Last night General Assembly had a great meet up called “Talk Data to Me” and brought together very impressive data experts from Coursera, Google and Clever, in addition to GA’s very impressive moderator Vrushank Vora, an expert in deep and machine learning.
Data science is an exploding field and has recently been impacting most industries globally and making a significant dent in education too. A recent article from edsurge (https://www.edsurge.com/research/guides/the-data-workout-how-it-s-impacting-teaching-and-learning) said “ If you think data—in education, or any field—is cut and dry, think again. Working with data in the classroom, especially, can be either exhausting or exhilarating—depending on your fitness level. Data can be big, but also quite small. It’s often quantitative, but is increasingly qualitative. It’s predictive, but not always inclusive. It’s private, but not always protected. But one thing’s for certain: data has enormous power to impact teaching and learning. “
According to this panel of data scientists, there’s no question that data, when used the right way can effect the course of education in numerous and positive ways. However, educators and administrators have historically been leery of too much data. Initially data was used to prove negative aspects of education such as ineffective teaching and led to loss of jobs and funding. Another deterrent to widespread use is expertise. Marc Harper, data scientist at Google said “ knowledge of how to understand and use data in schools is limited and is a road block for integrating powerful data systems”. Edtech companies must continue to innovate and demonstrate how useful and essential the right data can be and pave the way for educators.
Another point that stood out during the evening was that data was the key to improving personalized learning. Katherine Wong of Coursera said “ The more data we have- the more we can meet students where they are and increase personalized learning”. Fields like “Deep Learning” have made great strides in modeling sequential decision making for creating playlists; i.e. student specific plans for personalized learning. Though there’s been much progress in this area, there’s more work to be done to include social and emotional learning into PL curricula.
Everyone agreed that better, shorter, more frequent formative assessments (and digital tools for gathering and analyzing assessments) were needed – but that students should have less standardized testing which takes up at least three weeks per year!! Tests also need to be more nuanced and designed to determine how a student came up with a particular answer. Multiple-choice tests, though easier to grade are more difficult to glean any substantive data as to learning processes.
So much interesting stuff here! I can really see the possibilities.
nbdassociates 