Learning Analytics

A Special Report collection from The Chronicle, The Digital Campus: Big Data, examines the risks and rewards of using big data in higher education.

Bryant's EDUCAUSE Review article, "Everything Depends on the Data" (2017), asks interesting questions about access to and ownership of student-generated data. (One is reminded of Foucault's concept of "institutional gaze.")

The EDUCAUSE ECAR Working Group Paper, "The Predictive Learning Analytics Revolution" (2015), delves further into the use of learning data to predict -- and influence -- outcomes. This paper also opens discussions around organizational capacity and culture, as well as alignment between institutional policy and applicable laws and regulations with respect to the collection and use of student data.

Writing for New America, Ekowo and Palmer assert: "Without ethical practices, student data could be used to curtail academic success rather than help ensure it." Their report, Predictive Analytics in Higher Education: Five Guiding Practices for Ethical Use (2017), outlines critical institutional capacities for the ethical use of predictive learning analytics. Their recommendations include developing a vision and plan that anticipate and mitigate unintended consequences, engaging change management strategies to cultivate institutional capacity, developing data governance practices, pursuing transparency in data collection and analysis, and "intervening with care."

boyd & Crawford's conference paper, "Six Provocations for Big Data" (2011), poses questions of big data that we can consider in an educational context:
Do the numbers speak for themselves?
Is the data objective?
Is the data representative?
Does analysis account for data limitations?
Was data collection ethical?
Is there data equity?

Balkan is a self-described 'cyborg rights activist' [note: in both his and Elon Musk's mind, we're all cyborgs now] and founder of Ind.ie, an ethical technology design start-up. Balkan's perspective on the nature of data leads to the provocative conclusion that big data -- or data mining -- is a form of resource extraction best described as 'people farming.'

Gray and Citron approach these philosophical questions from the legal perspective in their law review article, "The Right to Quantitative Privacy" (2013). They contextualize data mining as a form of public-private surveillance and discuss its implications under the Fourth Amendment. Of particular interest to our discussion is section I: Quantitative Privacy: The Perils of Broad and Indiscriminate Surveillance, including their introduction of the concept of the 'chilling effect,' a brief discussion of the disparate impact of surveillance on minority populations, and the increased risks of false positives resulting from inaccurate data.

The Algorithmic Fairness project is exploring the issue of disparate impact and codification of bias in computer programs, analytics, machine learning, and artificial intelligence. One of their recent blog posts asks: "Do we expect more from algorithms than we expect from people? And is that reasonable?"

Recently acquired by Hobsons (parent company of Starfish), the Predictive Analytics Reporting (PAR) Framework seeks to aggregate "independent institutional datasets into a single data resource." A specific application includes data sharing between 2- and 4- year institutions to produce "cross-institutional metrics for accountability that consider student outcomes."

The algorithms which underlie big data analytics are often derived from statistical inference. In a timely long-form article for The Guardian, "How Statistics Lost Their Power -- And Why We Should Fear What Comes Next," William Davies discusses statistical reasoning as a technology in and of itself, and then traces its relationship to big data.