Web Analytics for Online / Blended Learning Environments

For this post I got on Google Scholar and just went to town. I initially had no idea what to look for. At first I tried "'web analytics' education." This brought up a bunch of results that were related to training people in the art and craft of web analytics. Not exactly what I was looking for.

The next query I tried was "learning environments online OR blended 'web analytics'". This brought up results that were much more promising than what I had found originally. Here are my thoughts on two of the articles that seemed relevant to our topic.

A Model of Online Instructional Design Analytics

I wondered if this article were based on some older research, because it focused mainly on how HTTP server logs can be harvested to provide us with data similar to what we would expect to get out of Google Analytics. In this way, it was quite similar (in principle) to what we do with web analytics today. I imagine these methods were common among those pioneering the field of web analytics.

Although relatively old, this article has some merit because it outlines many of the major obstacles we still face today in doing practical web analytics. Perhaps the most conspicuous of those limitations mentioned is our general inability to compare aggregate usage statistics against the individual's usage of the system.

This article was interesting because, although the methods they espoused seemed a little quaint to me, it made me appreciate the ingenuity that the pioneers of this field had. We owe so much of what we have today to people who just did the best with what they had.

Web Mining for Self-Directed E-learning

The focus of this article was not "web analytics" per se, but the question of how we can improve the relevance of information obtained by querying "e-learning systems." The authors broad definition of "e-learning system" seemed to be any online, searchable repository of information. They specifically geared their discussion around those repositories such as Google, CiteSeer, LexisNexis, and any number of other similar sites on the Internet that attempt to provide "results" based on "relevance" to the query.

The authors stated that there is an increasing trend in the number of "self-directed e-learners," or those people who use the Internet as a knowledge-building tool. This includes novices who are seeking a broad but shallow summary of an entire field of expertise as well as those experts who are seeking depth of information. The purpose of this article was to propose a number of ideas through which these information retrieval systems could be made dynamic and self-updating, providing more relevant results to the user based on the user's usage of the system.

While much of their discussion did not relate directly to web analytics, they did have a section on mining "web usage data," and the idea that grabbed me here was their idea to collect web usage data for websites, and make this data available to the users of the website in real time. This way, people using a website could see how others are using it and what types of results they are getting.

This provides an interesting question for web analytics. Imagine a large, complex website with many possible navigation paths. Or, imagine an online course website with fairly straightforward start-to-finish navigation options, but with a number of different tools available to the user. In both cases, an interesting question would be, How would website usage trends evolve if people using the site were given a realtime view of the website's usage trends? Or, from another perspective, is there any way we could identify usage patterns of those who are doing well in the course and give this information to other users in a way that would encourage them to use the course in the same way? Could this be of any actual benefit to the other users?