Democratisation and big data have established themselves as key developments in social research processes. But I’m wondering, are they pulling in opposite directions? Democratised methodologies immerse researchers within communities, undertaking relational work up close. Big data, on the other hand, has been described as a gaze from 30,000 feet.
Voices advocating radical challenges to traditional research practice through democratisation have questioned the model of research that positions the people who are the focus of study as subjects, and those who research them as experts who can analyse and evaluate. Under the democratisation paradigm, research seeks to serve the needs of those who’ve traditionally been excluded from positions of power, to highlight the voices of those who are disenfranchised on the basis of their gender, race/ethnicity, disability or other characteristics, and to further human rights. Democratised methodologies are concerned with ensuring that people who experience marginalisation influence research at every level of the process, to identify what it is that’s important to research, and how the community may benefit from involvement.
The other development, the increasing availability and use of big data, potentially creates critical tensions for democratising research methodologies and knowledge production. The potential (and seduction) of big data is the scale and availability of large sets of data that may be analysed; it promises easy access to massive amounts of data.
On the one hand, this may make access to data more democratic, with marginalised groups able to obtain material relevant to topics they have identified as important to them, and to engage in analyses with transformative potential.
On the other hand, while big data gives the illusion of providing unmediated and direct access to people’s beliefs and experiences, in fact it’s just as socially mediated and constrained as any other form of data. Indeed, the background frameworks structuring what knowledge gets collected in the form of big data and how it’s analysed may hide the transparent interpretation of human experience that’s central to democratising methodologies.
But can and how might democratisation of the research process and big data fit together? As big data enables new methods in knowing and defining social life, these new ways of knowing need to be critiqued for their limitations as they emerge. Perhaps it’s here that democratising methodology can step in and enable marginalised groups to have an input and make a difference.
I’m not sure that I would entirely agree with you because I would pose a different question – who is research for?
I don’t think democratising methodology is a one size fits all. At the level of community research, action research etc. where the methods and choice of methods is relatively accessible, then I would agree with you. But when we get to complicated, highly technical statistically sophisticated methods/methodology then this is a bit like democratising high energy physics.
But it is at the level of what we ask and for who we are asking it, some level of democracy is appropriate. What do we research and why? Can the results be made more open and accessible? Even then, I worry that democracy can itself be destructive of well being when people vote or choose in a knowledge vacuum. I think we have a home grown example in Brexit.
The problem with big data is not method or methodology (well it might at a technical level), but who owns it, who interrogates it and who gets to choose the questions.
I can see the strength of the points that you are making, and I wouldn’t disagree with all of them all of the time! I agree that the research questions asked of big data and who chooses them are central issues, but most democratising approaches wouldn’t draw a distinction between that and methodology – where methodology is an understanding of how we can go about gaining knowledge about how the world works, and thus how research should be carried out.
I also want to stress that the background to this is inequality and power and how it relates to methodology. There’s a long history of marginalised communities who have been problematised and stigmatised through the damaging assumptions embedded in the methodological approach and the methods used by social researchers. There is the potential that democratised research (just like democracy generally) may produce knowledge that’s problematic or skewed, although probably in different ways to traditional approaches, and that knowledge is open to challenge. But there’s just as much, maybe more, potential that it produces greater illumination.
There’s a tendency to see uses of ‘big data’ in social research as somehow antithetical to democratised research but I’m asking whether and how the two may be brought together. Attention to methodology seems to me to be an entry point.
I absolutely agree with your second point, but how to get there? Firstly, I’d reiterate we are at one in respect of participatory research, action research etc. Methods and methodology, should and can be democratised. ‘Big data’ and (say) complex administrative datasets do present a bigger challenge. In respect of the former, a current methodological question might be is the general linear model still an appropriate methodological framework? The debate on this is technical and statistical and one I find challenging, but it makes a profound political difference to what is asked and how it is asked. So, how can we improve the links of accountability and decision making between that data scientists and the public, in order to avoid dangerous assumptions/ stigmatisation etc?
The answer, for me, is threefold: firstly we can and should democratise data in society, through more informed critical reflection in the media and promote a better understanding of data in government and the third sector. School students must study English (or French, Spanish, Chinese, etc) and mathematics, but they also should study data in society to enable them to have a critical understanding of how data are used to persuade or govern.
Secondly and inevitably, at the technical and operational level methodological and methods choices must be made by a trained cadre of professionals. To deny this begs the question of why do we have social science degrees, PhDs etc? However, we should resist any shift to wholly methods and methodological training in social science at the expense of scholarship and critical reflection. It seems obvious, but all social scientists should have a critical understanding of the ideological and political context of data, methods and methodology. In the UK, at least, the danger is that universities and the ESRC are often dangerously focussed on measuring impact within a very narrow normative ideological agenda. So what we do and how we do it is so often decided by those who fund it!
Finally we should resist the colonisation of the analysis of big data by those without a social science background, for at best their questions may be trivial, but at worst informed by whatever is the current ideological fashion or project of government or big business.