WordPress in red, blackboard in blue. Showing the number of CoI presences indicated per day in each platform. WordPress consistently showed more fully-rounded presences (social, teaching and cognitive) while blackboard more commonly was restricted to just 1 or 2 simultaneous presences. Interesting.

The Community of Inquiry coding data, aggregated and compared. The WP posts showed more indicators of cognitive, teaching and social presence, across the board. Also interesting is that the instructor was only minimally active in WP – the teaching presence indicators were made by students…

It’s only the most basic of metadata, but already differences in activity patterns are becoming apparent. Both images are on the same time- and size- scales. The WordPress discussions appear to be using significantly longer posts and comments, spread over much more time. Blackboard discussions appear to be shorter posts, over briefer durations.

Next up, I get to code each post for Community of Inquiry model “presences” – as described by indicators for social, cognitive and teaching contributions in the posts. I’ll figure out some way to overlay that information on top of the basic metadata visualization.

Xin, C. (2012). A Critique of the Community of Inquiry Framework. The Journal of Distance Education, 26(1). Retrieved from http://www.jofde.ca/index.php/jde/article/view/755/1333

Thanks to Stephen Downes for pointing this paper out. I’m up to my eyeballs, processing data for my Community of Inquiry based MSc research, and could have missed this.

The Community of Inquiry model provides a framework for describing interactions within a community or classroom environment. It involves using textual analysis and coding of messages to interpret the type of interaction for each message – whether it involves social, teaching, or cognitive components. As I’ve been coding the data for my thesis, I’ve been adding as many types of “presences” as are appropriate – a message may include a number of things, indicating social, teaching and cognitive presences in a non-exclusive manner. I’m imagining each message having its own little Venn diagram for Social/Teaching/Cognitive component, as per the CoI model. It’s a simplification and abstraction, certainly, but looking at the coded output, I think it’s still got a fair bit of fidelity to describe the interactions at a high level. In my data, I’m also adding coding to describe the type of content (links, images, attachments, embedded media, etc…) as well as how involved the message is (is it a simple one-liner? a 2 paragraph response? a multi-page essay?) – and I’m thinking about how to include data on the timeline of the discussion (how rapid were the responses? staccato rapidfire conversation, or long drawn-out periods of silence?) I’m still thinking about how to represent that kind of data for an online discussion, but I think there’s something there, there.

In a total aside, I’m reading Foundation’s Fear (by Asimov & Benford), and it’s rekindled thoughts about psychohistory and the like – models at such a high level of abstraction that individual actions do not matter. That’s kind of how my thinking as shifted with respect to the use of CoI for my online discussion data – it’s not the individual messages that are interesting to me at this level, as much as the overall patterns and flow…

Now, on to the article!

Abstract:

This conceptual paper critiques the popular Community of Inquiry framework (CoI) that is widely used for studying text-based asynchronous online discussion (Garrison, Anderson, & Archer, 2000). It re-examines the three main aspects of CoI (cognitive presence, social presence, and teaching presence) and their relationship, and further highlights the specificity and complexity of online discussion forums. The paper explains that the multi-functionality of communicative acts which often combine instruction, knowledge construction, and social interaction in a single utterance. It further argues that all online expressions are inherently social. It clarifies the confused relation of cause and effect in CoI and specifies the leadership functions of teaching presence and how they are intertwined with social and cognitive presence. A game metaphor and Gadamer’s notion of “play” are employed to explain the dynamics of online discussion forums. The article concludes with a discussion of the practical and methodological implications of its main arguments.

on the coding itself:

Each of the three components of the CoI is an abstraction from the whole, i.e., the communication process. Often all three aspects are performed simultaneously in a single communicative act, e.g., a sentence or paragraph, their precise function depending on what said previously leading up to that point, the contexts, and the dynamics of the discussion.

Agreed. I think it’s essential to be coding each message in a way that documents each aspect. A message may include a couple instances of “social”, one of “teaching” and a few “cognitive”. The Venn diagram, rather than a single “this is a Teaching Presence Message”… In addition, other components of the message and discussion also convey meaning – the rapidity of the responses conveys a message about engagement or urgency. The depth of the responses conveys a message about importance or complexity, etc…

I think that by documenting all presences or aspects indicated in each message in a discussion, it should be possible to capture the more fluid communicative functions while still providing data at an abstract level that can be analysed as a whole.

on abilities vs. performances

Why does this distinction matter? It has to do with identifying what is operative or at work in the medium as opposed to capacities or desiderata that may or may not be active forces in the communication process. Abilities are not the same as performances. Knowing a rule is not the same as obeying the rule. The CoI encourages one to think about what a successful conference would entail, but it does not adequately account for how to get there or make it happen. The communicative functions provide the operative means for accomplishing that. They are important for clarifying what one can and must do online so that participants are able to play their part in a successful educational experience.

I may be misinterpreting the CoI model, but when I’m coding my data I’m not looking at abilities or roles – I’m looking at what actually happened in the message(s) – the performances. I don’t think CoI is incompatible with that view, and it shouldn’t force an emphasis on potential (and undocumented) abilities or the nature of institutional roles.

The Communicative Functions are interesting, and might make it more explicit that the actions are what are important, but they should be compatible with the CoI coding of presences.

on desired states vs. modeling behaviour:

Desired states of affairs such as “open communication” are descriptors that may or may not apply in any given situation no matter how much they are referenced as goals. This is why modeling behavior is important. A teacher could say, “Feel free to speak up” but until someone does so (e.g., by saying something daring) and is accepted as normal (e.g., acknowledged by the teacher), the norm is only an unrealized ideal. Again, this is an example of how communicative functions (CFs) operationalize a rule or a desired state of affairs. “Presence” must be understood as just such an operative component of the discourse. In this sense, the cognitive, social, and teaching presences are best specified in terms of the functions that make them “present.” “Presence” is an effect of what I call “functions.” In other words, CFs are actions that cause the effect of someone or something being present.

Bingo. This is something I struggled with as well – how do you know if a message is something that indicates “open communication?” In my coding, I have been looking at the actual messages as indicators of this – it’s not open communication because someone says open communication is encouraged, it’s open communication when a message shows that someone is communicating openly. This could be as simple as asking a question, or challenging something that was posted earlier.

on social presence, and confusing cause and effect:

The three categories of social presence also confuse action with outcome. Affective expressions are communicative actions that can cause certain effects. Open communication and group cohesion are not actions but positive effects or outcomes of successful forum management. Putting the three together confuses cause (what one does) and effect (what the action leads to) and obscures what a participant should (or should not) do online in order to achieve the desired state of affairs.

Now this is interesting. The indicators given for defining “social presence” are somewhat muddy or confused. Some look at actions directly, some look at indicators of desired outcomes. Is this just semantic nitpicking, or an indicator of something deeper?

I’m wondering if it might be interesting to analyse my data using Communicative Functions in addition to CoI…

on teaching presence:

Earlier I raised the concern that from a strictly communicative perspective, “instructional management” and the “design and organization” of content and activities such as “setting curriculum and methods” (Garrison, et al, 2000, p. 89) contribute to the learning experience but do not constitute the teacher’s presence online since the only way to be present online is to express oneself in words. This is important because no matter how thoughtfully a conference is designed and structured, without the active involvement of the teacher in the process of dialogue, its success is left to chance.

This is consistent with what I’ve seen – it’s not the initial design of an activity or discussion, as much as the continued authentic interaction by both teacher and student.

Xin maps the CoI teaching presences directly to Feenberg’s Communicative Functions. If they are able to map directly, doesn’t that suggest that they are roughly equivalent (but with different vocabularies)? If my goal is to gather a high level abstract analysis of the discussions, the exact vocabulary used to describe the interactions doesn’t matter all that much.

on multiple contexts in messages:

The highly context-dependent nature of online discussion also explains why facilitation is not just instructional, or social, or intellectual; it is often all of them simultaneously. Process and content are intertwined at the communicative level. Teaching presence both sustains the social relations of communication and advances understanding of the subject matter at hand. Because the messages that perform communicative functions often encapsulate cognitive contributions, the effective use of those functions fulfills an intellectual as well as a communicative role

Agreed. This is why I’m coding my messages using CoI in a way that can represent multiple contexts for each message. I may have to write up something to describe how I’m doing this – I haven’t found any solid resource on the mechanics of CoI coding at that level.

discussion and conclusion

Xin’s strongest point involves the confusion of cause and effect in indicators for CoI presences. Some of the coding is directly triggered by actions – someone uses an emoticon or shares an autobiographical narrative, and that indicates a social presence – and some are indirectly triggered by actions that are interpreted to imply desired goals or intents – someone appears to speak freely, so we infer “open communication” and a Social Presence.

This may just be confusion about how CoI can work – I’m using it as a hierarchy. Each of the 3 “presences” is represented by 3 categories, and each category is represented by a number of indicators. The wording used to describe the indicators may be a bit inconsistent (the cause-and-effect confusion Xin describes), but they can be interpreted consistently by a researcher to mitigate that.

Furthermore, the analysis of the communicative functions of online talks should be considered together with other aspects of interest — who said what, how, why and when.

Exactly. CoI is just one component of analysis – I’ll also be looking at other components to build the full picture.

Abstract:

The purpose of our research was to examine the influence of an online protocol on asynchronous discussions. A mixed-methods study compared two online graduate classes: one that used a protocol and one that did not use a protocol for the same discussion about a complex reading. Analysis of the data revealed that the online protocol more evenly distributed the presence of cognitive, social, and teaching elements necessary to create and sustain an online community of inquiry. Use of the protocol also promoted more shared group cognition and more student ownership of the discussion and empowered students to facilitate themselves, helping to reduce the instructor workload. These findings may enable educators to provide more dynamic interaction and richer learning experiences in asynchronous online environments.

purpose and context of the study:

Given the literature concerning online discussions, the use of protocols could prove particularly powerful to support both students and teachers. Concerning instructional design and organization, protocols may provide a very structured discussion prompt, which let students know their role, giving built-in supports leading to a progression in thought. However, it is still open-ended enough so that students can select their own topic, experiencing more ownership of the discussion. Protocols may reduce the instance of discussion monologues, since the protocol explicitly requires interaction. Because the protocol spells out who should post and who should respond at a given time, there is no need for a specific facilitator, as the participants essentially facilitate themselves.

they looked at 2 fully-online grad-level classes in a higher-ed setting. n=12 and n=14.

on protocol vs. non-protocol:

The protocol class received the “Save the Last Word for Me” protocol (McDonald et al., 2003), which was modified for an online discussion forum. In this protocol, at the beginning of the week, half the students were asked to post a quote from the reading, which they thought was important but particularly complex. They were told not to reveal the reasoning behind their selection of the quote. Then during the middle of the week, two students posted a reaction to each quote. At the end of the week, the students who posted the quote revealed their original interest in the passage and what they learned from reading the reactions from the other two students. This discussion was then repeated the following week with the other half of the students posting the quotes for discussion.
The non-protocol class received the open-ended discussion question: “How does thinking about a concept from multiple perspectives help or hinder the learner from gaining a deeper understanding?” The students responded to this initial prompt based on their readings and then replied to at least one other student’s response.

Results:

Overall, there was a more even distribution for the protocol group among cognitive, social, and teaching presences than in the non- protocol group, as shown in Table 2. Further analysis was done for the categories within each of these elements as described in the following sections.

so, with protocol, more people were more actively engaged, rather than a few keeners hogging all the fun.

Based on our analysis of this data, there were several interesting differences between the protocol and non-protocol discussions. For example, the protocol group had a more balanced distribution of the three presences than the non-protocol group. This is more reflective of the COI model, which emphasizes that all three presences have to meaningfully interact in order to facilitate a successful learning community (Garrison et al., 2000, 2010). The protocol group also demonstrated more shared group cognition than individual cognition. This increased interaction likely resulted from the structure of the activity and the defined roles for communication that prompted students to reply to one another to problem solve how to interpret the quotes selected.

This would be handy to look at differences in discourse between 2 different classes. Less relevant for my research, since both groups are in the same class, with the same instructor, doing the same activities. But if they weren’t, perhaps varying protocol used by different instructors could explain varying levels of engagement…

There’s a PROTOCOL to be followed!

Years passed, and I’m now planning the research proposal for my MSc thesis. And it turns out that the Community of Inquiry model is probably the best fit for what I want to do to investigate differences in discourse between two cohorts. More info on my research proposal at a later date…

Basically, what Community of Inquiry does is to take a look at the discourse of a community, from three overlapping perspectives:

* cognitive presence
* ability of participants to construct meaning through sustained communication
* social presence
* ability of participants to project their personal characteristics into the community
* teaching presence
* design, facilitation, and direction of the community processes

You take the discourse of a community, crunch it through some latent content analyses, and get an idea for how the participants fit together along the three perspectives.

The really appealing thing about the COI model is that it has been [used by several researchers](http://communitiesofinquiry.com/references) to investigate various communities, so it’s got some validity and rigour behind it. That’ll make my job much easier, as I won’t have to spend as much time designing and defending the analysis framework…

I found it pretty interesting that my poking around with research proposal planning lead me right to Randy’s work. What are the chances that you’d happen to be working for one of the key researchers in a field?