Which Is More Effective… Centralized Or Decentralized?

On 15 March 2019, the Government of Ontario announced the Education that Works for You – Modernizing Classrooms proposed policy.  From an e-learning perspective, the proposal calls for:


The government is committed to modernizing education and supporting students and families in innovative ways that enhance their success. A link to e-learning courses can be found here:

Starting in 2020-21, the government will centralize the delivery of all e-learning courses to allow students greater access to programming and educational opportunities, no matter where they live in Ontario.

Secondary students will take a minimum of four e-learning credits out of the 30 credits needed to fulfill the requirements for achieving an Ontario Secondary School Diploma. That is equivalent to one credit per year, with exemptions for some students on an individualized basis. These changes will be phased in, starting in 2020-21.

With these additional modernizations, the secondary program enhancement grant will no longer be required.

As you might imagine, we have received a lot of questions about these issues.  However, the most frequent one has been whether a centralized model of e-learning is more effective than a decentralized model.  As such, we wanted to provide some background and data focus on this query.

Canadian e-Learning Context

As outlined by the State of the Nation: K-12 e-Learning in Canada annual reports, there are a variety of organizational models that are used across Canada.  Less populated jurisdictions tend to utilize a centralized model of e-learning (e.g., Newfoundland and Labrador, Nova Scotia, New Brunswick, the Northwest Territories, and the Yukon) – often operated directly by the Ministry of Education or by a body designated by the Ministry of Education to have a provincial/territorial scope.  Other jurisdictions have a more decentralized model where school districts or school board operate e-learning programs (e.g., British Columbia, Alberta, Saskatchewan, and Ontario).  Finally, there are jurisdictions that have some combination of the two models (Quebec and Manitoba) – although in the case of Manitoba the centralized models use legacy forms of distance education, while its online learning program is primarily district-based.

While Ontario has been classified as a more decentralized model, this does not tell the full story.  Unlike British Columbia, Alberta, and Saskatchewan, in Ontario it is only the operation of the e-learning program that is decentralized.  The resourcing of these e-learning programs is quite centralized.  According to the Ontario profile on the State of the Nation: K-12 e-Learning in Canada website:

Since 2006, the Ontario e-Learning Strategy has guided the Ministry of Education to provide school boards with various supports necessary to provide students with online and blended learning opportunities. The Francophone version of the strategy, Apprentissage électronique Ontario, was released in 2007. Under this policy, the Ministry provides school boards with access to a learning management system and other tools for the delivery of e-learning, asynchronous course content and a variety of multimedia learning objects, and a variety of other technical and human resource supports (including a “Technology Enabled Learning and Teaching Contact” in each school board). School boards delivering either online or blended learning must sign a “Master User Agreement” to access all of these services.

Essentially, e-Learning Ontario (a unit of the Ministry of Education) centrally provides all of the tools and content needed to deliver e-learning, and even provides each school board with human resources to encourage the use of these services.  The only decentralized role in the existing system for school boards is basically is the determination of which courses will be offered, the selection of the individual teachers to provide instruction in those courses, and the enrollment of students into the centralized learning management system.  However, even these individual school board-based programs cooperate with other school boards throughout the province as a part of the Ontario eLearning Consortium, Ontario Catholic eLearning Consortium, and/or Consortium apprentissage virtuel de langue française de l’Ontario to maximize their online offerings by sharing course offerings, resources, and students.  The reality is that the existing system in Ontario is already highly centralized.

Which Is More Effective?

To date, there has not been a lot research published on the effectiveness of e-learning based on the organizational model.  Most of the research has focused on whether distance and online instruction is as effective as face-to-face or brick-and-mortar instruction; although that often has more to do with who is enrolled in the online learning than the effectiveness of that modality (see pages 525-528 of Barbour [2019] for a good summary of this issue and these findings).

From the jurisdictions that utilized the more centralized model, we can report that Barbour and Mulcahy (2008) found that rural students enrolled in the Centre for Distance Learning and Innovation (CDLI) performed as well or better than any of their classroom or urban counterparts in their final course scores during the programs first three years of operation.  Similarly, the authors also reported that e-learning students performed as well as any other group of students in their grade twelve public examination results.  More recently, data from the Newfoundland and Labrador English School District from the 2017-18 school year showed that students enrolled in CDLI courses had a successful completion rate of 86.8%, as compared to 80.9% for students enrolled in those same courses not in CDLI (i.e., in the traditional classroom environment).  This data indicates that students can be successful in a centralized organizational model.

Prior to the introduction of the CDLI, several school districts in Newfoundland and Labrador maintained online Advanced Placement (AP) program.  Barbour and Mulcahy (2006) examined student performance in these online programs and found that online students performed as well or better on the AP exam than their classroom-based counterparts.  In the highly decentralized province of British Columbia, a Ministry of Education presentation reported on the course completion rate of students in that province based on whether they completed a distributed learning (i.e., e-learning) course.

Click to enlarge the image

As the above slide indicates, during the 2009-10 school year there was a 7.4% difference between the completion rate of students that had taken one or more e-learning courses compared to students that did not take any e-learning courses.  During the 2010-11 and 2011-12 school years the completion rate was approximately the same, and by the 2012-13 school year students that had taken one or more e-learning courses had a higher completion rate than the students that did not take any e-learning courses.  This data indicates that students can be successful in a decentralized organizational model.

In terms of data specific to the Ontario context (as that is where this specific proposal is being considered), there is even less research available.  However, it was reported in the 2010 State of the Nation: K-12 e-Learning in Canada report the  Consortium d’apprentissage virtuel de langue française de l’Ontario had only a 4% failure rate during the 2009-10 school year.  Data collected by this CANeLearn-partnered project has found that the Ontario eLearning Consortium has also consistently had a 90%+ completion rate.  This data indicates that students can be successful in an organizational model that is resourced centrally but operated in a decentralized manner.

Main Takeaways

Before we conclude, it is important to point out that the research comparing student performance based on the medium or model of instruction is fundamentally flawed.  As University of Southern California professor Richard Clark wrote more than 35 years ago, the medium or model “are mere vehicles that deliver instruction but do not influence student achievement any more than the truck that delivers our groceries causes changes in our nutrition” (p. 445).  To put it another way, Kent State University professor Rick Ferdig has often written that researchers are constantly asked does e-learning work, whereas the better question is under what conditions can e-learning work?  This would run counter to the media reports around the work of University of Toronto doctoral candidate Beyhan Farhadi, and her study entitled “The Sky’s the Limit”: On the Impossible Promise of E-Learning in the Toronto District School Board, where she has taken the position that e-learning may not be appropriate for all students.  Is it e-learning or the conditions under which e-learning is taking place?  The position that e-learning, any type of e-learning, may not be appropriate for all students is the same thing as saying that classroom-based learning may not be appropriate for all students.

The truth of the matter is all students can learn effectively in any setting.  This statement does not mean that every student could have success in the specific model of e-learning currently employed by programs in Ontario, in much the same way that not all students are having success in a specific model of classroom learning or setting.  What works for one student or one group of students may not work for others – regardless of setting or context.  The real challenge for any educational program is how to design, deliver, and support learning opportunities in a way that can be effective for each student; and that will look different for various populations of students.  All that teachers, schools, school boards, and Ministries of Education can do is to ensure that the system is well resourced and professionals in that system are well trained in leveraging the affordance of all of the tools at their disposal to ensure that they can provide quality learning opportunities for all students.