What Moodle can tell us about our learning?
Currently I work on a new contribution for this year's ECTEL conference. This paper is again on log-file analysis and how interaction footprints can support personal learning processes. This time I focus on the information that is buried in the activity logs of the Moodle platform. This work is related to the new moodle plugin on that Steinn and I currently work as part of our activities in the GRAPPLE project. In this posting I discuss the role of activity tracking in a learning management systems.
The most important finding from studying the related literature was that activity tracking is a tool for instruction that might have positive effects on learning as well. This insight is not limited to the moodle platform, but it can be generalized to most research in the field.
In LMS and LCMS the role of the instructor as a "course administrator" is over emphasized compared to the role of the students as active learners.
This emphasizing of the relevance of the instructor is replicated in most of the research papers on using activity logging in learning management systems. Some of the papers mention that activity logging can be also beneficial for the students, but wherever this statement appears it remains mostly an alibi that is only mentioned but never elaborated any further.
Only a few studies report on giving the learners access to their activity tracking. These studies are mostly related to open learner models (OLM). In these studies activity tracking is used to explain behaviour of an adaptive system to the learners. The underlying models represent the conceptual structure of the learner model that is used for the adaptation algorithms. The open learner model can be quite complex structure.
A more educational perspective on applying activity tracking in learning management is related to concept maps and "learning spaces". Based on the activity tracking of a learner it is calculated what parts of a course's concepts have been learned, and what parts are still to be explored. This idea of "learning spaces" depends on a clear expression of the concept structure in relation to the possible activities in a online course module.
The different approaches of activity logging have in common that they rely implicitly or explicitly on the deficit paradigm of learning. In the cases where the learners have access to information that is based on activity logging, the information is not only used to show the learners what they have done so far, but -usually more importantly- what they still have to do.
This paradigm can also be found in those studies where the activity tracking is only exposed to the instructors. In these cases the greatest benefit of using activity tracking for monitoring the learning processes is that the instructors are enabled to identify those students who have the biggest deficits in the cohort. Although supporting learners with difficulties is an important task for moderators of online courses, the underlying deficit paradigm has been questioned in the educational domain for decades.
Why Moodle?
An alternative to the deficit paradigm is the "competent learner" paradigm. While the deficit paradigm presumes that a learner is lacking something and therefore has to learn in order to compensate this deficit, the "competent learner" paradigm understands that every learner has already acquired a rich set of knowledge and competences and that learners learn in order to broaden and to strengthen their knowledge and competences. The "competent learner" paradigm is closely related to the ideas of (radical) constructivism. The fundamental idea behind this is that we cannot avoid learning, but that we may get bored with some educational approaches or even offended by them.
We choose the Moodle framework for several reasons. Of course, Moodle is one of the most widely used open-source learning management systems. Secondly, at OUNL has some history of using and extending moodle. Thirdly, the GRAPPLE project also supports development for moodle. Fourthly, moodle has a central activity logging that offers detailed activity logs for learners and that is well integrated with most plugins. Finally, moodle claims to support learning by providing a social constructionist framework (see also here). Although I guess that the last statement is mostly to make teachers feel happy, we choose moodle for that reason, because we think that moodle users are open to applied constructivism in education.
What Moodle can tell us about learning?
Moodle's activity logging is simple and provides detailed data on the activity in the system. Each entry of the activity log provides the following data.
- the user who performed the action (who)
- time of the action (when)
- the course in which the action is related (where)
- the sub-system in which the action took place (did what)
- the resource object that is related to the resource (with what)
- the type of action (how)
In theory the log should also provide data on hierarchical activities (in terms of hierarchies of the related resources), but apparently this information is not usable because the related fields contain the same information as the field for the resource.
This data alone provides rich information about the activity in the system. However, many aspects of the educational processes are not represented in the system's logs but are stored in other parts of the system. This includes particularly the results of quizzes, test, surveys, grades, and discussions.
One of the nice things of the moodle activity log is that it is a system wide resource. This means that the information that can get extracted is not limited to a single course, but can include all courses in which a learner is enrolled. This allows more holistic views on the learning processes than the course centered information that were suggested so far. In overall the activity logs of moodle can provide information on several aspects of learning processes. In the following three aspects are elaborated.
- Social context
- Learning activity
- Learning rhythm
Social Context
The social context describes the social relations of a learner. For moodle two types of social context can be distinguished: Assigned relations and emerging relations. Both social relation types can be extracted from moodle's activity logs. The perspective that we take is that all social context is rooted in the learner. This context we call self.
Assigned relations are those social relations that are created through curricular constraints. To forms of these assigned relations can be distinguished in moodle: the peers in working group and the peers that are enrolled in the same courses as the learner. The first group we call "group members" and the second group we call "fellows". Both types of social relations have in common that they are rooted on the membership in groups and courses. However, this does not necessarily imply that the learners are in direct contact.
Different to assigned relations are emerging relations always based on direct contacts between learners. Again two forms of emerging relations can be distinguished. The two forms are defined through the social distance of a contact from a learner. The first group are called "contacts" these are all fellow students with who a learner was in contact. In moodle these social relations emerge through collaborative activities, such as Wikis and forum discussions. The second type of emerging relations we call peer learners. These are the contacts of the contacts of a learner.
Learning Activity
Learning activity is the refers to the effort of the learners in their learning. I prefer to use effort in the rather broad sense that includes the time spent in using the system as well as the number of actions that were performed in a system. Moodle can provide only estimates for the amount of time that a learner spends in using the system. This is a limitation of almost all web-based systems that do not apply active learner monitoring. For the actions that a learner has performed moodle provides more accurate information. For the different types of actions that are possible with moodle, the activity logs provide a good overview on what has done.
Although general information on learning activities is provided through moodle's activity log. Many interesting aspects of the learning activity is not available in the logs, but stored in the databases of the different moodle components.
Rhythm
Rhythm refers to repeating events and activities. The concept also refers to the intensity and changes of the activities over time. The rhythm adds a temporal dimension to the learning activity. The rhythm represents the structure of the learning through events and processes. Events are occurrences of activity within a predefined (relatively short) timeframe (e.g. the lecture time or a field trip), where as processes refer to interrelated activities over longer period.
Moodle supports different types of events, such as quizzes or seminar sessions (defined in the course agenda). Moreover, moodle offers process like activities, such as assignments or discussions. In addition to these the activity logs can be used to identify event and process phases for a course or a social context.
How to support the "competent learner"?
The previous section showed the potential of the data stored in moodle's activity log. I also introduced the "competent learner" paradigm. However, social context, learning activities, and learning rhythm appear unrelated to the idea of the competent learner. Infact, the underlying activity tracking can be considered as neutral to the different educational paradigms. The differences between the educational concepts in using the activity tracking is entirely based on the aggregation, the arrangement, and the presentation of the underlying data.
Aggregation describes the process to collecting and accumulating the data in the activity log into information. The aggregation process often includes statistical processing and clustering of the data. However, the literature often ignores this processing and treats aggregated information equally as the activity tracking data. In our terminology, the processing can be separated into different aggregation rules, which can be accessed through an aggregator. An aggregator is a small script that executes an aggregation rule on a data set.
Arrangement is the process of selecting and relating the results of different aggregators. The arrangement of aggregated information highlights that one way of aggregating data does often provide not enough information to support a learning process. More often the combined results of aggregators will provide information that is meaningful to the learner.
Presentation describes the ways of how learners (or instructors) have access to the information. While aggregation and arrangement refer to the content, presentation refers to the form to which the learners have access to the information. Much of my recent work dealt with visualisation as a form of presentation. However, information visualisation is only one way of presenting.
Aggregation, arrangement, and presentation are the core aspects of the information processing pipeline that allows humans to access the activity tracking. The selection and arrangement of aggregators as well as the way of presenting the information are key aspects of what is communicated.
In order to support the competent learner, two main principles have to be applied on this information processing pipeline. I call these principles perspective and contrast. The perspective principle considers the learner's current disposition. The learner's perspective is relevant for supporting the competent learner because it integrates and values the learner's prior experience as part of the learning process. In the social software world there are many applications that reflect this principle in their designs. However, the learner's perspective alone appears not to be sufficient for supporting the learning process. Considering the learner's perspective alone might lead attractive results, but in the long run "attraction" needs also some "meaning". The contrast principle offers a simple way to add meaning to a perspective by showing the information in contrast with similar information. That way learners can compare and analyse sets of information in a single presentation so they can define achievable goals and organise their learning activities themselves.
As a first example we implemented a small activity indicator that uses two aggregators of the learning activity type. The aggregation rule of the aggregators is the same, but this rule is applied on different social contexts. This allows the learners to compare their learning activity with the activity of their peers.