Implementing Connectivism by Semantic Technologies for Self-Directed Learning

Tham, J., Duin, A. H., Veeramoothoo, S. (Chakrika), & Fuglsby, B. J. (2021). Connectivism for writing pedagogy: Strategic networked approaches to promote international collaborations and intercultural learning. Computers and Composition, 60. https://doi.org/10.1016/j.compcom.2021.102643

Broad Topic: This article is seeking to determine if a measure can be created as it relates to the connectivism learning theory and the learners’ determination of connected concepts.

Theoretical frameworks used to ground the piece: “Connectivism describes knowledge as a set of symbolic mental constructs and learning as the process of accessing and storing these symbolic representations in the learner’s memory” (p. 1034). The researchers are trying to employ a network connected ontology to describe how connectivism can be measured. In short, how can learning performance be determined through connections that are formed when attempting to learn a subject? The use of ontologies as a knowledge representation would allow a visual graphic in which the interlinking concepts can be connected, thus describing the relationships between areas of interest.

The authors are seeking to answer two questions as it relates to concept importance: 1) How well connected are the concepts in a network, and 2) how necessary is a single concept for understanding connected concepts based on its underlying semantics? Essentially, how well do learners perceive the information they are receiving? The authors, with the assistance of an ontology engineer, determined that the level of “need” and “detail” would be the two deciding factors when detailing the determination of concept importance. A complex mathematical formula is created that is beyond the scope of this investigation, yet the relationships within the ontology are numerically defined. “Taking into consideration the type and number of connections for concepts, the concept importance measure is a suitable measure to rate the potential of concepts in a domain to explore new information in line with the idea of connectivism” (p.1039).

Seminal constructs or concepts in the article: George Siemens’ paper, Connectivism: A learning theory for the digital age, serves as the seminal construct for this paper. His work established the theory and laid the framework for other researchers. The researchers are also using the following articles to base their work on: Vas, R. (2007), “Educational ontology and knowledge testing,” as well as Weber and Vas (2016), “Applying connectivism? Does the connectivity of concepts make a difference for learning and assessment?” In applying these articles, the researchers are interested in the self-directed learning that many learners apply when gaining new information. They note that learning is not an isolated process, and much information is gained via the use of various communication techniques. Knowledge objects, and the technology environments where information is stored, would benefit from a schema that categorizes the ontology in a personalized manner. “Even if the implications of connectivism on individual learning is still unexplored; it is expected that better information access supports learning and a measure explaining the importance of concepts in information sharing helps to gain better insight into the learning process” (p. 1035).

Broad research or theory objectives including hypotheses, propositions, etc.: The researchers are interested in exploring the concepts deemed important by the learner, thus determining the domain ontology of the learner. “Accordingly, the measure integrates a set of factors: the number of connections of a concept and an interpretation of the semantics of the domain ontology. These factors – composing the measure – are together called the “importance dimensions” (p. 1037). The researchers use a multiple-choice exam in hopes of mapping the importance measure of each topic explored. Their goal is to document the various topics that are visited by the learner as they study for a domain-specific exam. The knowledge network that is created seeks to determine if the concept importance measure forms a correlation between performance on test questions of the same subject.

Research methodology: Data was collected in 2016 and 2017 from an undergraduate course in management information systems. In 2016, 267 students for both mid-term and final-term examinations were sampled, and in 2017, 278 students for the mid-term and 276 students for the final-term made up the population of students. The mid-term exam consisted of 20 questions while the final was comprised of 30 questions. The 2016 exams (mid and final) covered 56 concepts and 246 total questions. Similarly, the 2017 exams covered 49 concepts and 217 total questions. Two professors and the ontology engineer selected the concepts for the methodology at which point they then screened each question and matched it with a concept. A subontology was extracted and used to calculate the importance value for each concept. A graph was created that separates the ontology into smaller clusters that allows for better visualization of the concepts (graph to follow paragraph). “The size of the node visualizes the degree of the concept importance, and the color of the nodes, references each to one of the nine calculated modularity classes. Only the concepts that were matched against the examination questions are shown with labels” (p. 1040).

Findings, implications, conclusions: The goal of the process was to see if concept importance could be a predictor to how well the learner answered test questions. The researchers used a linear regression model as well as a quadratic function to see if concept importance could be a predictor test competency. No significant results could be found even though trends were determined for all aggregation levels. To explain this occurrence, the researchers suggested the “small number of concepts and observations for each individual and the randomization of the questions” (p. 1041). Variations at the concept level resulted in an unsteady trend that led to outliers affecting the data points at the class level. For the 2016 data, the p-value for the mid-term was .18 above the .05 threshold, while the final p-value was .9 above the .05 mark (each was stated as non-significant for the regression model used). Likewise, the 2017 data resulted in a p-value .31 above the .05 threshold for the mid-term and .44 for the final (both are non-significant for this model). Multiple reasons for these results are given by the researchers but for the purpose of this study, only a few will be mentioned: 1) recorded observations are limited, 2) unequal questions as it relates to associated questions, 3) unequal domain level coverage, 4) controlled environment in terms of test-taking yet uncontrolled in terms of test-prep, and 5) different learning environments and how the learner perceives information.

Prescriptive and descriptive contributions: I found it interesting that the researchers were attempting to find a measure on how the learner deems important concept connections. They also noted in their research that “no single strand of theories can currently fill the gap between the ideas of connectivism and its implementation or derive implications on the learning process” (p. 1043). This seems to point to a need for connectivism to be further researched in terms of application, and whether the theory can stand on its own in terms of an educational learning principle. While they did not find any significant results, their work is encouraging in that given a larger sample size with more variables (test questions), results might indicate a different finding. In terms of our learning technologies, I feel like the researchers were trying a novel approach in attempting to measure concepts and their importance to the learner via a relatively new (and largely ungrounded) theory. The idea of connectivism appeals to me due to its core premise in that learning is made through connections that are established between nodes. Yet, like the authors in this study, the impact of this theory needs to be vetted and more firmly established before we as a field accept it within the learning community.

Critique of ideas and research: A few points to note: Students learn in different environments and with various technology; the concept important measure is subjective and not well grounded; information exploration is not well-defined in the research or within the connectivism theory. These three areas, alongside the other points I mentioned in the findings, lead me to believe that an attempt to map an ontology with the learning experience would prove to be a mammoth effort due to the individualistic nature of learning and the replication needed to provide support for the study.

Relevant gaps and research opportunities: This study serves as a great starting point for future researchers in attempting to find a measurable observable involving the connectivism theory. While the researchers noted their lack of significant results, credit to them in attempting to find a way to link a theory with an applicable measure. The trends that were noticed, higher importance to lower passing of concepts in the mid-term, while lower importance to higher passing of concepts for the final, indicate a need for starting points in what is deemed fundamental knowledge. The authors noted that highly difficult objects are challenging to learn, yet the performance related to the objects increase over time. More research needs to be done on the concept importance measure and if it is valid in various projects.

Learning Theory and Pedagogical Practice for Networked Information Landscapes

Dunaway, M. K. (2011). Connectivism: Learning theory and pedagogical practice for networked information landscapes. In Reference Services Review, 39(4), 675–685. https://doi.org/10.1108/00907321111186686

Broad Topic: I am studying Connectivism as a learning theory in online learning. This article, written in 2011, was one of the first papers I read when learning about Connectivism in a previous course. The author expands on the theory set forth by Siemens and Downes while applying the approach to information literacy instruction.

Theoretical frameworks used to ground the piece: This article draws from the original Connectivism article written by Siemens and Downes while drawing information from the following articles: Mackey and Jacobson (2011) and Thomas et al. (2007). The two aforementioned articles dive into the topics of information literacy and how computer-mediated communication has changed the framework in which prior theories were presented. In these articles, two frameworks emerge that are relevant to the connectivist learning theory: metaliteracy and transliteracy. Metaliteracy, as explained by the researchers, builds on prior information literacy theories yet believes that they do not “consider collaborative media production and the impact on learning, which is why we need an expanded metaliteracy model with an emphasis on active production and sharing of new knowledge through technology” (p. 679). This expansion allows for Connectivism to be explored as a potential link with metaliteracy. Transliteracy takes a narrower approach in that the authors believe the focus on information literacy should be “on how communication takes place; transliteracy refers to skills that enable learners to access information and communicate across multiple platforms, tools, and technologies” (p.679). Again, the use of technology is the center piece of this theory and has the potential, like metaliteracy, to unify with Connectivism.

Seminal constructs or concepts in the article: George Siemens’ paper, Connectivism: A learning theory for the digital age, serves as the seminal construct for this paper. His work established the theory and laid the framework for other researchers. His paper proposed certain tenets such as: 1) Learning is process involving the connection of nodes (aka information sources), 2) learning and technology are intricately linked, 3) Diversity of opinion, and the ability to perceive connections, are fundamental to leaning, 4) connections must be formed and maintained and, 5) evaluating information is a skill that represents the learning process. These facets emphasize the learners’ ability to process knowledge that is stored and biased via technology. The ability of the user to make connections is central to the connectivist theory, thus adding to the notion that learning occurs by developing a network, establishing connections, and learning through the pathway that is formed.

Broad research or theory objectives including hypotheses, propositions, etc.: Connectivism is a modern learning theory to deal with direct impact the technology has on learning. The change brought on by technology calls for a learning theory to reflect our way of thinking and our way of off-loading tasks to a system. These principles of Connectivism guide the author’s research in that the learning environment should allow the student to gain from the instructional experience. Since this paper covers information literacy, as experienced by a librarian, the “Web 2.0’s emphasis on communication, connection, and content creation has contributed to educators’ interest in the ways that web technologies interact with students’ learning processes” (p. 678). According to the author, information literacy theories are more concerned with the social framework that information and libraries exist within, and not the technological background in which information resides.

Research methodology: The author conducted a literature review with two key tenets arising from her work; metaliteracy and transliteracy. Both models recognize the impact that technology has on learning with slight differences. Metaliteracy focuses on the connection between different literacies, while transliteracy is concerned with the connection process between the user and the technology. “Metaliteracy and transliteracy are frameworks for understanding information literacy that emphasize the importance of communities, connections, information networks, and information technologies; these concepts are central to the principles of the theory of connectivism, which postulates that communities, connections, information networks, and information technologies are central to the learning process” (p. 680). The author attests that Connectivism could explain how students acquire the aforementioned literacies.

If the purpose of Connectivism is to postulate that learning is the network that is established, then it follows that those practitioners of information literacy should acknowledge that students learn by connecting to the vast array of networks available. In terms of applicability as it pertains to libraries, if students can be encouraged to view libraries as part of their personal learning network and, as it relates specifically to information literacy, students would gain transferable skills that enable them to learn across a range of media. The author further explains that sources such as Wikipedia act as a pre-search tool for students prior to engaging with in-depth research databases. She notes that by recognizing user-generated content as “potential nodes in students’ personal learning networks and point students towards scholarly resources and library databases that represent other locations of knowledge” (p. 682). As it relates to librarians, the author further specifies that they have the ability to direct the learner to library resources and potentially cultivate the creation of a connecting node in the students’ personal learning network.

Findings, implications, conclusions: Due to this article being a literature review of information literacy and Connectivism, there is not specific findings to report on. In terms of a conclusive statement, the prevailing theme is that the information literacy framework has been significantly impacted by technology and that metaliteracy and transliteracy can provide a framework for helping students learn. The learning theory behind the instruction process, Connectivism, can help librarians understand how students are learning and engaging with their formed networks.

Prescriptive and descriptive contributions: Connectivism provides a framework for how students learn through personal networks, how teachers can play an influential role in shaping a classroom, and how technology can help design instruction that facilitates modern modes of learning. The author suggests that a single theory cannot explain the entire learning process, but points to a layering approach in which Connectivism can build upon prior learning theories (such as Constructivism, Cognitivism, and Humanism). In our field of learning technologies, I find this statement to be pertinent in that we are interested in how students learn. Multiple theories hold valid information that can be utilized in learning the process of how a student engages with their surroundings, forms patterns that build remembrance, or utilizes their available resources to shape their knowledge repository.

Critique of ideas and research: The author’s understanding of the impact that technology plays on a student’s ability to learn was very important. This article was written in 2011 when many competing ideas were battling for dominance as it relates to learning. I think she did a good job in recognizing the role that media plays when forming learning connections and that not all learning happens in a formal setting with formal resources. More research needs to be done as it relates to verifiable and empirical proof that Connectivism is a theory of learning and not just a perspective on how technology impacts the process. Some critics have argued that Connectivism is more of an instructional theory than that of a solidified learning theory. This statement could be true due to the systematic nature that surrounds technology (also, technology can be viewed as merely a tool and not a true learning mechanism).

Relevant gaps and research opportunities: At the time of publication, the idea behind the paper (Connectivism) was fairly new with many 2.0 technologies launching that had the ability to impact connections made to various sources. It would be interesting to see new literature as it relates to Connectivism and how our technological society has impacted the learning process. Are our sources factual? Do we take for granted the scientific process and ingest low-hanging information due to laziness? Has technology hindered our ability to think for ourselves, or are we evolving, and technology is our new system companion? Have more researchers accepted Connectivism as a theory and worked to iron out the flaws within it? Or is it still considered an outside perspective compared to Constructivism? Is there room for a melding of the two with parts from each theory combined? Regardless of the end result, the fact is we do learn via a digitized network and are directly impacted by the expansion of technology.

Selected Readings and Research Notes

This collection highlights selected readings that influenced my academic work and professional thinking. Each entry includes a citation, a brief summary of the source, and my interpretation of its relevance to learning, technology, design, and human behavior.

Dunn, W. N. (2018). Theories of the policy process. In Handbook on Policy, Process and Governing. https://doi.org/10.4337/9781784714871.00013

For this article, the author is elaborating on and describing the various methodologies and frameworks that comprise the totality of the policymaking process. This process is not always grounded in a specific framework and the author is expounding on the role of the individual, as well as that of the collective group, to promote or change a policy. Multiple-streams, punctuated-equilibrium, and advocacy coalitions theory are discussed and how each has a different effect on the decision-making process. “In the advocacy-coalition and the punctuated-equilibrium theories, policy change depends on what has occurred around the issue over a long period of time” (Schlager, 2018, p. 310). Multiple-streams theory differs in that the combining of like items, or streams is needed to successfully enact a policy change. In short, the careful and detailed process of deciding on a framework and theory should combine the quantitative and qualitative components in determining the policy explanation.

Greeno, J. G. (1994). Gibson’s affordances. Psychological Review, 101(2), 336–342. https://doi-org.libproxy.library.unt.edu/10.1037/0033-295X.101.2.336

Gibson’s Affordance Theory states that there are “affordances” or clues within our environment and how we perceive and interact with those clues are the basis for a decision or action to be made. Greeno (1994) states that “affordances are, in this view, preconditions for activity, as I believe is made clearer when they are treated as conditions for constraint” (p. 340). Humans modify their environment to obtain the desired response one is seeking. When we see a doorknob or a door leading to the destination we are attempting to arrive at, we pick up on affordances that prompt us to act, i.e., turning the doorknob or opening the door. The more we learn about our surroundings and how we perceive and interact with it, then our actions begin to make sense and our decisions more logical. Greeno expanded on his thoughts and added conditional constraints to his theory in that the constraint must follow certain conditions that exist in the background.

Mikropoulos, T. A., & Natsis, A. (2011). Educational virtual environments: A ten-year review of empirical research (1999–2009). Computers & Education, 56(3), 769–780. https://doi-org.tamut.idm.oclc.org/10.1016/j.compedu.2010.10.020

For this study, the authors reviewed empirical research articles that dealt with the topic of educational virtual environments. Various databases were searched for specific keywords (virtual environment, virtual reality, etc.) and a total of 53 studies, published between 1999-2009, were found. From this, the authors were also curios to know which articles followed the tenets established by Jonassen’ seven principles of constructive learning. 33 of the studies were conducted in an educational environment and 40 of the studies refer to STEM topics. Interestingly, the authors noted that social science instructors also used VR but to a different extent (contextualization vs abstraction).  In conclusion, the authors agreed with Higgins and Spitulnik (2008) when they stated that “despite the promise of technology for instruction, research to date has indicated that teachers tend to use new technologies to support existing practices, rather than developing methodologies more suited to the unique affordances of educational technology” (p. 778).

Ögren, M., Nyström, M., & Jarodzka, H. (2017). There’s more to the multimedia effect than meets the eye: is seeing pictures believing? Instructional Science, 45(2), 263–287. https://doi.org/10.1007/s11251-016-9397-6

The authors in this study are seeking to determine the effect that multimedia (text, pictures, graphs, etc.) helped or hurt participants in a Vector Calculus course.  Meaning, when presented with a problem statement and a graph, was the subjects’ attention or thought process hindered or aided by the presence of a graph. The authors “found no support for an overall multimedia effect (H1). Instead, graphs had a beneficial effect on performance only when problem statements were to be confirmed (instead of rejected), which is referred to as the picture bias effect (H2)” (Ogren et al., 2017, p. 281). When a graph was present, students made more use of other information sources, evaluated the problem statement to a closer degree, and made better use of creating more efficient mental models. Participants also paused more, or were silent longer, when contemplating problems that involved a graph.

Reeves, T. (1998, April 18). Evaluating what really matters in computer-based education. Reeves: Evaluating What Really Matters in Computer-Based Education. Retrieved February 16, 2022, from https://www.eduworks.com/Documents/Workshops/EdMedia1998/docs/reeves.html

The purpose of this article, as stated by the author, is to “describe fourteen pedagogical dimensions of CBE that have the potential to provide improved criteria for understanding describing and evaluating CBE” (Reeves, 1998, p. 2). The following pedagogies are discussed further:

  • Epistemology (Objectivism vs Constructivism)
  • Philosophy (Instructivist vs Constructivist)
  • Psychology (Behavioral vs Cognitive)
  • Goal (Focused vs Unfocused)
  • Experiential (Abstract vs Concrete)
  • Teacher (Didactic vs Facilitative)
  • Flexibility (Unchangeable vs Modifiable)
  • Errors (No errors vs learning from mistakes)
  • Motivation (Extrinsic vs Intrinsic)
  • Differences (None vs Many)
  • Learner Control (Restricted vs Unrestricted)
  • User (Access vs Engagement)
  • Cooperation (Unsupported vs Supported)
  • Culture (Sensitive vs Non-Sensitive)

The author then applies these principles to two programs developed in the late 80’s and early 90’s, Writing to Read and the Jasper Woodbury Problem Solving Series, and subjectively determines where the programs align in accordance with the stated pedagogies. In short, WTR is a highly structured, traditional, behavioral based approach to learning, whereas Jasper is more collaborative based and ground in constructivism and cognitivism.

Reiser, R. A. (2001). A history of instructional design and technology: Part II: A history of instructional design. Educational Technology Research and Development, 49(2), 57–67. https://doi.org/10.1007/BF02504928

The history of ID can be traced back to World War II and the psychologist who were tasked with creating training manuals and screening processes for our US soldiers. From there, and because of the war, the mid-to-latter half of the 20th century featured a whirlwind of change. Wars and a depression walked jointly with civil rights and technological advancements. New demands in the form of military aircraft and space shuttles created a need for more complex forms of learning. The needs of the time called our brains to action, and we responded accordingly. B.F. Skinner’s theory on behaviorism was being replaced by Gagne’s theory of the conditions of learning and Bloom’s Taxonomy identified our principal designs of learning. The 1970’s saw an increase in the amount of instructional design models. In 1983, David Merrill developed his CDT theory and PLATO was one of the first computer systems to be integrated into the instructional realm. The 1990’s brought Constructivism to the forefront, process improvement, prototyping, and computer-based training modules. E-Learning was becoming mainstream with the prevailing thought being that it would reduce costs and establish a set of e-learning standards.

Scarantino, A. (2003). Affordances Explained. Philosophy of Science, 70(5), 949-961. doi:10.1086/377380

For this article, the author is attempting to clarify some of the nuances of Gibson’s Affordance Theory while also expanding on the nature characteristics of affordances. Scarantino (2003) states that affordances are more than just a singular event in which “their manifestation is always constituted by an event in which the affordance-bearer X and the organism O are both involved” (p. 957). The author believes that a distinctive class of new affordances, goal and happening, to describe events that are happening to someone, or events that someone does. The distinction brings into account the differing degrees in which an event can happen, the current ability of the affordance, as well as the certain and probabilistic outcomes that could occur.

Schober, M. F., Conrad, F. G., Antoun, C., Bowers, A. W., Hupp, A. L., & Yanna Yan, H. (2020). Interacting with Interviewers in Text and Voice Interviews on Smartphones. Interviewer Effects from a Total Survey Error Perspective, 179–190. https://doi.org/10.1201/9781003020219-17

In reading through this synopsis of a study, it appears the authors are attempting to determine the effect that interviews have on their respondents. Due to the truncated nature in which the material is presented, it is difficult to try and expound upon the theory or literature behind the topic of inquiry (but here goes). Thirty-two respondents were asked questions in the following format: Human voice, human text, automated voice, and automated text. In turn, participants responded more to the human voice and text more so than the automated version. As a result, the following claims are presented by the researchers:

  • Text interviews decreased the social pressure of the interview, created less small talk, and allowed the respondent to answer when convenient.
  • Participants shared less when an automated system was the medium
  • “As people’s communication habits evolve—including increased interaction with automated systems—previous wisdom about effects of interviewers may change.”

Wells, A. J. (2002). Gibson’s Affordances and Turing’s Theory of Computation. Ecological Psychology, 14(3), 140–180. https://doi-org.libproxy.library.unt.edu/10.1207/S15326969ECO1403_3 In this article, the author is examining the different views on affordances from various writers, but of importance is Turing’s mathematical model that forms the basis of the paper. “The complementarity between animal and environment is captured in the way that the set of instructions relating affordances to effectivities specifies the way that the animal behaves. Turing machines have both structure and dynamics and are, thus, capable of providing models of the animal, the environment, and behavior” (Wells, 2002, p. 161). In applying a Turing machine application to that of affordances, it is the hope of the author to point out the relational and “that affordances are facts of the environment and facts of behavior, but that they are facts of the environment and of the agent that are determinately linked to behavior via effectivities. The configurations of Turing machines are causally associated with their behavior but are conceptually distinct from the behavior that they cause” (Wells, 2002, p.170). Lastly, the configurations that are created because of the application shows the balance between the organism and its environment