(SpringerBriefs in Business Process Management) Learning Analytics Cookbook_ How to Support Learning Processes Through Data Analytics and Visualizatio
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
18 3 Responsible Cooking with Learning Analytics
Therefore, people may not act according to the privacy preferences they claim,
and they are usually unconcerned about data protection and privacy until their data
are breached (Spiekerman and Cranor 2009). Users’ concerns about privacy also
depend on the type of data that are collected, the context within which they are
collected, and their perceptions of the dangers in disclosing them (Pardo and
Siemens 2014). The problem with this paradox effect in terms of learning analytics
is that designers, developers, educators, and learners tend either to underestimate or
overestimate the effects of and the importance to data subjects of privacy and data
protection.
Tene and Polonetsky (2013) investigated the fundamental principles of privacy
codes and legislation and argued that the principles of data minimization and
individual control and context should to be somewhat relaxed in a big-data context,
where individual data points are not connected to individual data subjects, because
of the potential benefits to society (e.g., public health, environmental protection)
and, along with society, themselves as individuals. At the same time, issues of
transparency, access, and accuracy must not be neglected. Tene and Polonetsky
(2013) discussed the distinction between identifiable and nonidentifiable data and
considered de-identification methods (anonymization, pseudonymization, encryption,
key coding) as important measures in data protection and security.
Schwartz (2011) developed a set of ethical principles for data analytics solutions
based on a series of interviews with data privacy experts, lawmakers, and analysts.
These principles include a set of overarching ethical standards related to compliance
with legal requirements, compliance with cultural and social norms, accountability
of measures tailored to identified risks, inclusion of appropriate safeguards to protect
the security of data, and setting of responsible limits on analytics in sensitive areas or
with vulnerable groups. In addition to setting up these generic principles, Schwartz
(2011) argued that ethical principles must be adjusted to the various sub-stages of the
analytics process, so the rules for how to tackle these challenges should be tailored to
each stage of the analytics while also seeking to maximize good results and minimize
bad ones for data subjects. Thus, in the data collection stage, care should be taken in
regard to the type and nature of the information acquired, particularly in terms of
avoiding the collection of sensitive data, while in the data integration and analysis
stage, sufficient data quality must be ensured and anonymization should be
performed. Finally, in the decision-making stage, the analytics results on which
decisions are based must be ensured to be reasonably accurate and valid.
3.3 General Ethical and Privacy Guidelines
The OECD provides comprehensive guidelines and checklists for those who are
seeking guidance on how to deal with privacy issues in analytics technologies and
other systems (Spiekerman and Cranor 2009; Tene and Polonetsky 2013). In 1980,
the OECD provided the first internationally agreed collection of privacy principles
that harmonized legislation on privacy and facilitated the international flow of data