In 2014, I was honoured to be invited (with my wife Susan) to be a visiting scholar at Beijing Normal University (BNU). BNU is arguably the preeminent research university in distance and online education in China. One of my commitments during this visit was to create a review of the thorny and complex topic of interaction in distance education.
As anyone researching distance and online education and many readers of this blog will know, interaction leading to active and engaged learning is a pivotal topic for teachers, learners and institutions. Interaction is multi-faceted (many actors, many modes) but also expensive in terms of student and teacher time. Thus, there has been a wealth of research on the topic during the 40 years that I have been an active researcher and teacher.
Thus, with no apology for the length, I link here the final 44 page review. Most obviously it is 7 years out of date, but I think distance education researchers and teachers will find something of value in the extensive research and references noted. Ironically when I returned to Canada I submitted the review study and was told it was published in China, However, I didn’t bother to promote or publish the review myself.
Fortunately, Grad student SCOTT A. HAUERT, saw the review in Chinese and asked me to make it available (in English)! I attempt to rectify this mistake by posting the review here under a Creative Common Public Domain license.
In this post I review an article that provides the first systematic review of the Interaction Equivalency Theory (EQuiv) that I formulated 15 years ago. The article is:
In 2003 I published an article in which I wrestled with the increased capacity for interaction that was becoming available to those of us designing online courses. I realized that synchronous, asynchronous, text, video, voice, mutli media, “smart” content and many more tools and toys were becoming available and flogged by ed tech companies. I also realized that many of these tools were expensive both in terms of money to purchase and support and in the time they took for both students and teachers to first learn to use, and then to effectively use. Perhaps I was being both simplistic and reductivist, but I speculated that though interaction is critically important in distance education, it can take many forms and further that one form can substitute for another. Building on Michael Moore’s notions of student-student, student-teacher and student-content interactions, I gave a very fancy name (Anderson’s Interaction Equivalency Theory) to the idea that if you could have a high level of one of these student interactions, you could reduce or even eliminate the other two. I further contented that adding the remaining two forms, may increase learning and persistence, but it would be more expensive (time and money).
The article was submitted and published:
Anderson, T. (2003a). Getting the mix right again: An updated and theoretical rationale for interaction. The International Review of Research in Open and Distance Learning, 4(2), 1-9. https://doi.org/10.19173/irrodl.v4i2.149
I got a few comments on the article and gradually noticed that others were quoting the article (not all positively) and a few researchers were using it in their conceptual rationale for their work. I see today that Google Scholar lists 934 citations to the article. So I was pleased with its modest interest and use. But I wasn’t even convinced myself it was entirely true. Like too many educational (and theological) theories, it could be used to explain a result in hindsight, but it is very challenging to design and implement experiments that could falsify the theory.
Partially to increase the validity and value of research in education that does not necessarily use control groups and other positivist methods, systematic reviews have recently become more widely used (for example see Martin, F., Ahlgrim-Delzell, L., & Budhrani, K. (2017). Systematic Review of Two Decades (1995 to 2014) of Research on Synchronous Online Learning) Thus, I was really pleased to see the first systematic review of the Equiv. Theory.
The Graham & Massyn (2019) article comes from Connie Graham’s PhD Thesis and is authored with her supervisor Liezel Massyn from the University of the Free State, Bloemfontein, South Africa.
Systematic Review Methodology
Like any good research project, this one starts with the research question . “How can the EQuiv be used to enhance interaction opportunities of non-traditional doctoral students?” It then provides the selection keywords (key words like doctoral education, EQuiv, interaction, non traditional students etc.) that were used to query the major journal databases and dissertation indices. The papers were further weaned to to focus on ones in which interaction and non-traditional doctoral students were highlighted, with or without the EQuiv Theory.
The context of this research is also of considerable personal interest to me, as I helped design, taught and researched with ‘non-traditional’ EdD students, studying at a distance, at Athabasca University. The paper is really three mini-systematic reviews rolled into one. The first is a review of issues related to non-traditional doctoral students. This section reviews studies that relate to completion, supervisor- student relationships, risk factors for dropout etc. The next section reviews interaction requirements in education in general and specifically with doctoral students. The final section reviews the EQuiv theory itself.
1. Doctoral studies with non traditional students
The doctoral research hones in on the often mythical relationship between the student and the supervisor. The the teacher’s role in this relationship has been described as “mentor” “master” ‘supervisor’. Graham & Massyn use the term ‘master-apprentice’ to describe the ideal form of this relationship. This relationship originally evolved on campus-based universities. At its best the student not only acquires content knowledge but also is socialized into the profession. This is accomplished by regular planned and spontaneous interactions between master and the apprentice doctoral students. This master/apprenticeship relationship is used in the training of doctoral candidates to help them to gain a deep understanding and opportunity to participate in the culture of their discipline tribe. This model/design has hundreds of years of university replication baring evidence that it can work. However, the Internet came along and caused us learn how to use mediated communications to create Equiv learning and socialization outcomes.
If we look at a typical doctoral student in the USA and in Canada today, they are studying all or some of their program online. In addition there are a myriad student-student online support interactions using social media. These students don’t often sit around the graduate coffee room and don’t get to be personally present when the cultural activities of the discipline are presented. However, they may (or may not) be meeting regularly with their supervisor via Skype, be following each others tweets and blog posts, and be following similar research topics or developments in their respective networks and forwarding them to each other. In addition they may be networking with professors and other graduate students around the world thus creating a new form of connected doctoral student.
Neither the “sitting at the knee” master-apprentice model nor the “connected model” works out in reality. Today the master is as often not on campus and private office conversations seem hard to arrange. Doctoral students have many demands on their time from vocation, family and health and are not readily available to benefit from face-to-face encounters.
On the other end, the master is often using different tools (University versus commercial provision) or prides themselves on NOT being on social media. Thus, the amount of personal interaction and socialization is extremely varied in today’s doctoral programs. This article begs the question, If the traditional student-teacher (master-apprentice) interaction is impaired does the Equiv theory help us to design compensatory interactions?
2. Interaction in Doctoral Education- especially at a distance
The second section deals with interaction in education with a focus on non traditional doctoral students. It is a good overview of this critical role of interaction in all modes of formal education. The usual student-teacher, student-student and student-content interactions are reviewed. I especially liked the section on student-institution interactions. I’ve usually considered this a subsection of student-content interaction. Especially for doctoral students an efficient and comprehensive web site or portal is critical to answer detailed procedural questions that every student bumps into. How many people of the candidacy examination committee? Which of the Faculty members would be the best member of my supervision committee? In days past these questions could be answered by informal conversation among grad students or hints from “the master”. But today a good web site is much more effective .
3. Interaction Equivalency
The third section of the review (longest and one of most interest to me) is on interaction equivalency. The review notes the earlier work by Simonson, Schlosser, & Hanson, 1999, that describes the necessity for distance education students having the “equivalent” experience in education as their campus based colleagues. This use of the equivalency in Simonson et al’ article is not what I had in mind. Distance education is not ‘equivalent’ to campus education in the sense that some experiences both on campus and off are not experienced by those not engaged in that mode. A lecture is NOT identical to a videocast, but they may have identical outcomes. Demanding literal ‘equivalence’ denies the unique affordances of both the live performance and highly mediated interactions.
The review then does a really good job of explaining the theory with some of the diagrams created by my colleague Terumi Miyazoe.
Graham and Massyn found a total of 25 papers directly using the Equiv theory that they summarize. The authors create a table in which they categorize and give examples of 12 different characteristics of the research papers such as learning method, type of students, interactions etc. None of the studies seem to directly falsify or uniquivacably support the theory, but most give a sense that it is a useful tool to think through a problem. As expected, the results are a bit inconclusive or as they state in the conclusion “the literature on the EQuiv is contextual, relative, and inconclusive.” This is not surprising as the Equiv is perhaps best used as a diagnostic or mnemonic aide to design and learning enhancement. One of the authors noted correctly that we never really provided a precise way to measure “high” or “low” levels and thus researchers have been forced to create their own metrics. And really, this is all I really wanted from the “theory”. Equiv is a designers’ (or teacher’s tool) that can and has inspired some empirical research but perhaps is best classed and measured by its efficacy as a design tool.
In the summary, Graham and Massyn present a new graphic that illustrates the Equiv in doctoral studies.
The dotted line shows potential (or likelihood) of challenges in quantity and quality of interactions with teacher that are routinely faced by non traditional, distance students. The diagram shows how an intervention, with enhanced S2S, S2T or S2C interaction, can address this deficiency and lead to social and academic integration and thus successful educational experiences.
To conclude, congratulations to Graham and Massyn for a useful contribution to the Equiv and by extensive online learning literature. As they note its use in the important and growing area of doctoral research, at a distance, is very under researched and there is lots of interest and room for ideas on how to make this experience more effective – for students, teachers and institutions.
I started out my teaching career as a “shop teacher” – teaching middle school students how to work and built with a number of technologies. Thus, it was a bit disturbing to listen to recent CBC radio broadcast listing jobs that have disappeared and to hear that ‘shop teachers’ along with elevator operators, typists and postal worker were disappearing.
Two articles from the latest issue of Online Journal of Distance Education Administration, caught my attention for the same reason. Will technologies soon reduce or even eliminate the relatively new job position of “online teacher”.
directly addresses the issue of substituting the traditional teacher role of answering student queries with a machine. The article notes that most online students do their work in the evening and on weekends when many teachers are not interested and often not available to answer questions. In an attempt to provide some sort of 7/24 service the instructors built a chat bot and seeded it with a database of questions culled from archives of questions asked by students taking the course in previous terms. Chat boots have rarely been used in education, though colleagues at Athabasca worked developing a Freudbot. But chat bots are becoming ubiquitous on online shopping sites where they provide a type of 7/24 customer support.
During the design-based research project the Bot was involved in 475 sets of interactions with the 56 students enrolled in 2 sections of an early Childhood Education course. As expected more than half of the interactions took place on weekends and 88% were during evenings. Of course some of the questions were ‘off topic” and used by curious students to learn the capabilities of the bot. But many were on topic with 65% of queries focussed on information seeking – much related to assignments and exams.
Of perhaps most interest was the students’ reactions, which were queried by 5 point Likert perception questions. Only about half (m=2.5) found the bot did effectively answer their question, but nearly all (m=4.8) noted the extra service provided 7/24 by the bot and over half (m=3.3) preferred asking the bot rather than emailing the teacher. The effectiveness of these type of bots naturally grows as the data base expands through use and the searching algorithms improve. Thus, one can, even now, see how machines will undertake at least some traditional teaching roles which in a positive sense can give teacher more time for more important learning diagnosis and help and reduce time spent on administrative type queries. It is also interesting to think about the “teaching presence” of of the female avatar (displayed at left) used in this study. In any case a very nice exploratory, design-based study.
The second employment related article compared adjunct teachers employed in for-profit universities as compared to those in the not-for-profit sector. This is an important problem as more than half of online courses in the USA are taught by part-time sessional instructors. This in itself has huge implications on job stability for teaching faculty, but the rise (in the US) of for-profit universities with a tradition of not offering tenured positions, doing no research and not training next generation of scholars is also threatening.
This study used an online survey to query 859 online adjunct teachers. The results were remarkable in the lack of statistical differences between the teachers at the two types of institution. No differences in age, satisfaction, education or a host of other variables. There were small differences in class size, but this was confounded by the higher percentage of graduate courses (with normally smaller numbers of students) in the public, not-for-profit universities.
The authors put a nice spin on the discussion by noting that the similarity means that professional development and support can be shared between the sectors – unlikely as that may be. But the study also adds empirical support to the notion that education can be and is being “privatized” with resulting decreasing employment for traditional, tenured faculty.
So what do I conclude from these two very good articles? I think one of the largest challenges for our global population (ranking right up there with climate change) is the capacity to meaningfully employ people in a context in which machines are more and more able to perform both mundane as well as high performance and high communication tasks. My Interaction Equivalency Theory speaks to this by noting the ways in which activity that used to be performed by live teachers (student-teacher interaction) can and is substituted by bots and canned media to create high quality student-content interaction.
I’ve been fascinated by the role of interaction all of my career as both a student, a researcher and a teacher. Michael Moore’s famous article details the role of the ‘big three’ (student-student, student-content, student-teacher) interactions and influenced Randy Garrison and I to explore the other 3 possibilities (teacher-teacher, teacher-content and content-content interactions). I’ve written a number of summary articles, a recent article on interaction in MOOCs and note that interaction serves as the primary indicator of ‘presence’ in the Community of Inquiry (COI) Model.
Many, many research articles have shown significant and positive relationships between interaction and a host of outcomes including persistence, achievement and enjoyment. However, these studies are almost always correlational and sometimes based exclusively on student perceptions. These are useful methodologies but marred by challenges of proving causation. Did the interaction cause the positive outcomes (causation) or do motivated students both interact more and get better marks (correlation)?
Thus, I was pleased to see an interaction study in the latest issue of IRRODL that used a quasi-experimental study to examine the impact of student-teacher interaction. The article:
Cho, M., & Tobias, S. (2016). Should Instructors Require Discussion in Online Courses? Effects of Online Discussion on Community of Inquiry, Learner Time, Satisfaction, and Achievement. The International Review Of Research In Open And Distributed Learning, 17(2). doi:http://dx.doi.org/10.19173/irrodl.v17i2.2342
The study was set in the context of a US undergraduate, fully online course with between 25-30 students in each of three sections taught by the same instructor. In the first instance there was no discussion. In the 2nd, the teacher posted a weekly discussion question and students were obliged to post at least one answer and one comment per week (typical forced participation that brought down the wrath of friend Jon Dron in a recent post). In the final section the teacher actively participated in the weekly discussions. In all cases the teacher was readily accessible via email.
A strength of the study was the multiple measures of interaction effects. These included:
completion of the COI Inventory – a Likert-scale, perception instrument derived from the original indicators of teaching, social and cognitive presence).
Student satisfaction measured by 3 lLkert questions
Time on task as represented by login time to the LMS
Student achievement as measured by final grade. The authors wisely excluded the marks for participation awarded in instance2 and 3.
Perhaps the least surprising outcome was that perceptions of social presence were significantly different with, as expected, higher perceptions of social presence in the more interactive instances. Also not surprising is that teaching presence increased in the 3rd instance, but not significantly.
The results became both interesting and surprising on the final 3 measures. In each of the three course sections, each with markedly different amounts of interaction, there were NO significant differences in terms of time on task (logged into Blackboard), student satisfaction or student achievement. There was a small (but not significant ) increase in student satisfaction in the 3rd instance with enhanced teacher participation.
The discussion section of this paper is also very good. They note that time requirements for participation required in the forums in instances 2 and 3 did not end up costing students more time – at least as measured by Blackboard logins. Finally, they note the obvious – teacher participation did not lead to increased achievement – despite the assumed time commitment required of the teacher.
These results tend to reduce support for the importance a number of the types of student and teaching presence described and promoted in the COI model. But the results provide support for the ideas I promoted in my Interaction Equivalency Theory. I proposed there that given high levels of one of the three levels of interaction, the other two could be reduced without loss of academic achievement. I also noted in my 2nd thesis of that work that likely increased satisfaction would results if more than one of the three forms were used (in this study, there was an increase but it was not significant in student satisfaction across the instances) but that it would come at increased cost (usually time to both teachers and students).
The study also did not look at attrition, perhaps because the numbers were small. I know from experience at Athabasca, when teaching and peer interaction is drastically reduced in self paced and continuous enrolment designs, that attrition almost always increases.
This excellent study concludes with recommendations for practice which include the note that student-student and student-teacher interaction are just choices and shouldn’t be considered to be hallmarks of all online (or classroom) courses. Good learning design counts!
