Many researchers and educational research institutions have recognised an ethical dimension in terms of personal privacy within the practice of learning analytics (LA). Examples are Avella, et al., (2016), Pardo & Siemens, (2014), McNeil, et al., 201, Slade & Prinsloo, (2013), Cardinali, et al., (2015), and Kay, et al., (2012). In 2015, JISC released a Code of Practice for learning analytics, drawing on data protection legislation, emphasising the need for educational institutions to “ensure that learning analytics is carried out responsibly, appropriately and effectively, addressing the key legal, ethical and logistical issues which are likely to arise” (JISC, 2015). These, along with the upcoming EU General Data Protection Regulation, requires LA practitioners to take stock.
It is of interest then to evaluate what concern is given to issues of personal data privacy and related ethical aspects in existing published academic literature involving learning analytics.
The author systematically searched for publications from academic peer-reviewed journals using the search term “learning analytics”. This task was challenging in several ways. Some articles were not empirical e.g. literature reviews, framework constructions, opinion pieces, etc. Some addressed multiple research cases. An example is Clow, (2013) who addressed a very large number of studies but none in any detail, meaning it was not feasible to expect coverage of ethics / privacy therein. Articles on MOOCs were omitted as their massiveness and the nature of student interaction therein is different to groups that are more traditional.
In total, 52 articles were deemed suitable. It must be emphasised that the search could not be exhaustive. They ranged from the years 2012 to 2016 inclusive, with one article from 2010 and one from 2017. Each article was perused with the view to identifying what (if anything) was said about participant ethics and privacy.
39 articles did not mention the ethical process undertaken with respondents, or privacy of their data. That is, three-quarters did not deem privacy important enough to discuss in their methodology. While this is not to suggest that ethical considerations were not taken, it is a surprisingly high proportion.
The remaining 13 varied in their coverage. At the lowest level were those who merely mentioned privacy as being important but didn’t actually say what they did. The most impressive, Dyckhoff, et al., (2012) drew on data protection law and recognised that research involves a trade-off of “data privacy” versus “pedagogical useful indicators". They curtailed some aspects of their research because of privacy considerations. Others carried out tasks such as anonymising the data but it was not always clear if the students had been asked for their permission to be part of the study.
Ultimately, personal data is becoming the life force of any aspect of education that moves online. There is a large-scale task involved in ensuring that we as educators are sufficiently responsive to and cognisant of student rights to data privacy. We do not want a situation where students are overly willing to give up privacy, and educators overly willing to take it from them.
