. Some also wish a better design and more intuitiveness and also more adaptive functionality, like visually highlighting important and urgent activities. They also propose acquiring more data sources like exam dates or the phone's calendar.
lead time follow-up time
learning activities (sorted ascending by creation time) Evaluation of integration of learning activities from di erent learning applications. For this part, we analyze learning activities, that were marked as done in the learning diary application together with activities from Moodle. Only one of the two learners (Learner B) used the identity management component for sharing a unique pseudonym between di erent learning applications. Learner B marked 75 learning activities as done in the learning diary application. Moodle activities were only present within the time of ve of those learning activities. Within those ve activities, 117 activity statements from Moodle are present in the LRS. In total the learner produced 296 activity statements in Moodle, meaning that 179 statements happened outside of activities that were marked as done in the learning diary. This observation can be interpreted in two ways: First, many learning activities happen without access to Moodle. Probably learners download course material beforehand but work on it outside of Moodle. This stresses the importance for including more applications to our architecture to capture more information on actually performed learning activities. Second, including Moodle activities in the learning diary leads to more complete information about learners' behavior, which stresses the importance of our proposed architecture. Even from the questionnaire, we obtained encouraging feedback for the integration of Moodle activities within the learning diary application. Evaluation of visualizing peer's learning behavior. The visualization of peers' learning behavior was also rated positively in the feedback questionnaire. Unfortunately, we have not enough users to get further insights on this aspect using data analysis. 5
In this paper, we proposed a mobile learning diary application which is innovative over traditional learning diaries in that it is able to send adaptive reminders, list activities from other applications and show statistics about other course participants. The analysis of data collected during a small eld study revealed that, even with reminders, learners show another behavior for organizing their activities than proposed by the learning diary application. It also showed that collecting data from Moodle is useful to provide a more extensive overview about learning behavior. Further, we obtained encouraging feedback for the functionalities of showing Moodle activities and course statistics from the questionnaire.
We acknowledge the limitation of having too few users for making signi cant statements about the usefulness of our approach. We consider this initial evaluation as a proof of concept implementation and there will be follow-up evaluations with more participants in the near future. We further want to work on minimizing the e ort for maintenance of the learning diary, for instance by recognizing learning activities autonomously using sensor data collected by the mobile phone. We also want to make the step from o ering planning, re ection and awareness functionality towards changing learning behavior actively by sending interventions that correspond to the learners current learning goals and actual behavior.