Despite robust exposure to ethics and human subject research in a variety of contexts and settings, I still get confused about IRB things sometimes. Also, I know that designing and performing internet research of any variety is an ethical tangle, which intensifies my lack of confidence. I learned today that my anxiety might be related to the low level of legal maturity found in learning analytic research. In other words, sentinel cases and other forms of precedent are still rapidly accruing, leading to high levels of moral ambiguity and the tendency for researchers to operate almost as they please until the law or public outcry tells them they have made a mistake.
Since I prefer to reduce my risk of being (ethically) notorious, I decided to take time out of my research activity to read about Internet research ethics in general and learning analytics research specifically. Two posts ago, I summarized the Association of Internet Researchers (AoIR)’s 2012 working committee report on Internet Research Ethics. It presented three internet-specific ethical concerns:
1. Studying participants in Internet culture is inherently different than studying human “subjects” in a clinical trial. There are different expectations for control and meaningful participation among many people on the Internet as compared to those submitting for medical testing. Internet participants are less vulnerable and possibly should be treated more like artists than vessels of study.
2. Privacy in the context of the internet is different than in other contexts.
3. It is challenging to identify the owners of data, who may (or may not) change over time and space.
This week, I was blessed with additional, learning analytics related ethics constructs to consider:
On August 25, Educause Review published a new ethical framework for Learning Analytics, which builds on the Pardo & Siemens article on the ethical and privacy principles for learning analytics, published just several months ago. The Internet gods were sending me a message.
The conceptual frameworks described in the two learning analytics (LA)-specific articles are consistent with each other and the AoIR report, although after reading the two LA papers, I concluded that the AoIR document is primarily meant for Internet ethnographic researchers. My assumptions come from the authors’ long, Lincoln & Guba–reminiscent approach to “human subject research.” In contrast, Pardo & Siemens deemphasized the issue of human subjects altogether, suggesting that: “The notion of subjects being put at risk by a study…is not that evident in learning analytics where large, often anonymized data sets are used for analysis” (pp 442-443).
Issues of privacy, however, are present in all three articles. Pardo & Siemens goes into more detail on privacy than the AoIR working paper, describing data collection as a “transaction” involving issues of trust, ownership, control (capability to influence flow of personal information), and limitations (possibility of preventing others from accessing information). The article provides a legal framework for guiding policy for LA, which includes:
2. Student control over data
4. Accountability and Assessment
Questions remain, particularly around the rights of students to change their data (control) once they see it (transparency). This reminds me of the same conversations qualitative researchers have about memberchecking.
Also, for clarification, “accountability” has to do with using the information effectively while “assessment” relates to institutional commitment to ongoing program evaluation and refinement.
If you don’t like this framework, there’s always Slade & Prinsloo (2013) which, despite being specific to LA, points out the importance of contextualization of data just as the AoIR working committee report did. The broad categories of ethical consideration (which are later broken down into six principles) are:
1. The location and interpretation of data
2. Informed consent, privacy, and the deidentification of data
3. The management, classification, and storage of data
And then finally, the article that triggered the email that made me think the Internet gods were smiling down on my desire to be ethical: Willis’ new framework for Learning Analytics and Ethics. While Slade & Prinsloo and Pardo & Siemens focused on legal frameworks, Willis went the philosophical route. This article does a lovely job of placing learning analytics within all the ethical lenses we know and love: utilitarianism, nihilism, utopianism, and ambiguity. There’s even a shout out to care ethics, which, to me, is a major component of the best motivations for learning analytics: LA researchers aim to provide a better, more holistic, more personalized education for all students. Or perhaps that’s just me being utopian.
I’ve just scratched the ethical surface, but I feel comfortable enough with my grasp of the major issues (if not any particular solutions) that it is time to move on. Next week (probably. if there’s not intervening email pointing me in a different direction.): What are Learning Analytics?