Since SLAM project is using advanced technologies, we also need advanced expertise to provide support for this side of the study. That's why we are lucky to have Abdul Moiz in our team to take care of some of the technical aspects.
Here Abdul answers on couple of questions about himself:

1. What’s your background, expertise and interests?

I am a currently a Doctoral student in University of Oulu and my previous educational background is MS in wireless communication engineering. My area of expertise are Software development and wireless communication engineering. I also have quite extensive experience in fixed network planning, design and implementation. My interest is frontend as well as backend software development.

2. What’s your role/roles on SLAM project?

I am working in SLAM as an application developer as well as backend server support. My main task is to transfer the real-time sensors data from Empatica E4 wrist band to the backend server via a tablet PC as an intermediate node. My responsibilities also include support to edX server.
 
3. What you expect to learn from this project?

I love new challenges and want to learn more and more new things. This project has exposed me to many more new technologies and I have learned a lot of things in this project. I expect to learn more in detail about the setup and management of edX server and how to make it more robust and fail proof.

The proliferation of Massive Open Online Courses (MOOCs) with large numbers of learners from all over the world has provided the necessary fertilizer for the learning analytics field to grow and develop: a vast amount of ‘learning’ (actually study) data. The interaction of students with the platform to meet their learning goals leaves traces that can be inspected and analysed to understand how (online) learning happens. Event logs storing when the students sign in and out, watch videos, solve problems, interact with each other, use both specific and generic tools, and so on. These events can inform us about students’ behaviour and habits.
 

Learning analytics is a hybrid academic and also functional discipline that benefits from knowledge, techniques, and tools from computer science, educational data mining, information visualisation, and the learning sciences among other areas. All of these disciplines merge and interact with the purpose of enhancing the learning process based on data. In the SLAM project we have realised that there might be another powerful discipline that learning analytics can benefit from, namely psychophysiology (the branch of psychology dealing with the physiological bases of psychological processes). The name may sound sophisticated if you’re not used to it, but it might become commonplace or even a trend for what it has to offer.
 

Truth is that psychophysiology is an intricate field because it deals with the most perfect, complex and mysterious machine ever created: the human body. In short, psychophysiology studies the psychology behind physiological responses and vice versa. Think about how our heart rate increases when we get scared or how we perspire less when we are at ease. Psychological or mental states can be categorised in three big groups: cognitive, emotional and behavioural. All three play key roles in the regulation of learning. It’s no surprise then that psychophysiology is a promising ally for learning analytics. There is so much our body can tell us about anything we do, and there is so much unknown from the physiological language to date, that the possibilities for exploration appear to be endless. There’s a lot to discover but also much already explored terrain. Some physiological responses have been connected for example to mental effort (eyes fixation duration, heart rate acceleration), attention (skin conductance responses), engagement (spontaneous skin conductance activity), and performance (heart rate variability, breathing).
 

Measuring physiological responses – that is to say making sometimes barely observable physiological responses visible - sounds like something you have to do in a laboratory. However, the measurement devices, the biosensors, are invading modern life as wearables in the form of smartwatches, wristbands, armbands… Adoption rates are rising mainly for physical activity tracking, typically having information on heart rate during exercising, but what if people could also use their biosensors to get feedback on their learning processes? With the help of such physiological data we in the SLAM project are aiming for a biofeedback dashboard to explore the support of collaborative learning in our next experiment.

 
This does not mean that learning scientists have to become psychophysiologists, but it does mean that a new tool for learning research to embrace is knocking at the door. Will you give it a warm welcome? Let us know in the comments.
 

 Héctor J. Pijeira-Díaz
 

 

We are happy to welcome Gayane Sedrakyan to the SLAM team! She's visiting us for the next few months, and we're looking forward to benefit from her expertise in learning analytics. Below you can read her introduction.

We have behind us two days of intense work on the SLAM project. Whenever Paul is in Oulu, we try to make the most out of his visits and progress as much as we can together. The first day we reviewed our conclusions from the pilot data collection. The main conclusion was that in order to find out learning relevant information from the physiological data we need to contextualize it and we can do that by combining it with video data analysis. The second day we spent at the seaside where we could focus solely on the next data collection which will take place this coming spring (follow us to see how things develop). The program included reviewing the scientific aims, and designing the experiment for next spring. We had fruitful discussions about the pedagogical design, learning environments and data collection tools. We decided to conduct the experiment in an advanced physics course, following an inquiry learning approach. We still need to find the learning environment that provides us with detailed log data while being flexible enough so that it would allow us to include the elements that would fit our pedagogical design. More discussions to follow on this… These intense sessions are a great way to make decisions that allow us to progress further, especially when working with international partners.

As Prof. Paul Kirschner stated on our earlier blog post, we are reaching for “new” objective forms of data on SLAM project. This also explains why research infrastructure and tools play a major role in this research. The SLAM pilot data was gathered in the LeaForum research laboratory. What is LeaForum and what does it have to offer for learning research?

Environment

LeaForum is a space located in the University of Oulu and it can be used for research, teaching and business purposes. At a quick glance LeaForum looks like a really modern classroom or seminar space but looking closely it can be quickly seen that there is more to it.  LeaForum allows many different kinds of uses, and it is especially suited for technology supported learning and interaction purposes. Modern observation and recording systems are smoothly integrated in the LeaForum facility.

Collaboration requires a flexible learning environment. This aspect has been taken into account already in the design process of LeaForum, and the architectural choices were made based on research. The spacious room is dividable in three sections and the multifunctional furniture allows a variety of group formations. For example in the SLAM pilot data collection we organized the setting so that students could comfortably work in small groups without disturbing each other.

Multimodal recording and analysis

High-tech recording equipment is also important part of LeaForums infrastructure. For example the MORE recording system is especially useful in learning research. It utilizes a multichannel approach in data collection and records audio using individual audio channels and video in 360 degrees. This ensures recording of the learning process as accurately as possible. The system also uses advanced signal analyzing techniques such as speech processing, motion activity detection and facial expression detection. The facial expression detection is used as also in SLAM pilot. The MORE system can also be used as a travel version in a classroom setting.

The laboratory has a wide variety of equipment that can be used in collaborative learning, such as tablets, interactive video projectors, interactive tables, whiteboards etc. Equipment also includes variety of possibilities for psychophysiological data recording. In the SLAM pilot we used wireless bracelets (Empatica E3) as a main recording device for psychophysiological measures.  We also collected eye-tracking data with SMI eye tracking glasses. Collaboration with the professional staff of LeaForum has made this all possible.​

Ecological validity

In future SLAM project heads towards even more ecologically valid environments, real classrooms. Modern and mobile research equipment of LeaForum will be used to capture the events as accurately as possible. Before that we have many challenges to face concerning research setting, environment and tools. The goal is to combine the objectivity of multimodal methods and learning analytics to situational “real world” research setting as well as possible. Let’s see how far we can get!



More about:

The SLAM project has begun and we are busy with our first pilot study. Our goal at this moment is to test the usability and functioning of a number of objective measures of affect and emotion to replace the usual subjective ones.

A well-known saying is: Why to do it the hard way if it is possible to do it easily? So the first question is WHY? There is so much research being done in the social sciences that makes use of subjective, self-report measures of how someone feels or experiences something, why then have we chosen not to do this too? Why do we find it so necessary to do it differently (other than that those who know me know that I pride myself in being different)?

Let’s say you want to measure how excited someone is with a birthday present. Of course you can ask her/him if (s)he was excited, but the chance is very large that (s)he will give you a ‘socially acceptable’ answer. And here is exactly the first problem with subjective self-report: People often give the answers they think the questioner wants to hear. But is it real?

We know that when people get excited about something, there are also physiological processes that mirror this. If someone gets excited, you’ll usually also see an increase in heart rate. Thus, if someone’s says that the present excites her/him but there was no increase in heart rate, then you might want to question the veracity of the answer. And if you can measure the heart rate, then you actually don’t even need to ask. Another example is if you become anxious (for example because the deadline is drawing near and the project you’re working on is not progressing well and thus, your grades will suffer). If this is the case, then you will probably begin to perspire, and the conductivity of your skin will increase. Thus, if you continually measure someone’s skin reactivity (i.e., their electrodermal activity) then you will be able to see if someone is becoming anxious and conclude that this is the case if the reactivity increases. This is the basis of a polygraph (lie detector).

And then you might want to know how excited (s)he is. Again you can ask, but along with the socially acceptable answer ‘extremely excited’ we come across a second problem, namely judging the degree of excitement. What does the respondent mean with ‘extremely’ and is the one ‘extremely’ the same as another? Thus, the second problem is the scale; how does the person benchmark her/his level of excitement? This, for example is one of the problems with an often used scale for measuring mental effort. When asked to report the degree of mental effort expended when solving a problem or carrying out a task, a learner doesn’t really have a benchmark to gage this. How much effort is a little or a lot and how do you compare it from one instance to another?

Going back to the earlier examples, if you have a base heart rate and a base galvanic skin response, and then continually or incidentally measure these two, you will not only be able to determine that someone is more excited or anxious, but also the degree of excitement and anxiety. And when you equate this with what (s)he is doing or saying at the moment, you can then see how a phenomenon that occurred during a collaborative session influenced one or more of the team members affective or emotional state of mind. Take for example your heart rate when exercising. By monitoring your heart rate during exercise, and equating it to the different exercises being done, you can determine how taxing each different exercise is.

In SLAM we want to measure a number of affective or emotional reactions that learners have when working with each other. We are studying the social regulation of collaborative learning, and as such, we need to get a good grip on the ‘social’ part of working in a group or team. How learners in a team work with and relate to each other is, to a certain extent, dependent on how they relate to and feel about each other. In order to interpret the processes observed during the learning experience, we need to know how the learners felt about each other and how they thought the others felt about them. I’m sure everyone reading this has worked with others in a team and had their behaviour modified by their feelings and emotions. Two of the project’s objectives are:

·         Make visible situational characteristics (i.e., cognitive, motivational and emotional) that determine the success of individual, collaborative and collective learning.

·         Find evidence for and representations of critical phases of students’ strategic regulation of learning (e.g., log traces, emotion tags, facial expressions).

The measures are:

·         Heart interbeat interval, Heartbeats per Minute / Blood volume pulse

·         Body temperature

·         Hand/arm movement / Accelerometry

·         Electrodermal activity

·         Eye movements / Pupil reactions

·         Recognition of facial expressions

·         Recognition of vocal expressions



In the pilot we continually measured these objective markers along with what was happening in the teams. With the aid of data analytics, the next steps are to (1) see which measures change with which social and cognitive activities, and (2) see whether different markers yield similar results, allowing us to choose the simplest for our further research.

With a comprehensive dataset in our hands, we undertake this exciting, ambitious journey to understand and scaffold learners’ self-regulation when learning alone and socially-shared regulation when learning in groups. Keep an eye on the blog as we will be sharing insights and findings from the SLAM project as they appear.

-Paul Kirschner

The first pilot study has been conducted related to the SLAM research project. The project aims at researching strategic regulation of learning both in individual and collaborative level via learning analytics and mobile clouds. Participants of the pilot study were 49 first and second year high school students from the Oulun Normaalikoulu. The pilot took place in Leaforum (http://leaforum.fi/) , where students had to work in groups to plan the perfect breakfast for different groups of people: marathon runners, cardiac patients and diabetics. The students used a learning environment accessed on IPADs to solve the task.

Multimodal data was collected from the participants while they were working: log data from the learning environment, eye-tracking data, data from physiological sensors and video data. The next step will be to start the analysis of these various sources of data to find out more about how students learn in groups, how they handle critical situations and what kinds of emotions they have throughout the learning session. The results from the pilot will direct the planning of the next experiment, which will be in the school context, in 2016.

This pilot was the result of the hard work of the many partners: the support from part of the Oulun Normaalikoulu staff and rector was essential to the success of this pilot, as was the help received from the Leaforum staff.