by Leon Wansleben, Assistant Professor, LSE Sociology

 

Last week, a curious incidence occupied my time more than I would have wanted or expected. I received an email from a German sociologist, who addressed me with a strange request: He wanted me to conduct 15 interviews for him on some specific question; the data would then be sent to Germany to be evaluated there; my remuneration would be 1500 euros.

I roughly estimated the hourly pay (including searching and recruitment of interviewees) but it was soon obvious that this offer was not attractive. I thus simply wrote that “given these conditions, it would be difficult, if not impossible, to find anybody willing to do this”.

Perhaps it would have been wiser not to reply. For a few days after my answer, I received another email: The German sociologist acknowledged that, when making his request, he had had no idea about the costs of recruiting an interviewer in the UK; could I please help with better estimating a realistic fee.

At that point, I felt that the flaw in his whole thing justified some attention. So I answered again, saying that, “this is not just a question of money: For any academic, even a recent graduate, it would be unwise to spend time in collecting data for others instead of dedicating that same effort into his/her own projects. The reward for any cooperating participant in a project should also be some gain of knowledge, not just money.”

 Why money shouldn’t buy data

Indeed, while the researcher’s original plan to buy 15 interviews for 1500 euros seems, at best, naïve, there are many areas in the social sciences where buying data is common practice: For instance, Bloomberg, Stockmaster, and other financial data providers sell price information to researchers in financial economics; doing research on the Greek debt crisis, I once myself requested a couple of analyst reports from Thompson Reuters; I was simply unable to pay the price (the overall sample would have cost more than 20.000 US dollars, some single reports 3.500 USD). Google, Twitter, and the likes also increasingly use their big data as a valuable asset, to be sold to academics, who are craving for data sets; the Economist Intelligence Unit, the Financial Times and other publishing houses have entered the same market.

What is wrong about all this? Why should we not allow for more professionalization and delegate the cumbersome work of data collection to specialized interviewers and paid participant observers? One can make an interesting case against buying data by drawing on Michael Sandel’s What money can’t buy. In this book, Sandel argues that money can negatively influence or even destroy our relationships with other humans and things through the mechanisms of exploitation and corruption. Exploitation takes place whenever an exchange of money against some good or service, while legitimized as being based on the free choice of both parties, actually draws on and reinforces inequalities between them. Sandel’s most striking example is surrogate motherhood, the exchange of money against the ‘service’ of bearing a child. This exchange is based on a voluntary contract between those with a child wish and a surrogate mother (in jurisdictions, where such contracts are permitted); but for Sandel, this ‘business’ is exploitative: Agencies choose surrogate mothers, who do not have many alternative possibilities for monetary income; and they offer a chance for the wealthy to expand the range of ‘goods’ that they can access with their money. Translated to the context discussed here, exploitation would occur if academics from Western universities hired academics in developing countries or simply chose those with failed careers to conduct research for them. It is also evident that economists usually have more resources for buying data or hiring teams of ‘data crunchers’ than researchers from other social sciences. This arguably reflects inter-disciplinary inequalities, but also reinforces them.

Sandel’s second concern is that monetary purchases can corrupt our relationships with the purchased ‘goods’. Here, Sandel’s prime example is the wedding speech, purchased online, which makes us appear as elegant and humorous speakers, but fails to symbolise any actual familial or amicable relationship; Sandel also cites studies that show how monetary rewards destroy, rather than support our original motivations for altruistic acts (blood donations etc.). Again, we can use this idea to think about the possible effects of widespread data purchase: It would certainly change the substance of data if some kind of ‘service researchers’ primarily acted out of financial motives rather than an interest in the respective study. While less obvious, even more standardised data may be corrupted by commercialization: As long discussed among researchers, financial rewards for participants in studies create sample biases and impact the participants’ behaviour and responses. It also seems clear that Twitter data cannot serve as a representation of public discourse, but partly reflects the corporation’s intentions in the design of this platform. (It goes without saying that the argument cannot be turned around: Non-purchased is not necessarily unbiased and uncorrupted.)

Following Sandel, one must conclude that there is something wrong with buying data – more commercialisation of data acquisition will likely reinforce inequalities and corrupt the very motivational foundations of research.

The limited accessibility of data is a sociological phenomenon!

But there is, to my mind, an even more compelling argument against data purchase: Our world simply is no big repertoire of data. The problems of making this world empirically accessible are one of the most relevant aspects of any research; they should thus be taken as sociologically relevant and meaningful.

Indeed, we learn more from our problems of acquiring data than we usually recognise. For instance, as I later realised, my own difficulties in getting access to banks in the context of my PhD were essential experiences in better understanding what banks are about: As I came to see, banks’ natural resistance to being researched reveals a fundamental aspect of their possibilities of existence. If the banks’ balance sheets were public and understood, the financial system would probably implode. Moreover, were it plain how banks fabricated the interest rates that they charge, it would be hard to keep the fundamental conflict of interest with their debtors in check. Lastly, if wealth could not be turned into confidential and impersonal bank account data, banks would probably be unable to protect so much of it with so little effort. Thus, banks, like many other social actors, can only sustain themselves through obscurity.

Any sociological research requires a critical engagement with actors’ resistance of being researched – much sociological insight can emerge from that engagement. Therefore, my hunch is that, without coming to the UK and encountering the difficulties of conducting his interviews, my German colleague will simply fail to realise what his ‘UK case study’ can teach him. And if he purchases the interviews, he will end up with data, but without meaning.