Discussion

Hypothetical Scenario for discussion

In the year 2022 All-Sydney University, Cabramatta, is having a crisis in student retention rates. The University Trustees have told senior management they must fix whatever is causing this, because if they don’t the State and Federal funding on which the university relies will be stopped, and that will compromise their capacity to offer courses and to fund student facilities. Senior staff in the University are in a panic but are unsure what to do. Student disengagement is a long-standing problem. And the teaching staff are at a loss about what to do; they’re having a hard enough time dealing with sudden curriculum changes and the introduction of digital innovations neither they nor the students fully understand. So, everyone is wondering: what do we do? Where do we even start?

Then the Deputy Vice Chancellor for Infrastructure has an idea. What if the new digital administrative and teaching systems aren’t the problem, but the key to a solution? Why not look at the rich range of data already collected by various automated and networked systems to figure out why students are dropping out of their degrees? There’s the data on GPA, grades, and attendance, and the data on misconduct cases and outcomes. More importantly though, in the two years since the University introduced a scannable student ID card system they’ve got a record number of individual data points. Those scannable cards register when specific parts of the campus facilities are used (gym; library; computer labs), digitally signals attendance in tutorials and lectures, logs purchases in the co-op and food court vendors (as the ID cards work for cashless buying), and record a full record of access to online learning resources. In addition to the data logged by the use of ID cards round campus, parking is monitored by license plate number recognition cameras at all the university entrances. The campus Wi-Fi network also tracks students’ internet use and can monitor individual student movement throughout campus grounds, using their mobile phone as proxies. All these new measures have over the last couple of years produced large quantities of ‘small data’ that—if machine learning and data analysis was used on that data—could identify the causes for the crisis in student retention: and with a high degree of accuracy about the specific individual factors that might make a student at-risk of dropping out. And once they know that, they can design intervention strategies to turn the situation around—promoting positive educational outcomes for all.

So, a local data science company, NoSeen, is contracted by the university to:

1. …identify predictors of student disengagement as an indicator for dropping out, and apply machine learning tools (text and data mining) to these predictors in order to flag at-risk students;
2. …equip teachers with fine-grained information to allocate resources and assist at-risk students by suggesting specific interventions, such as talking to the student directly, adjusting their workload and schedule, contacting parents or guardians, or automatically triggering contact by the student counselling service or academic course advisers;
3. …ensure transparency in the way the system is used.
The University gives NoSeen access to their existing databases, also providing access to new data as it’s collected. Implementing a broad policy of data analysis, NoSeen’s machine learning systems look at a large number of predictors for students who quit their study , ranging from various student demographics (e.g. race, ethnicity, gender, pattern of paid work, carer responsibilities, mobility, address), to academic factors (e.g. grades, GPA, examination results, history of disciplinary action, records of attendance); cross-referencing all this against data on teaching staff (levels of staff qualification, whether a unit is taught by casual/permanent staff) as well as data on units of study (Student Feedback on Unit and Student Feedback on teaching to identify where students might struggle; percentage of fail and absent fail rates per unit, per session of offer). NoSeen also harvested new data for a full academic year, allowing its machines to correlate the data of students who had previously dropped out with information about current continuing students in order to recognize patterns. NoSeen was then able to generate inferences that would not have been possible without the kind of data mining and analytics it was using.

Using all this data, NoSeen was able to identify eight key indicators that, in combination, predicted (with over 90 percent accuracy) whether a student would drop out of their study. The reasons ranged from administrative issues (e.g. too high study load or chosen wrong course, enrolment or progression issues) to external factors (e.g. balancing a job with school, domestic and carer responsibilities), as well as previously unconsidered factors (e.g. poor nutritional options on campus, distance to travel to campus). NoSeen then supplied teaching staff with profiles of at-risk students that helped them better understand why an individual student might be struggling, and that also suggested targeted approaches for helping him or her. These targeted approaches included things like tutoring, modifying assessment requirements, talking with the student’s parents or guardians and alerting students individually about carer support policies and processes, and automatically scheduling appoints with student counsellors.

By the end of the 2022 academic year, All-Sydney University had turned around its retention rate. The number of students graduating within three years of starting their course skyrocketed – and the dropout rate declined—from 21% to only 5% percent. In an interview with The Australian, the Vice Chancellor praised the work done by NoSeen, which he argued was instrumental to this improvement in student outcomes. This was the first public acknowledgement by the university of their use of intensive data mining and analytics.

Hearing about NoSeen made some students, staff and members of local community concerned. Not only had NoSeen been using student data to make its recommendations, but the whole thing had happened without anyone being directly informed about it. And the University are so impressed with the process and outcome that it plans to continue its data collection and use policies into the foreseeable future. Some teaching staff, students and parents begin to voice criticisms of, and opposition to, the new system—some of them confidentially expressing those concerns to journalists at The Guardian and the Sydney Morning Herald…

In 400 words approx., critically evaluate the hypothetical scenario above. Did All-Sydney University violate the privacy of its students by sharing their data with NoSeen? If so, is this breach justifiable, and on what grounds?

You should contextualize your response using the essential Lupton reading.
(HINT: it helps to remember the concepts covered in the week’s tutorial!)