Published in The Australian, 28 September 2016
Universities are being warned they need to consider their ethical and legal responsibilities when using learning analytics programs, with evidence emerging of universities using students’ data beyond their original consent.
With the focus shifting to personalised learning and a keen eye on attrition, universities increasingly are turning to learning analytics as the latest innovation to help identify students at risk of failure.
But academics at Central Queensland University say “underlying ethical concerns” have emerged since they implemented learning analytics in 2014, in one of the first institution-wide deployments in Australia.
In a paper published in the journal Educational Technology Research and Development, the CQU academics say ethical or privacy issues were “only cursorily considered” in the development of the program, as the university was already collecting the student data at enrolment.
CQU’s program uses data from different university systems along with students’ activity on the online learning platform to identify those at risk of failure. It enables lecturers to send a personalised email offering support or assistance to those students.
Colin Beer, a lecturer in educational technology at CQU and the developer of its analytics program, told the HES there was “an institutional presumption” that student data consensually gathered at enrolment could be used in the analytics program.
“But really it’s beyond the scope of that original consent,” he said.
There was also a lack of transparency, as receiving an email through the system might have been the first time some students learned their level of engagement was monitored and they were at risk of failure, the paper notes.
CQU had now “tightened up” its language around the consent gained from students during enrolment, said Mr Beer, who is doing a PhD on student analytics. “Once students enrol they have a much better idea of how their data will be used,” he said.
While the use of learning analytics is proliferating across Australian campuses, a report last year for the former Office for Learning and Teaching found there was “relative silence afforded to ethics” in the local programs it surveyed.
This lack of attention did not reflect “the seriousness with which the sector should consider these issues”, the report found.
In Europe and the US, the growing use of learning analytics has sparked debate about the ethical implications of the programs.
Niall Sclater, a British expert in learning analytics and a former director of learning and teaching at the Open University said care and rigour were needed to ensure staff and faculty took their ethical and legal responsibilities seriously.
“Some early adopters have perhaps been so caught up in the excitement of their projects, sometimes driven by pressure from senior management at their institution, that they’ve neglected to consider these issues properly,” Mr Sclater told the HES.
Analytics programs were often driven by individuals who believed technology could solve their problems, and institutions did not put adequate resources into areas such as policy, raising awareness, staff training and evaluation, he said.
“Developing a learning analytics policy or adapting existing policies is essential — this should involve a conversation across key stakeholders, including students,” Mr Sclater said.
The CQU academics said they continued to grapple with the best approach to consent.
Celeste Lawson, a senior lecturer and co-author of the paper, said the university was exploring different options including allowing students to opt out of the analytics program.
Other universities would confront similar issues as they more fully adopted learning analytics, she said.