If I had to make an early prediction about a hot trend for 2012, I'd point to "data." It's an easy pick: big data is poised to be a big deal regardless of industry in the coming year. Thanks to the data from sensors, cell phones, Web browsing, social media updates, photo and video uploads, credit card transactions, online purchases, e-book readings -- you name it -- we are producing a mind-boggling amount of data. One of the big opportunities in the coming years will be to make meaning out of all the data that we create and store. That's true in banking. It's true in commerce. It's true in healthcare. It's true in education.
Interest in educational data and analytics has been evident already this year. Nationally, we seem rather fixated on standardized test scores. Both Knewton and Grockit, two adaptive learning test prep startups, raised substantial rounds of investment this fall. And earlier this month, Knewton announced a major partnership between it and educational publisher Pearson, meaning that Knewton's adaptive engine would power Pearson's digital curriculum -- content that, according to the company's figures, reaches about 9 million students.
That's data on 9 million students -- how they read, how they progress through course content, where they require remediation, how they display knowledge gain. It's incredibly valuable data, particularly at that scale and particularly if it's actionable. That is, if Knewton (and Pearson) can actually deliver personalized course content that helps bolster student test scores.
Indeed much of the emphasis on student data has been on this sort of thing, creating algorithms and models to help predict students' responses. Earlier this month, Khan Academy intern David Hu blogged about some of the improvements he was working on to make the non-profit's study platform handle "mastery" better. Just this week, Grockit posted a challenge on the popular new data contest site Kaggle, supplying a dataset and asking the data scientists and machine learning researchers there to "improve the state of the art in student evaluation by predicting whether a student will answer the next test question correctly."
And now the education platform EverFi has indicated its plans to develop its own data analysis toolkit. The company says it plans on investing $2 million in a research initiative called Assembly that will offer students and schools insights into digital learning. While EverFi doesn't boast quite as large a dataset as Pearson's 9 million student consumer-base, EverFi is definitely operating at a large scale and it does have a wealth of student data. The company announced earlier this year that it has certified some 4 million students through its curriculum, and its products are in 46 of the 100 largest school districts and more than 500 colleges.
EverFi's curriculum is different than Pearson's. While the latter is focused on core curriculum areas (the partnership with Knewton deals with just math, reading, and writing content for now), EverFi's platform addresses other critical skills. These are skills that fall outside students' academic study: financial literacy, digital literacy, student loan management, and alcohol and substance abuse. Often the EverFi courses are part of freshman orientation or other back-to-school programs, and they're a blend of online and offline software.
The types of classes that EverFi provides have different goals and priorities than other sorts of college courses. It isn't so much about knowledge gain per se. It's about identifying risk (risk of student loan default, perhaps, risk of alcohol abuse). It's about changing behavior. As such, it's easy to see how a data platform built by EverFi think of learning analytics in a different way than a company like Knewton. What role does student attitude play? What role does student motivation play? EverFi's Director of Big Data and Research Todd Wyatt told me, "We're hearing the entire student voice."
Wyatt says that EverFi's Assembly product will provide students real-time, personalized feedback so, for example, they can see how their attitudes compare to other students' when it comes to things like alcohol, drug, or credit card usage. EverFi hopes too that by looking beyond just student test scores, that it will be able to offer school insights into aspects of student learning and behavior. It means a different sort of intervention or remediation than the type that Knewton and Pearson offer.
And that prompts other questions: what are colleges going to want to do with all their student data? What sorts of insights will they want? What sorts of action will they want to take on it?