MozFest – Learning Analytics

Learning Analytics are targeted at administrations that is trying to understand how teachers and students are interacting with the tools and processes inside the educational system.

(This blog post is just a collection of my personal notes and extra thoughts and is not something that a specific person said inside a discussion and that I probably misunderstood anyway).

There are multiple aspects to the concept of Learning Analytics, depending on who you are and what are you tracking.

Online Courses

Simple and straightforward tracking of students progress, how much time they spent on a specific piece of content, what worked for them and what didn’t. The basic premise being that you can have a feedback loop built into each iteration of the course and refine it as you go.

Besides all the research in the MOOCs area, I think there are two other interesting areas of general development in this field:

Commercial Tools like Summit Evergreen.

The problem that they are solving is that if people don’t finish their paid commercial course, they want their money back and they don’t learn and you can’t offer them better courses (and it’s overall a bad experience). So you closely track their progress and once you see that they’re not progressing, you start emailing them and trying to help them overcome whatever problem they have.

General research areas.

As we start to develop better online tools, we can also start publishing academic research papers on user behaviour. There is a persistent ethical question of when is ethically unacceptable to even build in tracking mechanisms compared to gathering everything where you anonymise the data after the fact. I do not think this part of community has already developed as well as understood all the issues. The fact that everyone is trying to publish papers so they can justify their research and get PhD’s does not provide good incentives for researchers to self-limit their access to data.

Motivational Characteristics

If you’re working in Web development, you’ve heard of a concept of *Dark Pattern* by now. It is a practice of interface that is designed to trick you into doing something you did not want, buying extra insurance or subscribing to newsletter. Whole experience of buying airplane ticket online also comes to mind.

It shouldn’t be a surprise for me, that there is a strong correlation between personality traits of certain age groups and how they react to different user interface elements. I do not have better data, but as far as I understand there are certain age and gender groups that will be strongly influenced by, e.g. the number of follower counts.

This has several interesting implications – do we try to design with such traits in mind or do we perceive it as a dark pattern and rather build more online version of a textbook instead of highly competitive online learning environment?

Designing Online Reputation Systems

I would love to learn more about different motivational characteristics of online users. But until I can find a good resource on this topic, I can recommend an excellent talk by Randy Farmer:

(and his book – Building Web Reputation Systems).

It’s about Educating Everyone

Some the issues and exposed problems might be mitigated by having better and more clearly presented privacy policies. Instead of just gathering data, get users to opt-in into collection and clearly present value to them and how this research helps everyone. Example of such data collection project done in an extremely transparent fashion was Mozilla Lab’s Test Pilot – https://testpilot.mozillalabs.com/ . Where they would show you a graphic of the data you’re about to send and ask you again if you’re ok with that.

Having similar approaches to the analytics tools inside different online services in this field could make it easier for both students and their guardians to understand what is going on.

To me it also seems possible that we’re much sooner see new legalisation in this regard in EU compared to USA. This will once again present problems when using both open and closed platforms that will try to upsell local educators on benefits that might legal risk to them.

The Future

Big Data and Everything Analytics is an emerging trend that we will soon start seeing much more research and work done on. It will also mean that we’ll start to talk to educators about things like A/B testing and funnel analysis. More complex technologies and concepts for everyone.