Stephen Whiffin: Balancing the potential for supporting educational practice with public privacy

Data sharing in education: complications in data ownership

Wearables have brought data collection on a person’s health outside of hospitals and into the everyday, collecting information on movement, heart rate, sleeping, and even blood oxygen levels—it has also driven discussions around the sensitivity of that data and how it might be re-purposed. Education data holds some of these same characteristics and is arguably all the more sensitive given the number of potential data points that are relevant to educators and the early age from which data collection begins.

What is educational data?

At its core, it’s easy to think about educational data as recording the results of standardised tests. This is, in fact, data that has been collected on students for decades and feeds into an evaluation of how their education is progressing. But in jurisdictions like Canada and Finland, data collected on students is much broader, focused on developmental indexes. Schools can ask parents to collect data on how much time a student spends watching videos or engaging in other activities. Schools can also monitor social media, where activities outside of school feed into an evaluation of student behaviour and need. School data isn’t just coming from the classroom.


In some ways, the data collected in the classroom is arguably more chaotic than what you find outside, which creates an even wider diversity of data in education. Teachers and schools collect, use and manage data for purposes individual to each classroom—upwards of 90% of data collected within schools on a student's work and performance can be unstructured.


What are the advantages to collecting this additional data?

This data, collected both inside and outside of the classroom, is being used to bring tangible improvements in education. Evaluation of student performance is done to identify where students might receive additional support so that they can improve at both a scholastic and social level. And as the data becomes richer, it affords greater opportunities to personalise and improve that support. Concretely, identifying learning disabilities takes years and costs large amounts of resources, but a smart sifting of a student’s data—perhaps with the support of learning algorithms or artificial intelligence—can shortcut that process to get children the help they need earlier.


Do students and parents control their data?

Within the General Data Protection Regulation (GDPR), the concept of consent is clear, but in an educational context, students (or parents) do not have a right to opt out of most data collection that will take place at an educational level. School systems will have obligations around the transparency of that data giving parents the right to request the data being collected and also transparency from schools on how that data is being used. But there is no tick box for parents to opt out in the same way that students cannot opt out of attending school until they reach a certain age. This creates added pressure on school systems to protect data and use it in an appropriate way.


How can data sharing work?

Given the sensitive data being collected—and the fact that these datasets are largely mandatory for all students—school systems have an obligation to ensure that data is being used in an appropriate manner. This makes them naturally more sensitive around sharing data with other organisations—whether that be other government departments or private-sector education providers—so as to improve education systems in general. It leads naturally to data silos and provides a built-in justification to avoid data sharing whenever possible given the sensitivities involved. It also means that when data sharing does take place, it raises questions of how providers are using that data and if it’s being used for other purposes, whether educational institutions have a right to withdraw access to that data.

This is one reason why data sharing agreements in education need to both clarify ownership as well as a right to withdraw access to datasets. This brings complications for third-party organisations looking to experiment and innovate, as the data’s sensitivity means that educators tend to the narrow scope of how data will be used and disposed of after use. Strict scopes on how data will be used can limit innovative potential, and mechanisms to adjust sharing agreements that allow scope changes while still respecting the data’s sensitivity should be explored.

Name

School District #43 (Coquitlam)

Sector

Education

Region

North-America

Countries

Canada

Time

-

URL

https://www.sd43.bc.ca

Business model

Education provider

Participants

Stephen Whiffin

Type of organisation

Educational organisation

Data sharing model(s)

Between education institutions (intra-sectoral)

Core impact

School District #43 helps set up data sharing agreements in education to clarify ownership as well as a right to withdraw access to datasets, respecting the sensitivity of the data.

Context

Teachers and schools collect, use, and manage data for purposes individual to each classroom. Identifying learning disabilities takes years and costs large amounts of resources, but a smart sifting of a student’s data—perhaps with the support of learning algorithms or artificial intelligence—can shortcut that process to get children the help they need earlier.

Stephen Whiffin: Balancing the potential for supporting educational practice with public privacy
Image credit:
2020, Robo Wunderkind

For questions and comments, please visit our forum on Futurium.