Today I attended the LAK (Learning Analytics and Knowledge) 2016 doctoral consortium, where there was a wide range of topics covered by a diverse group of nine PhD students. You can find more detailed abstracts and links to papers/posters online here, but below are my very brief summaries of each project, as well as a summary of overall advice shared throughout the day:
Angelique Kritzinger, University of Pretoria, South Africa
“Exploring student engagement in a blended learning environment for first year biology”
Angelique’s work focusses on analysing first year student engagement patterns in a blended course. Her research uses demographic data, prior learning data and engagement data to their consider relationships with outcome variables (i.e. marks). A CHAID analysis found that semester 1 tests could be predicted by factors such as home language, gender, ethnicity and prior learning variables.
Elle Wang, Columbia University, USA
“Bridging skill sets sap: MOOCs and student career development in STEM”
Elle is interested in understanding learner motivation, achievement and interaction in MOOCs for career development in the STEM field. Her research analyses data from Coursera and edX iterations of a ‘Big Data in Education’ course.
Korinn Ostrow, Worcester Polytechnic Institute, USA
“Toward a sound environment for robust learning analytics”
Korinn asks: how we can use big data in education to improve products we are presenting to teachers and administrators? Her research aims to establish a framework for developing and evaluating online tutoring systems for K-12 education. (Interestingly – a need for this was specifically highlighted in the recent LAEP expert workshop)
Héctor Pijeira-Díaz, University of Oulu, Finland
“Using learning analytics with biosensor data for individual and collaborative learning success”
Héctor’s PhD focusses on the SLAM project. It’s aim: strategic regulation of learning through learning analytics and mobile clouds for individual and collaborative learning success. He uses biosensors (wristbands/eye trackers), VLE data, videos and pre-post test data in hopes of creating a biofeedback dashboard for student emotions to support their learning processes.
Garron Hillaire, The Open University, UK
“Self-regulation using emotion and cognition learning analytics”
Garron’s research highlights a need for understanding the relationship between emotions and cognition. In the face of self-report and observation challenges, he plans to use a combination of self-report, sentiment analysis and physiological measurements to understand what emotions can tell us about learning, how emotions are displayed in data, and how students can use their own emotional data for self-regulated learning.
Michael Brown, University of Michigan, USA
“Contact forces: studying interaction in a large lecture hall using social network learning analytics”
Michael’s research focuses on tools and artefacts in social learning in STEM higher education classrooms. Using VLE data, his work highlights how instructors and students use tools in social learning in large classrooms and which factors influence peer interactions. He aims to also use actor oriented modelling to consider the co-evolution of behaviour and the structure of relationships.
Jenna Mittelmeier, The Open University, UK
“Understanding evidence-based interventions for cross-cultural group work: a learning analytics perspective”
My own research considers social tensions in cross-cultural group work. Social network analysis combined with learning analytics data has highlighted that diverse social networks lead to higher quantity of group work contributions, and qualitative interviews have demonstrated differences in student perceptions by academic achievement level. In the future, a randomised control trial will consider evidence-based interventions that encourage collaboration.
Caitlin Holman, University of Michigan, USA
“Designing success: using data-driven personas to build engaging assignment pathways”
Caitlin’s research incorporates gameful learning, and uses learning analytics in combination with qualitative interviews to build ‘personas'(detail-rich characters that represent the target audience). Personas can then be made available to teachers to prepare them for incorporating gameful learning. Her research asks: How do different types of students make choices? What can we know about students at the beginning to design better courses for them?
Hazel Jones, University of Southern Queensland, Australia
“What are the impacts of adopting learning analytics in different higher education academic micro cultures?”
Hazel’s work focuses on adoption of learning analytics at an institutional level. She asks: How do academic micro-cultures adopt learning analytics methods for informing their teaching? Which data strategies and learning analytics approaches are most effective for informing teaching between different discipline groups? The process of adopting learning analytics is analysed along with staff engagement by the use of longitudinal surveys in combination with LMS data.
Summary of overall advice for PhD students from the LAK 2016 doctoral consortium:
- Embracing what you don’t know. One key suggest was that researchers shouldn’t be afraid of learning to do a new kind of analysis. Just because we might not be sure of how to do a certain kind of analysis, doesn’t mean we should exclude it as an option in our work.
- PhD students as future supervisors. As PhD students will in a few short years begin supervising their own students, they should begin to think now about wording criticism in a helpful way and encouraging peers to deepen their thought process.
- Moving from interesting research to powerful research. Research should have meaning and relevance to the field, and should push knowledge into new domains. Doing research for research’s sake is the wrong way to approach your work.
- Multi-disciplenary backgrounds are an asset. Oftentimes PhD students straddle several fields or methods, making it difficult to feel ‘at home’ with any particular approach. However, this flexibility should been seen as an asset rather than a burden.
- Soft skills are just as important as technical skills. Successful academics focus not just one what they say, but also how they say it. There’s a strong need for researchers who are charismatic and can communicate their work in effective ways.
- Searching for a job can be a self-reflection tool. The job search can help you assess your own skills and reflect on your strengths and weaknesses. Although it can be a frustrating process, one positive aspect is the self-awareness that builds through the process.
- Good supervisors encourage their students to gain valuable experiences. In today’s competitive job market, simply completing a PhD is not enough. Employers are looking for graduates who have done more, and good supervisors are the ones who encourage their students to diversify their experiences.
Sidenote: One aspect that I found particularly interesting in today’s doctoral consortium was its juxtaposition to the recent LAEP learning analytics expert workshop I attended in Amsterdam (see my summary of Day 1 and Day 2). In the workshop, several “action points” were highlighted by learning analytics experts as essential next steps for progressing the field. Although in the workshop, expert attendees often discussed barriers to incorporating these notions into their own work, many of the PhD students attending today’s LAK event are indeed incorporating these elements into their research. Thus, perhaps there is a logical connection between PhD work going on in the learning analytics field and the needs of experts.