Deconstructing Disengagement

This study takes place in the context of a series of Stanford xMOOCs. It;s important to note here that the xMOOC model does not rely on the use of Social Media, and, as such, may have limited application to my Research project.

Stanford xMOOCs typically take a standard approach. Students engage with video lectures, and then engage with MCQ’s, problem sets, further reading etc. It is entirely possible to complete the courses with no peer interaction, and materials are housed on lms platforms, or closed platforms.

Where my proposal focuses on the use of social media, in cMOOC contexts, where social media are the essential element for peer interaction, and the platform comprising the bulk of MOOC instruction.

This in no way undermines the paper, it’s merely a caveat in terms of applying it;s findings to a radically different type of MOOC.

“.These MOOC sare structured learning environments that emphasize instructional videos and regular assessments, centralizing activities  on a single platform.” Here a comparison with, and distinction from cMOOCs is apt. cMOOCs do not take place on one platofrm, are not centralised, or assessed regularly, if at all.

What is shared is the focus on the relatively low engagement rates in MOOCs generally. In this study the focus is on completion rates, but the participation rates for both, as a percentage of those who sign up, are similar.

The paper makes the point that “given the heterogeneity of the population” – of participants, “we would be remiss to make a priori assumptions about the approriate characteristics or behaviors around which to categorise learners” Ty]he same poiint is made about assumptions with regard to what participants might valkue in terms of outcomes, or how they make their way through the course, This would seem to hold true for cMOOCs too. The large numbers, and open aspects tend to mean that there are few, if any, uniftying variables, and participants may engage in multiple meaningful ways that bear little relation to each other.

This is a paper to return to prepare for the subgroup profiling stage. At present, it’s focus on xMOOCs makes it less useful for my current work.

Posted in Uncategorized | Leave a comment

Cognitive Architecture and Instructional Design

Sweller, J., van Merrienboer, J. G., & Paas, F. C. (1998). Cognitive Architecture and Instructional Design. Educational Psychology Review10(3), 251-296.


“Learners who have a more automated schema have more working memory capacity available”


“Novel tasks may prove to be impossible to complete until prerequisites have not only been carried pout, but  also automated”

“without automation there may be insufficient working memory capacity to even begin learning and perfgorming the new task”

This hooks up qwell with Hsiao, and Kirchner et al, and their take on the transaction cost involved in collaborative learning, as well as the sense of the the importance of prior knowledge in learning.  This ties in, also, with Ying and Yang Cognitive Processes, and the issues and problems encountered when cognitive thresholds are breached, and efficacy defaults occur.

p259 – 260 Paper makes the point that cognitive load depends on the number of elements to be processed simultaneously in working memory, and that depends on the level of element interactivity. If elements can be learned in isolation, and are not dependent on each other, interactivity, and intrinsic load is low.  Low interactivity tasks are ones where items are learned separately, and not simultaneously, and intrinsic load is low as only one item is held in working memory, and not several.


Element interactivity is subjective. That is, it can only be determnined with relation to the level of expertise a person has. Prior knowledge lowers the level of interactivity by allowing schema to be automated. Novelty increases it.

High element interactivity tasks, for example, networked learning for a novice, will tend to be high instrinsic load.


automation frees working memory for learning, and mastering the novel aspects may be a prerequisite for learning. This speaks directly to the creation of something like  a Social Media Toolkit to deliver the ability to automate schema for Digital Literacy Novices.

Posted in Uncategorized | Leave a comment

Research Proposal Literature Review


This Literature Review broadly comprises three strands. Cognitive Load Theory and it’s relationship to collaborative learning comprise one strand. Self Efficacy theory, in the context of cognitive overload comprises a second, and potential Cognitive Load implications of Connectivist MOOC practice comprises the third.

The literature review is written in support of the deployment of a Social media toolkit as a preparatory resource in Connectivist MOOC contexts.

The research questions are detailed below

Primary Research Question

Can a Social Media Toolkit, used as a Preparatory Resource address technical and digital literacy shortfalls amongst technology novices in cMOOCs?


How will a Toolkit that targets Digital Literacy shortfalls affect novice motivation and participation in a cMOOC environment?

Will a scaffolding resource that addresses prior knowledge gaps in Social Media lead to increased participation?

Will the Toolkit lead to decreased, reported cognitive load amongst Social Media novices?

Milligan et al (2013) note that due to the novelty of Connectivism, there  “a small amount of empirical research, published in niche journals and peer reviewed conferences,” “supplemented by a large body of more anecdotal and reflective research published outside the traditional peer-reviewed journal system.” There is very little literature detailing self-efficacy concerns and Connectivist MOOCs, and very little empirical literature. To that end, there is something of a reliance on non Connectivist analyses of Self-Efficacy, cognitive load and collaborative concerns.

I focus on empirical and conventional literature, and avoid a reliance on the more informal tradition of Connectivist literature, in part because of it’s disseminated and uncatalogued nature, and, in part, because it can be difficult to form an overview when material is so diversely distributed.

My focus on Cognitive Load theory comes from seminal work by John Sweller, and subsequent developments of that work in the direction of collaborative Cognitive Load experiences. I have also focused on recent work tracing theoretical relationships between Cognitive Load, motivation and persistence, and Self-Efficacy.

Connectivist literature, and MOOC literature is confined to recent work. Connectivism itself is a new theory, and Connectivist MOOCs data from circa 2008.

Cognitive Load Theory (CLT)

Kirschner et al summarise aspects of CLT by saying that “individual learning depends on the limited processing capacity of the learner’s cognitive architecture and the cognitive load imposed by a task”  (Kirschner, Paas, Kirschner, 2010, p615-616). This limited processing capacity is a function of working memory which is able to deal with two to three items elements simultaneously (Sweller, van Merrienboer, & Paad, 1998, p. 259). Knowledge construction is a function of working memory. In order for skills, processes and knowledge to become automated and lodged in long term memory as schema they must first be processed in working memory. Such automation frees working memory resources, allowing complicated skills to be deployed with minimal cognitive effort. (Sweller, van Merrienboer, & Paad, 1998, pp. 256-257).

Intrinsic Cognitive Load

Intrinsic Cognitive Load is determined by levels of element interactivity – the number of elements to be learned, and whether these elements are dependent on each other. Elements which are dependent, and learned simultaneously levy high demands on working memory, and have high cognitive load. Independent elements which may be learned separately levy lower demands on working memory. Intrinsic load may be decreased by increasing expertise, or lowering element interactivity (Sweller, van Merrienboer, & Paad, 1998, pp. 259-260)

Task complexity and element interactivity are suggested to be subjective  (Kirschner, Paas, & Kirschner, 2008, p. 36) The same task may represent low levels of task complexity for a subject expert, and high levels for a novice.  Individuals with automated schema and expertise will have lower cognitive demands and more resources available for learning. (Sweller, van Merrienboer, & Paad, 1998, pp. 261-262).

We can assert that novices in an area of study are likely to have higher cognitive loads that those with expertise.

Cognitive Load in collaborative contexts

Several papers put forward the premise that multiple working memories may be shared amongst groups (Kirschner, Paas, & Kirschner, 2008) (Hsiao, Kester, & Sloep, 2013; Kirschner, Paas, & Kirschner, 2011). Working memories may be added together to  collaboratively solve complex problems that might be beyond the capacity of a single working memory. (Kirschner, Paas, & Kirschner, 2011, p. 616) (Kirschner, Paas, & Kirschner, 2008, pp. 36-37)

CLT  proposes a transaction cost inherent in collaborative learning. The transaction cost is the cognitive load involved in organising collaborative learning amongst a group of learners. That may involve the organising of knowledge sharing, group activity organisation, deciding who should pursue what knowledge etc.  (Kirschner, Paas, & Kirschner, 2011, p. 621)

This transaction cost can increase cognitive load as students need to “invest mental effort in the communication of information and the coordination of actions” in order to achieve collaboration. Consequently individual learners do better on tasks with low element interactivity, as the cognitive load cost of collaborating is far greater than the benefits. Groups do better on tasks with high element interactivity, as the benefits of collaborating outweigh the organisational cognitive load costs involved. (Kirschner, Paas, & Kirschner, 2011, pp. 621-622)

Kirschner et al admit to several limitations to their study. Their experiment was closed, and they minimised cognitive load incurred from off-task processes which may not mirror real life contexts. They also suggest that real world collaborations comprise heterogeneous  groups, in terms of expertise and knowledge, and, as previously noted, element interactivity is subjective (Kirschner, Paas, & Kirschner, 2011, pp. 622-623) making accurate measurement difficult. For my purposes, an additional caveat is that the experiment takes place in face to face formal learning environments, which bear little relationship to the collaborative mechanisms that characterise online Connectivist collaborations. Hsiao et apply this work to heterogeneous open groups of online and informal learners, which bears a closer relationship to Connectivist learning environments.

Cognitive Load and collaboration in open, online, informal groups.


Hsiao et al locate their research into collaborative learning and cognitive load in the context of Learning Networks and nonformal learning. They describe nonformal learning as “intentional learning based on personalized learning goals, exempt from externally imposed evaluation criteria and institutional supervision” and learning networks as “online, social network that is designed to support non-formal learning in a particular domain’’. “Learners have to take responsibilities to organize their own learning activities and to acquire knowledge from others to achieve their learning goals”. (Hsiao, Kester, & Sloep, 2013, p. 69)

Learning Networks and Connectivism, a comparison.

These definitions are similar to Connectivist contexts. Kop describes Connectivist learning as characterised by learners having to “set their own learning goals, find resources, and try out new tools and make them work.” The autonomous, instructor-independent aspect of Connectivism Kop admits “could be problematic”. (Kop R. , 2011) Downes has noted that assessment and formal evaluation is not a design concern for Connectivist MOOCS, and that their aim is, partially, to exist outside of traditional institutional practice. (Downes, Connectivism and its Critics: What Connectivism Is, 2008) Numerous commentators note the online and open characteristics of Connectivist MOOCS as essential and defining characteristics. Connectivist learners are expected to navigate the networks of online participants, in a self-directed manner, organising their own learning, activities and interactions. (Downes, 2008) (Downes, 2011) (Kop, Fournier, & Mak, 2011) (Siemens, 2004).

The case for higher cognitive loads in Learning Network Collaborations


Hsiao, quoting Sweller (Sweller, van Merrienboer, & Paad, 1998) argues that in Learning Networks, novel information must be attended to in order to proceed to long term memory. (Hsiao, Kester, & Sloep, 2013, p. 90) Sweller et al further argue that the processing of novel information may be a prerequisite for proceeding to more detailed information or knowledge. (Sweller, van Merrienboer, & Paad, 1998, p. 258) It certainly appears to be a prerequisite for forming schema, and committing novel information to long term memory, and achieving automation. (Sweller, van Merrienboer, & Paad, 1998, pp. 256-257). This novel information may be of several types.

In learning networks, there are no barriers to entry. Learner, backgrounds, expertise and motivations are likely to be diverse, this heterogeneity can add extraneous Cognitive Load, as the activities of finding other participants to interact with, and determining who has suitable knowledge in highly diverse contexts are likely to levy Cognitive Load demands and decrease Working Memory resources, and “detract from learning rather than increase it” (Hsiao, Kester, & Sloep, 2013, p. 91). It seems clear that novices who are unfamiliar with collaborative techniques and technologies may find themselves lacking in such prerequisites, and may suffer amplified cognitive loads.

Hsiao et al identify three facets of collaborative learning in learning networks that can detract cognitive resources from knowledge construction.

  1. Finding collaborators (and finding useful or apt collaborators).
  2. The challenges of online as opposed to face to face collaboration.
  3. Lack of knowledge with regard to collaborating and knowledge sharing skills.

These three characteristic difficulties would seem descriptive of potential cognitive load issues in Connectivist environment, specifically for novices.

Specifically, it may be the case that Hsiao’s three factors describe the cognitive load concerns for digital literacy novices in Connectivist MOOCs.

How this analysis informs the Research questions

It is theorised, in the research questions, that there is a cognitive load concern for Social Media novices in Connectivist MOOCs. The work of Hsiao et al, Kirscher et al, and Sweller et al seems to support the thesis that in contexts where participants are unfamiliar with new technologies, pedgagogies, and with the specific techniques of collaboration (in this case across social media) the cognitive loads can be expected to be significant, and that this load may well present barriers to participation and learning.

It seems reasonable to hypothesise that measures which would reduce such cognitive load, by allowing participants to develop the relevant schema would help reduce such loads. The Social media toolkit is an attempt to allow participants to master the novel skills and techniques which are a prerequisite to Connectivist learning, as Sweller et al and Hsiao et al both note and suggest.

The next section will attempt to support the supposition that a lack of digital literacy skills is a participation inhibitor amongst MOOC participants, with reference to Connectivist literature.

Barriers to participation in Connectivist MOOCs

As noted previously, Learning Networks and Connectivist contexts bear similarities in how they transpire, are designed, and occur. In addition, they may share participation barriers.

A lack of digital or critical literacies is often suggested as a barrier to MOOC participation. In the interests of clarity I offer a definition of digital literacy

“the ability to use information and communication technologies to find, understand, evaluate, create and communicate digital information, an ability that requires both cognitive and technical skills” (Office for Information Technology Policy’s Digital Literacy Task Force, 2013)

Digital Literacy, and the three Cognitive Load increasing facets of collaborative learning posited by Hsiao et al (see above)  seem to be essentially similar.

In Connectivism, Digital Literacies often transpire across social media (Downes, Connectivism and its Critics: What Connectivism Is, 2008) but they may also transpire across Moodles, LMS platforms, Youtube and wikis (Waite, Mackness, Roberts, & Lovegrove, 2013) or numerous and diverse platforms.

Waite et al narrate novice MOOC experience thus “novices felt initially overwhelmed by technical issues, multiple channels, and a need to be able to multitask, which required too many initial participatory skills” (Waite, Mackness, Roberts, & Lovegrove, 2013) If critical literacies are lacking there is a risk of lack of participation. Some learners found the abundance of information (a characteristic of Connectivist MOOC design) to be overwhelming. (Waite, Mackness, Roberts, & Lovegrove, 2013)

Connectivist commentators suggest that digital literacies are considered a prerequisite for engaging in courses, (Milligan, Littlejohn, & Margaryan, 2013) (Fini, 2009) (Kop R. , 2012, p. 3). McAuley et al note that “lack of familiarity with the digital skills privileged and rewarded within the MOOC will limit participation” (McAuley, Stewart, Siemens, & Cormier, 2010, p. 56), and yet, there is typically little support for the development of such literacies. Some commentators seem to expressly rule out technical or pedagogical support for novices -“we expect students to be able to manage complex and rapidly changing environment – in other words, to be able to manage through just the sort of chaos we are creating”. (Downes, Connectivism and its Critics: What Connectivism Is, 2008)

McAuley et al argue that “MOOCs – like most digital communities and networks – operate on the assumption that people have contributions to make and know how to make them in an appropriate manner” and suggest that novices may not find the support necessary to develop digital literacies and cope with the course technology and pedagogy, as such structures are not formalised in Connectivist MOOC courses. (McAuley, Stewart, Siemens, & Cormier, 2010, p. 56)

One Connectivist commentator does specify a requirement for novice scaffolding in digital literacy “scaffolding is necessary to build confidence and self-efficacy and to ensure novices will feel confident and competent in using technologies and are supported throughout the course.” (Kop, Fournier, & Mak, A pedagogy of abundance or a pedagogy to support human beings? Participant support on massive open online courses, 2011)

It seems likely that the assumption of digital literacy amongst participants by course organisers, and the lack of support for developing such literacies are likely to lead to increased cognitive load for novice participants and higher transaction costs in attempting to work collaboratively. Assumptions that individuals are likely to have the requisite digital literacies seem problematic at best, and may well be misplaced in some cases. Recent studies, for example, on the digital native meme are indicating that the digital native generation may not be as digitally literate as supposed. (Margaryan, Littlejohn, & Vojt, 2011)

How this relates to the Research question

In the above section, I attempt to indicate that there is an identified issue with digital literacy gaps in Connectivist MOOCs. The case is not to argue that the lack is universal, but that participants characterised by Digital Literacy unfamiliarity face definite barriers to participation and engagement. The primary research question asserts that there is a digital literacy gap lack amongst novices, and Connectivist commentators seem to agree, and claim it as a barrier to participation. It would seem that a Social Media Toolkit would be a good candidate for addressing this gap, and monitoring the possible effects on motivation and participation as a consequence of addressing such a gap is the stated aim of the research.

Assumptions of Digital Literacy, the case against.

Fried notes that where laptop use in lectures is unstructured – not guided by instruction – laptop using students spent an average of 25% of lecture time engaged in off-topic activities (Fried, 2008, p. 910), self-report understanding less of lectures, and perform lower than a non-laptop using control group on standardised testing (Fried, 2008, pp. 910-911). Fried puts forth an argument that unstructured and unguided technology use may, in fact, be damaging to cognitive load management and to learning, where distraction and tangential activites occur.

A recent study of 74 Finnish first year trainee teachers used a limited range of technologies in designing ICT inflected lessons, entailed minimal social media usage, and used social media merely as a passive information source rather than for active knowledge creation, further noting that “typically these student teachers will not necessarily be the ones to invent new ways of using technology in their work, but rather prefer to wait for other users’ comments and ideas before adopting” (Valtonen, Pontinen, Kukkonen, Dillon, Vaisanen, & Hacklin, 2011, p. 14). It must be noted here, that Valtonen et al focus on the first year of a teacher training program in a rural university. Although their sample is significant (74) and representative of Finnish youth, the fact that students have not yet completed their first year of training may have had some impact on their lesson design choices. It is also unclear if we can take a small rural university to be a representative sample of digital native youth as a whole.

Margaryan et al  assert, in a study of two UK universities that, regarding assertions about digital natives learning styles and multitasking cognitive abilities that , “there is no empirical basis to them” (Margaryan, Littlejohn, & Vojt, 2011, p. 429)  and find that large numbers of students in their study (over 80%) never used blogs, podcasts or chat in formal or informal learning (Margaryan, Littlejohn, & Vojt, 2011, p. 434) and that “Students did not appear to understand the potential of technology to support learning. Instead they looked to their lecturers for ideas on technology enhanced learning” and suggest students had largely conventional expectations with regard to their learning (Margaryan, Littlejohn, & Vojt, 2011, p. 436) They argue that the primary motivating factor in determining the digital literacy levels of students in their educational use of technology is not age, but lecturer and institutional use of technology, and prior educational experience. (Margaryan, Littlejohn, & Vojt, 2011, p. 438)

A potential weakness in Margaryan et al may be that the above analysis relies on a small sample ( 8 students selected for interview), and, although the quantitative analysis is significant (160 students responded) the conclusions with regard to technology use and preferences were drawn from the smaller sample.

ECARS’ recent survey of over 100’00 students in 14 countries seems to lend some support to theses positing a lack of educational digital literacy amongst students. While noting that students are generally confident in their use of educational technology (Dahlstrom, Walker, & Dziuban, 2013, p. p10) they found that students tend to use limited and conventional websites and technologies in their learning (Dahlstrom, Walker, & Dziuban, 2013, p. p11). In addition, a majority of students – almost three quarters – had no idea what a MOOC was, and those who did expressed a preference for blended learning environments while testing out MOOC experiences (Dahlstrom, Walker, & Dziuban, 2013, p. 15). ECAR’S study is a measure of student preferences, and, while the sample is large (over 100’000) it is perhaps wise to note that the survey does not monitor actual technology use, merely reported use and preferences.

Regardless of individual study weaknesses, when taken in tandem with the barriers to participation noted by Connectivist commentators, the assumption that a proportion of MOOC participants will not have the requisite literacies, nor the tools to develop them seems supported. We can, it seem, assert that there is likely to be a class of novice participant for whom Connectivist MOOCs will present challenges they may not be prepared for, and will experience potentially troublesome levels of cognitive load in attempting to negotiate. It would seem wise to assume that there is a constituency for whom a Social Media toolkit, informing them of both technical and pedagogical uses of Social Media in a structured way might be useful.


Self-Efficacy and Cognitive Load

Self-efficacy is the self-judgement of one’s ability to achieve, and a sense that tasks are achievable, (Bandura, 1982, p122).Bandura’s ideas regarding self-percepts of efficacy as a major predictor of student success, and the role that high cognitive loads can have on individual persistence, motivation, effort, and learning success (Bandura 1977, 1982) are key points in understanding the relationship between digital literacy novices and participation barriers in Connectivist MOOCs.

The relationship between Self-efficacy, difficulties, and persistence is well documented. Bandura suggests   “Those who have a stronger sense of efficacy exert greater effort to master the challenges” and “when beset with difficulties people who entertain serious doubts about their capabilities slacken their efforts or give up altogether”. (Bandura, Self-Efficacy Mechanism in Human Agency, 1982, p. 123) Lowered self-efficacy may result in task avoidance or lack of success. People with low self-efficacy tend to “shun or fail those (tasks) that exceed their perceived coping abilities.” (Bandura, Self-Efficacy Mechanism in Human Agency, 1982, p. 126).

Clark argues that “perceived difficulty is primarily (but not entirely) a function of cognitive load” (Clark, 1999, p. 6) and in contexts where we are challenged to process increasing amounts of data, increased mental effort is necessary (Clark, 1999, p6) and tasks with high cognitive load tend to produce low percepts of self-efficacy in students, as a function of the large amount of new knowledge which will need to be mastered (Clark, 1999, p9)

We can argue here that the transaction costs for novices in Connectivist MOOCs are likely to incur such loads, and possibly reduce self-efficacy amongst those with undeveloped schema for digital literacy, reducing effort, persistance and resistance to difficulties, such as higher transaction costs in collaboration, or feelings of being overwhelmed.

High Cognitive Load, initially, may be beneficial, but beyond a certain point, is likely to undermine a student’s sense of the achievability of a course of action or task (Clark, 1999, p6). At levels of too high or too low cognitive load, participants are likely to deploy less effort, be less motivated, and engage in unhelpful or undesirable behaviours. Clark suggests that “Cognitive Overload” – where “cognitive load exceeds working memory capacity” (Clark, 1999, p10) – may occur where the novelty of a task is too high.  When a task is “perceived as impossible, self-efficacy issues lead us to avoid the goal at hand” (Clark, 1999, p10) “an automated “default” occurs that forces learners away from the immediate learning goal and towards novel or different performance goals” ” (Clark, 1999, p11). Clark states this efficacy default results in undesirable and unhelpful behaviours – abandonment of learning, distracting thoughts and goals, mistake making, reversion to inadequate learning strategies, or a state or sense of helplessness. (Clark, 1999, p16).

How this relates to the Research questions.

The Research Questions imply a possible link between cognitive load and motivation, participation and motivation in Connectivist MOOCs. Although a detailed analysis of motivation and cognitive load is beyond the remit of this review, it is hoped that the discussion provides a theoretical basis for investigating the possible effects of lowering cognitive loads, and transaction costs for novices. Analysis of tis will be based on monitoring participant engagement across the course quantitatively – measuring tweets, numbers of posts made, number of seminars viewed, and by a qualitative analysis based on questionnaires and interviews.


High Cognitive Load, transaction cost, and lowered motivation seem likely to be the case for at least some MOOC participants, specifically, Digital Literacy novices.  As noted previously, Connectivist literature does draw a link between a lack of requisite digital literacy, and feelings of confusion, being overwhelmed, and lack of confidence in one’s ability to participate, to a degree where such a literacy lack may be an actual barrier to participation. It seems likely that assumptions with regard to levels of digital literacy amongst participants may lead to contexts where digital literacy novices will experience heightened levels of cognitive load, and may consequently suffer confidence, performance and participation disadvantages.

It seems likely that a possible solution is to engage a preparatory resource (in this case the Social Media Toolkit), with the aim of inculcating such digital literacies, and lowering cognitive load and transaction cost amongst participants, freeing up working memory to engage with the transaction costs involved in collaborative learning in open, nonformal learning networks, and allowing participants to leverage the collaborative and collective working memory benefits that can make collaborating on complex tasks a beneficial exercise.

It is the aim of this research to deploy such a resource, and to measure it’s effectiveness, both quantitatively (in terms of overall participants, and participant activity across social media and MOOC platforms) and qualitatively, by drilling down into participant experiences via questionnaires, and interviews.


It is intended to submit the resulting paper to The International Review of Research in Open and Distance Learning for consideration The IRRODL has a history of interest in, and publication of Connectivist and Connectivist related research and papers, and would seem a good fit, both in terms of topic, and in terms of audience. Currently, the Research Articles sublission section details compatible word length requirements with the DIT guidelines (3’500 – 5’000 words).

The journal is associated with Athabasca University, which, as both a distance learning University, and a University with links to numerous Connectvist commentators and publisers, seems well positioned to both publish, and provide expert peer review.

The journal is peer reviewed, and is published on an open access policy, with all material licenced under Creative Commons.

Information Managment Strategy

At present, the full breadth of the literature for the project is contained in an annotated bibliography housed on this blog. Currently over 40 articles have been annotated. On a more longterm basis, literature annotations may be moved to diigo, or, both forms of scoial bookmarking. A blog is suitable at present, however, as the depth and breadth of reading and annotation is suited to a blog format,  the blog is easily accessible from the multiple device types I use, and is integrated into the browsers on those devices. As blogging is also a focus of the research project, it is additionally intended to use this blog as a part of the course of instruction the research will take place in.

Target User Group

The target usewr group are educational professionals, of all levels. Given that cMOOCs are open access, with no barriers to entryit is not expected to be able to control the main body of participants. Participants may be geographically distributed, and their practices may be at all levels, and of all types. The main MOOC will be hetergenuous. The Social Media Toolkit  however, the actual foicus of the research, will target digital media novices. Educators who self assess as having low levels of competence, or confidence with social media, and/or with social media based pedagogies.

Project Plan


It is envisaged that quantitative data collection (Twitter posts, blog posts, seminar attendance) will be automated, and collation should also be automated. All posts will be tied to unique identifiers (Twitter handles, gmail addresses, blog addresses) and can be collated automatically. This will be uswed to identify potential subgroups, in terms of Social Media participation, to be atargetted with questionnaires.

This data will be available as soon as the main MOOC ends.

Qualitative data will be collected at point of exit from the Social Media toolkit, as a questionnaire, and post MOOC questionnaires will be completed by late Febraury. Collation of of quantitative data (primarily questionnaires, with minimal, or possibly no interview data) will be completed by Mid/late March.

There will be an 8 week window for paper completion. It is envisaged that an analysis of the quantitative data as part of the paper will already be significantly underway by this stage, as will the literature view, and aspects of the theoretical framweork underpinning the results from quantitaive analysis.

Works Cited

Bandura, A. (1977). Self-Efficacy; Toward a unifying Theory of Behavioral Change. Psychological Review, 84(2), 191-215.

Bandura, A. (1982). Self-Efficacy Mechanism in Human Agency. American Psychologist, 27(2), 122-147.

Clark, R. E. (1999). Yin and yang cognitive motivational processes operating in multimedia learning environments. In J. van Merrienboer, Cognition and multimedia design. Herleen, Netherlands: Open University Press.

Dahlstrom, E., Walker, J. D., & Dziuban, C. (2013). ECAR study of Undergraduate Students and Information Technology. Educause Center for Analysis and Research.

Downes, S. (2008, October). Connectivism and its Critics: What Connectivism Is. Retrieved April 12, 2013, from Stephen’s Web:

Downes, S. (2011). Connectivism and Connective Knowledge. . Retrieved April 26, 2013, from Stephen’s Web:

Fini, A. (2009). The technological dimension of a massive open online course: The case of the CCK08 course tools. The International Review of Research in Open and Distance Learning, 10(5).

Fried, C. B. (2008). In Class Laptop Use and it’s Effect on Learning. Computers & Education, 50(2), 906-14.

Hsiao, Y. P., Kester, L., & Sloep, P. (2013). Cognitive Load and Knowledge Sharing in Learning Networks. Interactive Learning Environments, 21(1), 89-100.

Kirschner, F., Paas, F., & Kirschner, P. A. (2008). A Cognitive Load Approach to Collaborative Learning: United Brains for Complex Tasks. Educational Psychology Review, 31-42.

Kirschner, F., Paas, F., & Kirschner, P. A. (2011). Task Complexity as a Driver for Collaborative Learning Efficiency: The Collective Working-Memory Effect. Applied Cognitive Psychology, 25, 615-624.

Kop, R. (2011). The challenges to connectivist learning on open online networks: Learning experiences during a massive open online course. The International Review Of Research In Open And Distance Learning, 12(3), 19-32.

Kop, R. (2012). The Unexpected Connection: Serendipity and Human Mediation in Networked Learning. Educational Technology and Society, 15(2), 2-11.

Kop, R., Fournier, H., & Mak, J. (2011). A pedagogy of abundance or a pedagogy to support human beings? Participant support on massive open online courses. The International Review Of Research In Open And Distance Learning, 12(7), 19-38.

Margaryan, A., Littlejohn, A., & Vojt, G. (2011). Are Digital Natives a myth or a reality? University student’s use of digital technologies. Computers & Education, 56(2), 429-440.

McAuley, A., Stewart, B., Siemens, S., & Cormier, D. (2010). The MOOC Model for Digital Practice.

Milligan, C., Littlejohn, A., & Margaryan, A. (2013). Patterns of Engagement in Connectivist MOOCS. Journal of Online Learning and Teaching.

Office for Information Technology Policy’s Digital Literacy Task Force. (2013). Digital Literacy, Libraries, and Public Policy. Washington D.C.: Office for Information Technology Policy’s Digital Literacy Task Force.

Siemens, G. (2004). Elearnspace. Connectivism: A Learning Theory for the Digital Age. Retrieved May 15, 2013, from Elearnspace:

Sweller, J., van Merrienboer, J. G., & Paad, F. (1998). Cognitive Arictecture and Instrustional Design. Educational Psychology Review, 251-296.

Valtonen, T., Pontinen, S., Kukkonen, J., Dillon, P., Vaisanen, P., & Hacklin, S. (2011). Confronting the Technological Pedagogical Knowledge of Finnish Net Generation Student Teachers. Technology, Pedagogy and Education, 20(1), 3-18.

Waite, M., Mackness, J., Roberts, G., & Lovegrove, E. (2013). Liminal Participants and Skilled Orienteers: Learner Participation in a MOOC for New Lecturers. Journal of online Teaching and Training, 9(2).

Posted in Uncategorized | Leave a comment

Task complexity as a driver for collaborative learning efficiency: The collective working-memory effect

Kirschner, F., Paas, F., & Kirschner, P. A. (2011). Task complexity as a driver for collaborative learning efficiency: The collective working-memory effect. Applied Cognitive Psychology25(4), 615-624. doi:10.1002/acp.1730

This is an experimental confirmation of the expected task comolexity effect with regard to cognitive load. Specifically, high task complexity can be shared advantageously amongst multiple working memories, and collaboration proves to be beneficial, but with low task complexity, the cognitive load incolved in collaborating is too complex. Specifying the exact levels at which these effects occur is not necessarily easy – groups will be heterogenuous, their levels of expertise, and therefore element interactivity and resultant cognitive load will vary, making it difficult to measure.

Additionally, collaborations are messy and difficult things, and real world groups go off piste, off task and take tangents.

This speaks to the resource constuction. The simple tasks are solo, the complex tasks are shared. But the paper takes place in a formal, standardised, and face to face setting, and not an informal, unstructured, online setting. Hsiao et al are a better fit for Connectivist Contexts.

p622 Kirschner et al note that, outside of experimental contexts, where collaboration/communication can bbe kept on task, transaction costs – the cognitive load costs involved in communication to achieve collaboration – may be higher, due to off task communication.

They also note task complexity ( and hence cognitive load) is related to the person’s subject expertise, and varies from person to person. This, of cousre, makes it difficult to specify the task complexity for a group colaborating, and makes accurate predictions of tsak complexity, and related efficiencies difficult, if not perhaps functionally imnpossible. (p 622-623)

p622 Their research finds that “learners profited from having learned from high-complexity tasks in collaboration, while learners profited from having learned from low-complexity tasks individually”

It must be noted, however, that there may still be collaboration issues that are yet to be resolved. For example, to qwhat degree do varying levels of expertise within a group effect learning efficiencies, and cognitive load. There is possibly an increased load in assessing peer expertise, and peer need, as Hsiao have noted, but varying levels of expertise may also provide positive effects for some students ( the benefits of peer learning, and peer teaching).

Posted in Uncategorized | Leave a comment

A Cognitive Load Approach to Collaborative Learning: United Brains for Complex Tasks

Kirschner, F., Paas, F., & Kirschner, P. A. (2009). A Cognitive Load Approach to Collaborative Learning: United Brains for Complex Tasks.Educational Psychology Review21(1), 31-42. doi:10.1007/s10648-008-9095-2

“Intrinsic load provides a “base” load that is irreducible other than by constructing additional schemas and automating previously acquired schemas—in other words, by an increase in expertise or by deconstructing the task so that less elements interact (see Ayres 2006; Pollock et al. 2002).”

This is, in part, what the Toolkit is intended to achieve. The number of interactiogn elements ( achieveing requisite technical and pedagogic literacy, soucing knoweldgeable peers, communicating with those peers, interprewting online communications) is quite high for novices. Lowering that complexity is a way of lowering the intrinsic load, by both allowing people to acquire basic expertises, and by splitting up the  work of acclimatisation so that the tasks are now separate, but related, as opposed to occuring all at once. This refers to p36.

The Group as Information Processing System

“it could be argued that assigning high complexity tasks to groups of learners allows
information to be divided across a larger reservoir of cognitive capacity and might result in
more effective and efficient learning than assigning them to an individual learner” as Hsiao et all note, the reality may be more complex – quote their provisos here.

Kirschner et al also argue that task complexity is key in whetehr collaboration will facilitate leveraging group working memory for more efficient learning. With simple tasks, it seems the load involved in collaboration may outweigh the advantages, and efficiency decreases. With complex tasks, where collaboration is structured, the cog load incolved in colaborating is minimal in comparison to the task complexity, and so, collaboration may aid learning efficiency here.

Again, Hsiao et al provisos are apt to quite.

On p39, authors assert that cognitive load has a pivotal role to play in determining factor in determining whether a learning task should be collaborative or individual ( i.e. the intrinsic load) but ultimately the interplay bertween social, cognitive and motivational factors will need to be investitaed, with cognitive load as an aspect.

Hsiao et al note additional aspects of the transaction cost of collaboration specific to online engagement. The lack of face to face cues, the need to learn how to use the tools for online collaboration, and the etiquette and procedures for so doing – finding out who is knowledgeable, who isn;t and how that fits in with your context.

This seems to dovetail with the difficulties several other commentators describe in MOOCS.

Posted in Uncategorized | Leave a comment

Cognitive load and knowledge sharing in Learning Networks, Interactive Learning Environments

Ya P. Hsiao , Francis Brouns , Liesbeth Kester & Peter Sloep (2013) Cognitive
load and knowledge sharing in Learning Networks, Interactive Learning Environments, 21:1, 89-100,
DOI: 10.1080/10494820.2010.548068

Hsiao et al adopt a definition of non formal learning

“Nonformal learning is here defined as intentional learning based on personalized learning goals, exempt from externally imposed evaluation criteria and institutional supervision (Livingstone, 1999; Van Merrienboer, Kirschner, Paas, Sloep, & Caniels, 2009)”

Livingstone, D.W. (1999). Exploring the icebergs of adult learning: Findings of the first
Canadian survey of informal learning practices. The Canadian Journal for the Study of
Adult Education, 13, 49–72.

Van Merrie¨nboer, J., Kirschner, P., Paas, F., Sloep, P., & Canie¨ls, M. (2009). Towards an
integrated approach for research on lifelong learning. Educational Technology Magazine,
49, 3–15.

and a definition of Learning networks that’s comparable to Couros’s PLNs.

‘‘online, social network that is designed to support non-formal learning in a particular domain’’ (Kester et al., 2006a; Sloep, 2009; Sloep et al., 2007). Learners have to take responsibilities to organize their own learning activities and to acquire knowledge from others to achieve their learning goals (Kester et al., 2007).”

Sloep, P. (2009, 11–13 June). Fostering sociability in learning networks through ad-hoc transient communities. Paper presented at the Computer-Mediated Social Networking Conference, Dunedin, New Zealand.

Sloep, P., Kester, L., Brouns, F., Van Rosmalen, P., De Vries, F., De Croock, M., & Koper,
R., (2007, 14–16 March). Ad hoc transient communities to enhance social interaction and
spread tutor responsibilities. Paper presented at the The Sixth IASTED International
Conference Web-Based Education, Chamonix, France.

This link is made explicit , as the Connectivist Literature is minimal, and little work on cMOOCs, Connectivism and Cognitive Load has been done. The identification of both non-formal learning, which fits with Connectivist concerns with non formal networks, and Learning Networks, whose architecture reflects the networked architecture of Connectivism, is useful.

“During knowledge sharing in LNs, collaboration is a means to achieve reciprocal
understanding between learners and to construct knowledge by performing activities
through interaction with others.” p90. Here is another explicit parallell.

The Introduction argues that s, for some participants in collaborative knowledge construction in these informal learning networks, some scaffolding and structure is needed.

The paper focuses on a Peer Support System that dirests peers to useful peers based on knowledge requests. So, it;s focus is not exactly inline with mine.

Page 90 has a basic outline of Cognitive Load knoweldge construction, the focus of my resuearch. It argues that ”

Learning takes place via schema construction, elaboration, and
automation stored in long-term memory. During the learning process, novel
information must be first attended to and processed by working memory before it
can be stored in the long-term memory (Sweller, Van Merrie¨nboer, & Paas, 1998).”

Furthermore, working memory has limitations,, and can process only two or three novel items at a time.

Intrinsic, germane and extraneous load, when added together, need to remain within the limits of working memory processing power if learning is to be effective.

The upshot here is that cognitive load is limited, novices will need to master basic techniques, ideas, strategies and practices before they can continue, and bottlenecks in cognitve processing are the key issue.

(Paas et al., 2003a; Sweller
et al., 1998) (Van Gog & Paas, 2008; Van Merrie¨nboer & Sweller, 2005)

In Learning domains where there are no entry requirements, heterogeneous groups tend to be formed, where peoiple have different lkevels of k=subject knowledge, backgrounds, competence levels and purposes ( summary of Hsiao, p92)Hsiao et al assert that this can add exraneous Cognitive Load, as the activities of finding other particiapnts to interact with, determining who has suitable knowledge are likely to do this, and “detract from learning rather than increase it” P91.

To sum up, without support that deals with heterogeneous group composition
and online communication, knowledge sharing in unstructired LNs imposes additional
extraneous load. It does so because extra cognitive resources have to be devoted
to finding a suitable collaborator and finding out how best to communicate with
others online, and managing the unfamiliar tools, reducing the available resources for processing the information and knowledge that is available.

Task Complexity and Cognitive Load

“Non formal learners choose their own learning actions” in contrast to the structured task design of formal learning. “Collaboration is initiated ad hoc because of the perceived complexity of the learning “task” – this is a description of Connectivist Learning.

Cog Load theory does support collaborative learning, in specific and circumscribed ways. Learners, seeking to simplify complext tasks might seek to lighten their load through knowledge sharing, and collaborative help. See Kirschner et al.

“According to this argument, the created joint working memory has
more processing capacity and can therefore deal with more complex learning actions
than can each individual working memory. However, these cognitive benefits can
only work well if collaborating learners know how to share knowledge with each
other. As argued, without support learners, LNs need to allocate extra cognitive
resources to organizing and maintaining knowledge sharing.” p93

In short, extra learning may be triggered by the activities of explaining and elaborating, and a group can be considered as sharing working memories, and spreading intrinsic load, freeing up cognitive resources for learning. But, this is only the case if, as noted the sharing is structiured. If unstructured, the sharing is likely to increase cognitive load.

Hsiao et al identify three facets of the networkeed lerning experience that can detract cognitive resources from knowledge construction.

Finding collaborators (and finding useful or apt collaborators), the specific rigours of online as opposed to face to face collaboration, and lack of knowledge with regard to collaborating and knowledge sharing.

These both increase cognitive load, and undermine the potential cognitive load advantages specific to collaboration.

Posted in Uncategorized | Leave a comment

Implementing Web 2.0 in Secondary Schools: Impacts, Barriers and Issues

Crook, C., Fisher, T., Graber, R., Harrison, C., Lewin, C., Cummings, J., Logan, K., Luckin, R., Oliver, M., & Sharples, M. (n.d.). Implementing Web 2.0 in Secondary Schools: Impacts, Barriers and Issues. Retrieved from BECTA website:

I focus on the barriers section, as that seems most in line with my research concerns.

“Some teachers see great potential, and are enthusiastic proponents –
58.5% believe that popular Web 2.0 resources should get more use in the

“Only 26% of teachers had used a
social networking site in the last 24 hours (14.3% in the last week), 71.1% had never
written or edited a blog, and 43.4% had never used instant messaging” “Dated 2008 – which, in some sense. limits the applicability of this statistic, especially for my research, detailing twitter as it does (founded 2006, and breaking through in 2009/2010. That said, surveys consistently show that Twitter has amongst the smallest uptake in terms of Social media amongst educators of the main social media sites.


“In some schools, insufficient levels of technical support including specialist support for the Web 2.0 tools is still a barrier to staff uptake” p 64

“A crucial driving factor was a sense of community that had been stimulated by
noticing the innovation of other teachers and, sometimes, having one’s own
innovation brought into view for them. In short, the collaborative and publishing tools
of Web 2.0 serve not only as the content of innovation but also the medium in which
that innovation is exchanged, noticed and rewarded.”

The HEA report calls for the creation of a network od digital curators amongst educators. Laurillard calls for peer networks (as does Conole) as a necessary part of encouraging, supporting and disseminating innovative practice, and use of technology and social media amongst educators.

“The informants make it clear that becoming a member of a community of practice
can be crucial in increasing the awareness of possibilities. One innovator relates this
to his personal history:” p71

Both these quotes loan themselves to supporting the Connectivist/Couros idea of a PLN. What’s missing, however, from all these commentaries is data supporting the calls for PLN development. Is a peer network, at least initially, the key? Or is it, as Kirschner et al would have it, something that is necessary when a certain level of expertise is achieved, and best displaced by direct instruction for novices?

“Effective staff development opportunities are key to Web 2.0 adoption, and
56% of teachers indicated that they would welcome more guidance in the
use of Web 2.0 technologies. More than a third (36.9%) of teachers report
that they never receive training in the use of new technologies including
Web 2.0; 26.7% say they only receive training ‘rarely’.
• Innovators commented that it was important that any such training and
support came from the bottom up and not as a top-down prescription.
Moreover, it was probably wise to start from modest aims.
• Web 2.0 can be the medium of exchange: 32.5% of teachers frequently or
occasionally use Web 2.0 to share resources and ideas with other



Posted in Uncategorized | Leave a comment