Bass Connections EHD
Coursera and the Future of MOOCs
With the rise of open, online, publicly available education, educators have begun to question the legitimacy and practicality of this new form of learning (research opportunities, cost effectiveness vs. quality of education, completion rates, etc.). Our Bass Connections team in the Education and Human Development theme has developed and launched two Coursera courses in introductory chemistry and statistics, developing modules, investigating alternative means of conveying information, and probing the future of online education through mixed-methods research. Student team members are involved with all components of the project as well as serving as Community TAs for the courses. The syllabus for the 2014/2015 academic year for the research group can be found here. More information about Bass Connections EHD teams can be found here.
|Team leaders||Team members|
|Dr. Mine Çetinkaya-Rundel||Dr. Dorian Canelas||Maria Elena Carvajal||Kun Li||Heather Shapiro||Anthony Weishampel|
|Statistical Science||Chemistry||Chemistry||CIT||Statistical Science||Statistical Science|
Alumni: Brittany Cohen (Statistical Science), Abdul Latif (Religious Studies), Clara Lee (Chemistry), Shocao Mo (Theater Studies), Will Trautman (Chemistry)
What are common characteristics of the students, who complete the project, class, and qualify for distinction?
by Anthony Weishampel
Using supervised learning and clustering algorithms we are investigating common characteristics among the students who complete a goal in the DASI course. We have classified three goals for the students: the project, class, and complete with distinction. This analysis aims to provide a better understanding of the types of students that will complete the class.
Skills: Exploratory data analysis, big data, computation, machine learning.
Can we use sentiment analysis of student feedback to determine whether or not obtaining a certificate is the only way to measure success in Coursera courses?
by Heather Shapiro and Clara Lee
This project aims to develop a deeper understanding of who our MOOC students are, why they are taking our course, and what factors impact their success. With the initial help of Clara Lee, who has since graduated, both students interviewed roughly 20 students from both DASI and Chemistry courses. These students varied in age, experience, and geographical location. The next process is to code the transcripts from all of the interviews to understand more about the student experience. By breaking down specific phrases into different attributes such as positive and negative, we can then use this to perform a qualitative analysis on the data.
Clara and Heather conducted the interviews for the statistics and chemistry courses, respectively, and Heather is taking the lead on the analysis of the data.
Skills: Survey design, qualitative analysis, sentiment analysis, NVivo.
The IRB number for this project is C0103.
What is the best method for teaching R? Does the method depend on the student? Does the students’ background in statistics or programming affect which method would be best?
by Anthony Weishampel
On the development end, Anthony has converted the R labs from the Sta 101: Data Analysis and Statistical Inference at Duke into a format that is more suitable for Coursera. This entailed updating some of the labs to use datasets that might be more of interest to an international audience as well as converting the lab questions to multiple choice questions that can be automatically graded within the Coursera platform. He also developed a survey to be added to the end of each lab to gather information about the student experience with the labs, and reviewed the interactive web-based DataCamp versions of the labs developed by Dr. Çetinkaya-Rundel and the DataCamp team.
With data coming in from the first run of the course we are working on comparing the survey results and performance outcomes from the R/RStudio and DataCamp versions of the labs as well as investigating the relationships between students’ performance and opinions/experience with the labs. We are also investigating if these relationships are dependent on the students’ background of statistics and programing experience. We believe that these analyses will provide insights into effective methods of teaching statistical computation.
Skills: Instruction design, teaching computation, assessment.
How do engagement, SRL, and perceptions of motivational strategies differ between students with a mastery goal and a performance goal? Are students aware of their own cognition, behaviors, and motivation as well as their self-regulated learning (SRL) strategy usage in learning MOOCs? What are students’ perceptions of the motivational tactics that are applied in the course?
by Kun Li
Kun Li is working on a design-based study focusing on the problem of low levels of engagement in Massive Open Online Courses (MOOCs) and aims at mitigating this issue by incorporating ARCS motivational design model in MOOCs. She is also interested in examining MOOC students’ initial motivation and goal orientation when starting a course as well as how self-regulated the students are while taking the course.
ARCS, short for Attention, Relevance, Confidence, and Satisfaction, was first proposed by John Keller in the seventies of the twenty’s century and modified in his following publications. In ARCS motivational design, designers need to draw learners’ attention, and more importantly, hold learners’ attention for some time during instruction. Relevance means learners should be well informed why they need to learn the content, how the content is related to the learners needs. Confidence, the third factor in ARCS model, is the degree learners believe themselves can success; and it can affect learners’ success during the learning process greatly. Satisfaction is the degree that learners feel achieved and satisfied with their learning results (Keller, 1983, 1987a). The ARCS motivational strategies that were used in MOOCs are mainly implemented in the course announcement/email, course key pages and exercises features.
The IRB number for this project is C0534.
Skills: Instructional design, qualitative analysis, design-based research.
by Brittany Cohen (alum)
During the 2013-2014 academic year Brittany worked on developing applets that are used in the Data Analysis and Statistical Inference course. These interactive applets are developed in R using the
shiny library. The aim was to provide a visual tool for challenging concepts in the course. Links to applets are listed below, and all associated code can be found on GitHub.
- Distribution calculator
- Central Limit Theorem for means
- Central Limit Theorem for proportions
- Diagnostics for simple linear regression
Skills: Statistics pedagogy, R, shiny.
by Abdul Latif (alum)
During the 2013-2014 academic year Abdul Latif interviewed two chemistry professors and the videos (below) were released as part of the Introduction to Chemistry course on Coursera.