I was having a conversation with a scientist friend today about my doctoral research and explaining how part of my dissertation would involve grounding my research in a particular theoretical stance. He gave me a very perplexed look, which is not really a surprising reaction from someone who is not a social scientist. I’m sure it sounds very odd to some people to think that one’s findings and data are subject to interpretation based on one’s particular theoretical bent, but I do think it makes some sense when you’re dealing with social phenomena.
Anyway, I had written a blog post for one of my first semester doctoral classes that very much addresses this topic, so I’ve decided to repost it here. This blog was a response to one of our class readings, so it makes some assumptions about shared vocabulary – in case you’re not familiar, here are some terms you’ll need to understand this:
- epistemology: an understanding of knowledge and how it is constructed, which informs a theoretical basis, which in turn informs research methodologies and therefore specific methods. It’s sort of the scaffold on which knowledge is built
- objectivism: an epistemology that suggests that meaning and knowledge exist apart from human interpretations of things. In other words, there is one objective Truth that can be discovered.
- constructionism: an epistemology that suggests that meaning is constructed by human interpretation, in which case there may be multiple valid ways of constructing meaning from the same observations.
So with that in mind, here we go!
As I read Crotty’s description of epistemologies in his Introduction to The Foundations of Social Research, I naturally found myself wondering which epistemological stance most closely fit my natural approach to research and knowledge. On the one hand, objectivism deeply appeals to the scientist in me. Not only that, but I also spend my days surrounded by objectivist research in my work at the NIH. In biomedical research, there’s no room for constructing meaning about research. Either a drug works or it doesn’t; a bacterium is present in the culture or it isn’t; a reaction occurs or it doesn’t. Granted, there are plenty of ways to “game the system” to ensure you get your desired answer out of your data – outcome switching and p-hacking come to mind – but by and large, the professional world I currently inhabit is pretty strictly objectivist.
Nonetheless, I think that constructionism might be a better fit epistemologically speaking when it comes to my own research. The phenomena I’m interested in are highly social – how do researchers’ communities of practice, ways of constructing knowledge, and attitudes and experiences shape the ways that they interact with data, specifically in terms of sharing and reusing data? In fact, I argue that the social factors are possibly the most important barriers to the problem of data sharing and reuse. While it’s true that some technological barriers exist in this problem, at least work is already being done to address some of those problems. On the other hand, the social issues are much harder to quantify and therefore harder to address. I think when it comes to questions like the ones I’m asking, it’s really difficult to talk about objective approaches, so I find myself feeling drawn to constructionism.
So what’s a constructionist to do in an objectivist world? I guess partly I’m trying to situate myself within my appropriate research community and figure out how to interact with other research communities that take different epistemological approaches. Biomedical researchers may not be my research “tribe” insofar as our methods and epistemologies are concerned, but I’m keenly interested in how my research will eventually be perceived by this community because I expect my work will have implications for them, such as in terms of how they train their next generation of researchers to interact with data successfully, how they incentivize sharing and reward certain types of academic labor and research work, and even how they approach their work of data gathering at a very fundamental level.
I was on a conference call at work last week with a group that is charged with evaluating usability of a data catalog. It was easy to tell which of the participants were scientists by training. They were essentially trying to brainstorm how they could come up with a randomized controlled trial and determine some sort of gold standard for data discovery that they could use to compare to the new data catalog. It was obvious from the discussion that they were uncomfortable with the thought of employing what they considered “non-scientific” methods and skeptical about what kind of meaningful results could arise from qualitative approaches. They wanted hard, numeric data, and other types of evidence were not part of their approach.
Moving forward, I will be interested in seeing how I can reconcile the qualitative methods I may take and the constructionist approach I may adopt with the objectivist epistemologies commonly adopted with the biomedical research community.