Small Data in A Big Data World: An Insight into the learning Science behind McGraw Hill Education Products
This presentation was part of the Big Data series from the The Alexander von Humbolt Institute for Internet and Society as well being supported by The Vodafone Institue The institute aims at a better understanding of the interdependence between the society and the Internet. You can see more photos here
By partnering with educators around the globe, our learning engineers, content developers and pedagogical experts are developing increasingly open learning ecosystems that are proven to improve pass rates, elevate grades and increase engagement for each individual learner while improving outcomes for all.
The vision is to unlock the full potential of each learner.
The mission is to accelerate learning through intuitive, engaging, efficient and effective experiences – grounded in research.
At McGraw-Hill Education the contribution to education is that the key to unlocking a brighter future lies within the application of our deep understanding of how learning happens and how the mind develops. It exists where the science of learning meets the art of teaching.
Educators have been and always will be at the core of the learning experience. The solutions we develop help educators impart their knowledge to students more efficiently. By harnessing technology we can enhance learning inside and outside of the classroom and deepen the connections between students and teachers to empower greater success.
Ever since the first recorded human civilizations began, we have always looked for ever more sophisticated forms of gathering, collecting and storing the vast amounts of data we produce. From the very early writing systems to vast storage units in the industrial age, up to our present day real-time, social and sentiment tracked lives, we have diversified Big and Small Data to suit our needs.
To this day, Big and Small Data remains sometimes elusive. I am sure many of you here are maybe experts on the topic and whilst I do not claim to be, as I am primarily an educator and consultant, I will endeavor to give you an overview within the scope of McGraw Hill Learning Products.
So what does Big and Small data actually mean and what can it do in the context of digital learning and education today?
Big Data can be seen as a blanket phrase that raises issues of security and accessibility but also how to sort, compress and understand as well as store masses of seemingly unrelated information. It can also be understood as a collection of data sets that are so large and complex that it becomes difficult to process, using on-hand database management tools or traditional data processing applications
Today, it is about collecting as much relevant data from as many sources as possible, analyzing them in real time, and making an optimal decision based on current circumstances. We have to be aware that the decision is never absolutely 100% accurate and always has an element of error due to human judgement.
When analyzing vast data sets throughout history, Big Data can be defined by the three V´s: Volume, Variety and Velocity three V’s: Velocity, Variety and Volume. was identified by a 2001 research report, META Group (now Gartner) analyst Doug Laney defined big data as being three-dimensional, i.e. increasing volume (amount of data), velocity (speed of data in and out), and variety (range of data types and sources) other data scientists such as Mark van Rijmenam is Founder of Datafloq have argued that even these three V’s are not enough to analyse that mass of data and go further. This is beyond the scope of this presentation, however, I would encourage you to look further into that. So going back to the 3V’s the elements increase through each historical period, correcting previous assumptions about the Big Data sets.
The three V´s can be applied to different topics so that we can understand better, how big data has historically changed.
So to go back to what is small Data? Small Data is usually defined by the fact that it is not big data. Information is collected for a specific purpose. In this case to improve specific learning outcomes and the curriculum. The is about how the students interact with the curriculum to see whether the student has mastered the objective and how confident the student is about his or her mastery of the material, the meta cognition. It does not include family history, medical records or retail purchases. It doesn’t include all the other websites that the student has visited IT only includes the student’s interaction with the curriculum. However small data is only part of the story.
We can compare Big Data and Small Data in this way.
The major part is that education is inherently social and dialogical. There aren't enough teachers to go around and give the kind of personalised learning that students need and unfortunately the model doesn`t scale.
The use of small data instead of Big Data was a deliberate one by MHE, in a world where many organisations, such as retailers are using big data instead. With small data, students and teachers can navigate the knowledge spaces in a more effective way by providing purposeful questions that drive outcomes. MHE collects the minimum amount of data in order to understand what those outcomes are and how they can be achieved.
Currently Data science is active within education to the point where we can strike an analogy of the development of the telescope in Astronomy and the way that changed how we saw data in space to the way in which education is changing.
Taking our view of the universe as an example we can see how Galileo´s heliocentric views affronted the authority of the Catholic Church. Taking it one step further, we can even see how Galileo`s efforts, by today`s standards have radically changed.
As time passes, the volume, variety and velocity of Big and Small Data changes, so humans can make ever more accurate analyses, assumptions and decisions.
Developments in Astronomy have identified a new position of where our Milky Way is in relation to other galaxies.
It has become known to us that we are actually part of a massive super-cluster that means we are only one of the 8000 galaxies that surround us. This helps us not only to understand our current position in our vast universe but also more about where we came from and where we are going.
With current tools it is possible to explore observations that do not conform to an expected patterns or assumption. So by analyzing Big and Small Data in Education like the universe and our place within in we can explore new frontiers that we couldn`t even imagine, once beyond, are now within our reach.
Small Data can be defined in that it is collected for a specific purpose. By gathering data about the learning outcomes it is possible to enable rich findings, as we analyse the volume of that specific learning data, the variety and the velocity.
- What is Small Data?
- Strategic use of learning data with a direct correlation to performance
- Assignment scores, time on task, progress in an adaptive learning environment
Data and analysis in service of the learner and instructor. The way in which that data is collected in the system is to help both the instructor know how they can support the student’s learning objectives but also then how the students can understand where they are struggling and in which areas they need to gain support or do more reading or ask more questions. So the small data acts as a foundation to support both the instructors and learners instead of being the be all and end all. Small Data can begin with adaptive learning and end with analytics and insights, but of course that is only part of the story, the other part is how then the teachers and students can interact with that data in a more meaningful way.
Some of you are probably aware of the increase of blended learning environments and the flipped classroom approach.
In traditional teaching and a study by study by Sally Mann (2009) more than 59% students found found all lectures in their course boring and 79% find some lectures boring.
Because of the ability to use the adaptive learning approach and gain insights through small data, instructors can use the flipped classroom approach so that the foundational lectures are done at home. This is doubled up with gamified activities and foundational exercises that builds upon their conceptual knowledge but then allows for the students to become more engaged with the content, by realising what some of their pain points, misunderstanding or challenges may be. As the students interact with their coursework that is assigned by their instructor the system collects formative data on student performance and engagement through peer grading, mastery based assignments as well as ungraded formative assessments. This velocity, variety and volume of the data help the lecturers in turn understand more about where they can support their learners before the class and continually through the course starts so that they can create more engaged learning spaces where the activity is Focused in the Classroom, the students are actively Engaged in learning, Efficient there is more Teacher/Student interaction, there is collaborative learning as well as more time for assessment and activity, more time for questions and deeper understanding of concepts in the course which allows for more agency in the students learning as they are able to integrate previously learnt concepts with the newly acquired concepts.
With learning analytics we are able to understand better when the students are able to move onto the next topic, but also when students may be at risk of not completing the course and even what grade the student is likely to receive without intervention from the teacher. It is worth remembering at his point that like big data, small and specific data can has margin for error. These analytics that give feedback and are predictive cannot be taken in isolation, but rather used effectively as tools understand and consideration with other elements such as the social environmental factors then we can gain perhaps a more accurate understanding and build better learning environments for teaching and learning.
How does this help the student. Imagine that the student has grown up with a fear of mathematics, but with the small data and adaptive learning the student can be better informed about the way in which the student can stay on the path to positive progression so that the maths become not only achievable but also fun and then the student turns into a confident learner.
Beyond that the students paths through the curriculum are fed back into the system to drive the further improvements into the curriculum. Not only then are we able to help the students become confident learners we are also able to get feedback on the learning moment. we have information at the learning objective level about how the student is doing so that we can support the learning and teaching in a more effective way.
For example if several students are struggling with a particular learning objective, that has been identified through the system, through quizzes and meta cognition, the time spent on the questions and the level of mastery (which would see how many questions were needed before they were able to understand the concept) the teacher could create space in the lesson to focus on the most challenging learning objectives for those students whilst designating other learning objectives to the students who were not struggling in this area.
Finally small data needs to be taken as it is, a tool to understand and more effectively teach and learn, not something that can solve all the problems at once. We are on the verge of gaining much better insight into how we learn, but we need to proceed cautiously and with small baby steps, taking into account the risks and regulations to protect those who seek to learn as well, but I think it is important to move forward, always preciously balancing our needs to understand with awareness of those ethical concerns that are taken into account as we act as pioneers into this complex and ever changing learning universe.