3rd week: case examples

During third week I’ve read some docs and watch some videos about organizations that are using learning analytics to improve their achievement.

Basically, they are using LA to dig on to their data and localize the areas they need to attend in deep or to diagnose the studentens trying to foresee their probabilities of success (mainly to diagnose students in risk of fail).

One thing very important is a basic consideration that organizations needs to rethink BEFORE doing anything: Wich is the value of the student fail?

Some high education institutions, like the ones we have in Spain, consider some degree of fail. Some teachers participates on this idea on the believe that  a high level of educational quality and proficiency entails high levels of students failure. A lot of “failings” means that the requirements are hard to achieve and not all the students are capable of this: subliminally they conceive teaching as selection process  instead of elevation process.

But nowdays, immersed in a economic crisis that are forcing to rethink what we are doing in Universities, we can’t afford such high levels of fails. Every student counts. The public (and private!) money is forced to be spent in a better and more efficient way. And it is necessary to  check the failing causes and to correct them.

In that sens, the case studies web have read about are very enlightening. Those institutions worried about how to help students to achieve high levels of proficiency detects which of them are in risk of fail and points the efforts on those problems.  An example to follow.

 

2nd week: concepts and definitions

Some important things to considere are the different concepts involved in elearning analytics:

Learning analytics: it is about measure, collect analysis and evaluation of data about learners in order to understand and improve the learning process. SOLR propose the following definition:

the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimizing learning and the environments in which it occurs.

Academic analytics: use of the processing of big data in order to take decisions about educational organization. (to make predictive models) (Probably, my phd is more focused on this concept!) CITE: http://net.educause.edu/ir/library/pdf/PUB6101.pdf

Educational Data Mining: discipline about the development of methods in order to explore data from educational process or to research about methodology and others.

So, The main difference between learning and academics analytics is the main object of study: in the case os learning analytics the target is to study how the learning process is developed. The goal is to improve learning. Academic analytics has a wider target, trying to deep in all data related to the educational process (from enrollments to graduations rates) in order to take strategic decisions.

 

Why is so important to work on learning analytics? (if you care about higher education…)

A carefully reading of the documentation provided by the course drivers gives you a set of reasons about why is important to make learning analytics.

May be a single phrase could resume the main idea:

“the economy, stupid”

Do you remember? it was the 2nd point in a famous banner in the Little Rock campaign Headquarters during Clintons’s race to the White House in 1992. It’s the economy again: the e-learning has become so complex and resources-consuming that it is neccesary to be very rational about how to proceed: you have to apply your limited resources where most needed and you need to be very effective doing this.

The cost of education has increased a lot during last years. It is very important not to waste your money and your time and, specially, the student’s money and time.

In the European context, Public Education is a concept related to the welfare state. But in United Kingdom,  the cost of higher education has increased far away this concept. This shows a possible future for all other countries like mine, Spain.

In the U.S, the situation is not very different:

In addition to good targets and goals precise formulation, It is necessary to apply accurate surgery to do the needed intervention. For example,

  • to alert the teacher staff when it is important or critical to support a student by providing coaching or
  • to provide exact information about the effectiveness of one methodology used in a specific situation with real students.

How to be so precise? How to know how, when, where and to whom it is necessary to attend?

Learning Analytics could shed light on this.

L.A. could dig up the data you need to take strategic decisions. To show up relations between variables, to point weakness and strengths. In a word: to know yourself.

I like this quotation “Online earning without embedded analytics is like a car without wheels”…”embedded analytics turns online learning into an engine for both scaling access and improving retention, persistence and completion”.

 

My starting point… why I’m attending Learning Analytics and Knowledge 2012 #lak12

Massive Open Online Courses are one of the more interesting  things had appeared in educational world during last years. From the Stanfords AI MOOC to this one, everybody around the world is excited with the idea of make real learning from prestigious institutions and to experiment the future by themselves.

Actually, I strongly believe that it is a flexion point for higher educations institutions in the same way that the MIT Opencourseware project was in 1999. (by the way, you can consult a very interesting report by diggin in http://old-internetng.dit.upm.es/wp-content/uploads/File/MIT.pdf ).

Nowadays, higher education institutions are facing a non very clear future: economic problems, new social frameworks and globalization makes necessary to re-think the whole idea about what we are doing inside and outside our campuses. Old recipes are not longer valid.

It so happens that  I’m in the process of writing my thesis report for my Phd.  I’m working trying to establish some relationship between the use of the LMS in my university with some “quality” parameters like “success”, “efficiency”, “efficacy”, measured from grades or the cost related to the gradding process. (I’ll explain in further post the quotaton mark in “quality”, “success” and others.

I’m working with a big quantity of  data: 11 years of grades, all the degrees taught at my university, all courses… near 1.000.000 student-course enrollment data.

I’ve done some approximations to the final research with a set of 64.000 enrollments in Architecture studies and I’ve presented this for my “DEA” (Diploma de Estudios Avanzados. A first research work previous to the final thesis)

At the present, I’ve obtained some results but I’ve more questions than answers… Everybody says it is normal at this stage, but I need some more analysis tools and this is one reason of why I’m enrolled in this course.

During last week I’ve been reading and watching all the papers, docs and videos proposed by the course drivers… It has been more interesting that I expected for. I would like to point:

-the video from Marisa Mayer (Google) It is enlightening how a big firm like google is providing great tools to make possible to play with big data and to process it in a very easy way.

-The document from Baker and Yacef. This document is a very good introduction to the topic and, following some citations inside, I’ve found two more documents very interesting too, for example the meta-study from Romero y Ventura about Educational Data Mining until 1995

-And the Educause Center document titled “Academics Analytics: the uses of management information and technology in Higher education” where it is possible to find why we needed to use learning analytics and other educational analytics from the management point of view of our organizations.

One of the most interesting things of this course is the possibility of sharing knowledge via the videoconference system each week we have  scheduled. It is a privilege to attend the meeting with people from all around the world, with different points of view and needs… I’m waiting for the next one!

follows bibtext record for the work from Romero y Ventura
@article{romero2007educational,
	Author = {Romero, C. and Ventura, S.},
	Date-Added = {2012-01-30 11:01:17 +0000},
	Date-Modified = {2012-01-30 11:26:43 +0000},
	Journal = {Expert Systems with Applications},
	Keywords = {Educational Data Mining},
	Number = {1},
	Pages = {135--146},
	Publisher = {Elsevier},
	Title = {Educational data mining: A survey from 1995 to 2005},
	Volume = {33},
	Year = {2007}}