When you are using MOOC learning, they learn with MOOC study - Massive open online course

When you are using MOOC learning, they learn with MOOC study
What is  MOOC? stands for Massive Open Online Course
MOOC is an online learning wave up from the end of 2011 originated in Silicon Valley. This wave is not only open up a new era of student participation in online classes, but also opens up a new field of scientific research --MOOC. Scientists from the system by means of massive amounts of data and the opportunity to join the MOOC experiments to explore everything in the course of "learning".

In the current issue of the journal "Science", Harvard University School of Education Anchorage Justin (Justin Reich) published a review article, emphasizing the promotion with MOOC deepening study conducted MOOC also need to begin the necessary changes. How to improve learning outcomes, how the curriculum is more efficient? To clarify these issues, MOOC researchers need to design better research system. source

(Maya Blue/ Compile) in edX · CEO Anant Agarwal (Anant Agarwal) seems large open online course (Massive Open Online Courses, referred MOOC) should become a "learning particle accelerator"[1] . In recent years, as more and more people pay attention MOOC around the "learning" to start a new research direction also will be started. However, the results of these studies for teaching little inspiration - and not because of big data "big" will be able to answer questions of interest to us.

MOOC want to effectively promote research and learning about scientific developments, researchers, curriculum developers, and other leading staff needs to be improved in three directions: transfer from research to the study of the participation of the learning process itself; from individual Comparative study of research to curriculum across the background, after engaging analysis (post hoc analysis) steering multidisciplinary experimental design.

Learning, or just click on?

What teaching measures being able to explain how to improve learning outcomes can MOOC study rare. We have a TB level mass data tell us what online click on the student, but the student's brain changes what we know almost nothing.

4 study online learning platform Udacity, Khan Academy (Khan Academy), Google curriculum developers (Google Course Builder) and were used as an example edX[2-5] Relevance, they have to assess each student's research and action learning (such as test scores or course completion rate) between. These four studies have taken a similar approach to measure the behavior of students, will come down to the massive data into a simple, individual-level summary of the variables, such as the number of attempts (Udacity) answers to the exercises, long (Khan Academy) during site visits, weekly amount of task completion (Google) or each student clicks (edX) event log. In studies conducted in these platforms, the complexity of the behavior of the learner is lost[6] . By simple comparison or regression, these four studies come to the same conclusion: the learner's level of activity was positively correlated with their learning outcomes.

But we do not need to prove the event log on the trillion effort associated with success. Causal link "learn more" and "learn better" between the still unclear. In addition to warn students more actively involved in learning, they do not provide any practical suggestions for curriculum design. MOOC research need to adopt a more diverse generation of research design, to explore more deeply the reasons to promote student learning.

Watch the video does not equal learning

Early studies only focus on one of the reasons MOOC student participation and course completion rates is that most programs do not have the right MOOC assessment framework to support study and research. Ideally, MOOC researchers grasp assessment data should have three characteristics.

First, the assessment should be carried out in a number of different times. In MOOC course, crucial test before class. You know, the course participants varied background in both the novice, there are professionals in the field. Secondly, the assessment data should cover many aspects of the learning process, to get a high score in the test students often quantified in terms of conceptual understanding is generally stagnant or professional thinking. Finally, the assessment methods should be included in the curriculum has been proven effective research, and teaching effectiveness to be compared to other settings.

Some recent studies have reached MOOC these requirements, they help us understand what the learners benefit most from MOOC, what teaching materials to learn the most helpful . Voluntary participation in online learning, the distinction between "participation" and "learning" is especially critical. Want to correct the misconception that we have crossed the chasm between intuition and scientific facts.

Unfortunately, learners may prefer to see the video - they use a more simple way to show the course material. Although students can directly eliminate misunderstandings media to learn more, but their evaluation of teaching videos more active, because they intuitively describe the phenomenon . Student participation data collected may pursue courses, participation of developers to create media experiences pleasant, but let the students intently watching a video, does not mean that they can learn something.

MOOC poke superficial research needs a new generation of data to explore the participation of learners practical learning process to optimize this process. Source: Science

Rethinking Data Sharing

Another problem is that the student data sharing. Regulations to protect student privacy, trends and monopoly consider data protection aspects of data are prevented sharing of data. Although researchers can analyze the differences (in the same course) among students, but cannot compare the differences of different courses. For example, Neiss Turk (Nesterko) and other people in the study found that middle course exists between job submission frequency and completion rates positively correlated MOOC, But they studied 10 courses not only arrange different job submission deadline on course enrollment, course content and other aspects there are also differences. To conduct meaningful teaching methods post hoc analysis, data from hundreds of course essential.

But share learner data is not a simple thing. Some recent studies have various needs for data processing to personal characteristics, in order to protect student privacy, found that if you want to ensure that no leakage of student status required obfuscation will greatly distort the data, causing it to no longer meet the study. To achieve the purpose of scientific research MOOC data sharing, a large number of policy adjustments need to be made, at the same time develop a social science data sharing new technologies.

One measure could be to protect the privacy and anonymity of data separation, which allows researchers to share identity data, but the data protection system researchers for more stringent regulation. One of the technical aspects of the solution is the difference between privacy (differential privacy), that enables organizations to save student data in a standardized format, this format allows researchers to query the database, but only the result of an integration of the output.

Beyond A / B Testing

Due to the lack of comparable data across the curriculum, only limited experimental designed to assess the effectiveness of specific teaching methods. MOOC courses at the earliest, the researchers joined the A / B test (Translator's Note: A / B test refers to the same target set univariate two programs, the implementation of test results comparing the two programs) and other experimental design . One of the most common areas of experimental interventions is a universal built-in test. For example, a study set up in the discussion area MOOC medal system, active in the discussion area of ​​student performance can be a virtual medal. The students were randomly assigned to study in different forums set up medal, which inspired a number of forums more forums.

The experiment was a machine-learning courses conducted, but it can also be carried out in the United States or biological literature courses. These experiments have areas of universality has been widely applied in MOOC research. However, this universality has limitations: The study was unable to push forward the scientific discipline of learning, can not get the optimal learning order to find the best way to doubts.

Motivation of learners in a well-designed curriculum is undoubtedly a good thing, but if the overall teaching means a MOOC course is wrong, then the built-in test will only increase the useful work done by the learner. Research on teaching specific subjects designed to explore MOOC learning mechanisms in the discipline, they may be a prerequisite for effective universal interdisciplinary research.

In order to assess the measures of teaching, instructors Fisher (Fisher) teaching. "Copyright"This door will assign students to a class of two courses, a curriculum design based on the US case law, the other main global copyright issues. He adopted the final exam scores, student surveys, assistant feedback to assess the effect of the course, and draw conclusions: Compared to research on global issues, for in-depth exploration of a single copyright system to help more students. The conclusion of the legal aspects of online education provides a feasible measure.

With MOOC study matured in specific areas of research and pervasive are important, but research in specific areas may require more effort. Built experiment more easily into the academic structure, and focus on specific areas of research requires an interdisciplinary team, including content experts, instructional designers, and expert assessment, but this aspect of the continuing shortage of talent[18] . MOOC more complex research needs support from educational institutions to carry out, including universities and funding agencies.

Study on Improving the threshold

MOOC courses for the first group, it allows millions of people to get basic course material is undoubtedly an achievement; while the first batch of MOOC researchers, the data can be purified for analysis is also a success. In the early stages, towards the path of least resistance is to explore a very sensible approach, but it may also make the study a "inertia", braving the risk of a road went black.

MOOC for such a young field of research, the use of participation data, using data from individual courses, using a simple built-in test is reasonable experimental design strategy. However, if you want to continue to develop this area, researchers need to overcome those early studies avoided out of problems. These problems can not rely on individual researchers to solve. To improve MOOC research requires the joint efforts of universities, investment institutions, journal editors, conference organizers and curriculum developers.

Universities should give priority to those programs to solve a fundamental problem in the field of teaching and design. Journal editors and conference organizers should give priority to publish across agencies, research study to explore the effect, not only concerned with the participation of indicators; design should focus on research and experimental design, rather than just focusing on the post-mortem analysis. Investment institutions should follow the above principle of giving priority to education has the potential to change the open scientific institutions to support entrepreneurship. 

Share this

Related Posts