Big Data in Education Needs Better Outreach, National Report Says
As a field, education researchers have a lot to gain from the rise of big data in education—and also a lot to lose if concerns over student privacy make districts and parents wary of data being collected and used.
That's why the National Academy of Education has called for researchers to learn more about how to protect data, and communicate with districts and parents better about how their students' data will be used.
"Education data use and education research have always gone together," said Andrew Ho, Harvard University education researcher and a member of the National Academy committee that wrote the report. "What's new is the curation and scale and preservation of the data. ...We need to keep an eye on learning process data as an opportunity that comes with a set of risks."
Big data in education generally falls into two categories. First is administrative data, the demographics, test scores, grades, and paperwork that schools have collected for decades, but which longitudinal state databases have made much easier for researchers to access. The second is "learning process" data, the often minute-by-minute information generated as students work through tasks. It might include keystrokes, logs of time students spend on individual problems, and even eye-tracking.
The report suggests that parents and even school staff may conflate administrative and process data. For any project, researchers need to educate school staff and the broader school community on:
- Collection, including whether the data is collected generally or for a specific study;
- Use, including the purpose and particularly the value of the study;
- Security processes in place to protect the data; and
- Sharing of the data, both in who will access it for study and how the results of the study will be provided.
Caution for Researchers Working With Companies
Academic researchers are trained in the ethics around working with people in experiments. But the report notes that researchers should tread carefuly when working with data collected by companies through online browsers, apps, and programs.
"Researchers should be not just asking for the data, but being an advocate for transparency among vendors," Ho said.
For example, back in 2014, Facebook saw widespread backlash from a study in which it altered teenagers' social newsfeeds to show them either more positive or negative news. The study did show that the site's users tended to post more positive comments after seeing a positive newsfeed, and negative comments when their feed was negative, but critics noted that the site had not informed users that it would be changing their feed, or allowed them to opt out.
"Facebook and Google do not have academic-style institutional review boards; they have nowhere near the commonly articulated standards that [academic] researchers use," Ho said. With the emotions study, he said, "The researchers got the data from Facebook, said, 'Look at this great experiment,' and published the results without thinking about the ethics of it."
"We can see our role as an audit, an evaluation of whether all the exposure students are having to these vendors is worth the risk," he said. "Education reseachers have to stand up for the standards we are held to."
- The Ethics of Educational Experiments
- Analytics in K-12 Schools: Big Data, or Big Brother?
- More Districts Mine Data to Refine Teacher Hiring
Want more research news? Get the latest studies and join the conversation. Follow @SarahDSparks