Computation, digitisation and the source of social values

When the word ‘ethics’ is used in an e-research context, very often it is closely followed by the word ‘regulation’. This is understandable, since many of the capabilities of e-research technologies leave us perplexed as to whether to go on applying existing regulations or whether to make a change, and if so, which changes. But even if we did have a perfect set of regulations, perfect in formulation and in transparency so that it was always perfectly clear when and how to apply them, we would still not have exhausted the ethical dimensions of e- research, and at most, all we could be sure of is that e-research which complies with the regulations is not unethical-(on some presumably perfect conception thereof), and that researchers or their institutions cannot be sued for it.

But there are many more aspects to research that have ethical significance than those that can be regulated. In part, the guidelines for researchers in different disciplines try to address the gap between the regulations and the ethos of the discipline, it’s values and norms, and this is a very important area of debate that needs to engage with the institutional regulations and legal framework for research, but also with the technologies and media through which the research is undertaken and presented.

So, what more might we want from e-research that is not only not unethical but in some sense ethical? What do these terms mean anyway? One way of understanding them is in terms of the principles of least harm (non-maleficence) and benefit (beneficence) that have been so prominent in the medical model of research ethics. I will not attempt a detailed discussion of these principles, or whether they are appropriate as the basis for e-research ethics in the social science, and I mention them here as only one expression of a central way in which ethics is often understood: that is, as the capacity to benefit and harm each other, as individuals and as social groups. Without specific conceptions of benefit and harm the principles are purely formal, without any content or grip on actual situations.  Substantive conceptions of benefits and harms need to be fleshed out by specific desires, desires, wants, needs and fulfillments of human beings or social groups.  What we can desire, want or aim at is dependent on a wide variety of factors, including physiological, environmental and also, very importantly, social, economic and political factors. The particular form that desire for career progress can take, for example,  depends on gender, education, institution, and so on.  A clear example is the way in which women could not have career fulfillment as a mode of self-fulfillment until relatively recently: because of gender conceptions, women’s needs and desires were conceived of in ways that excluded this or made it difficult (with exceptions only proving the rule).  These social and self-conceptions are a very important source of what can count as desire or need or goal in terms of which any person or group can be benefited or harmed. They are sources of values which put meet on the bones of formal ethical requirements.

This is where social science can play a very important role, as it is by no means merely a neutral observer of the social field but an active player in it, and a powerful one at that, since it serves to bring the cachet of science and apparent objectivity to what might otherwise appear as mere subjective opinion. Social science both brings along with it specific conceptions of values in the social domain and contributes to forming those values.  In this way it contributes to the constitution of the sources of what counts as ethical practice and action, as benefit and harm. Social science by any means makes some sources of values easier to articulate, and makes others more marginal; and thereby makes more available some ways of benefiting and harming, and others more difficult.  It’s this domain, the source of values, that I hope to see coming to the fore in a Manifesto on Ethics in e-Research for social science.

The question facing social scientists using e-research methodologies and techniques is this: what difference is made to the sphere of social values by the use of computational and digital methods in social science?

Some examples:

A typical technology of e-research is the construction of huge data sets, very often consisting of quantitative data or — in virtue of the type of technology involved, data which is easily translatable into quantitative data — which can be computationally mined, processed, aggregated and disaggregated.  The UK government COIN initiative (Combined Online Information System) is one way in which social science will have new ways of using large data bases for studying issues relating to governance. However, we do know that we tend to measure – very carefully and with a great deal of apparent precision and objectivity – what we care about.  Consider all the different ways of measuring quality of life and which measures are used by whom and for what. A question for e-research ethics is: to what extent will the technologies of e-research enable a greater awareness of which values are beingacted upon, and make other values more visible, more apparent, and more available to be acted upon or enacted in the research? To what extent will the picture that we get of our own values depend on what happens to be computable?

A common method  in e-research in the natural sciences is agent-based modelling.  As the name implies this depends on the notion and understanding of agent. Agent based modeling is often seen as making methodological individualism more tractable as a basic social science methodology, since it makes it possible to make predictions centred on individuals as the unit of analysis.  It is also the case that understandings of agency in terms of rational self-interest lend themselves particularly well to the modelling and simulation approach.  If this way of doing research becomes as prevalent in the social sciences as it is in the natural sciences, what shifts in our self-understandings might ensue and how might this affect the sphere of values?

Computationally enabled network analysis also makes more available  social science research based on individuals, but this time understood as nodes in a network of relations.  The notion of social capital, measured by the number and types of relations that any one node has with others, has emerged from network analysis as a measure relating to individuals’ likelihood of success or failure in different areas, and therefore as something which can enter into a calculation of benefits and harms.

Privacy is a topic that has received a great deal of attention in the ethics debate, since the Internet has shifted the boundaries between public and private spaces. Social science by e-research means avails itself of these shifts to learn about Internet behaviour, but at the same time does not leave the field the way it found it, reinforcing the directionality of the shifts.

Thus there are ways in which we can conceive of ourselves which are extremely difficult or impossible without computational means of doing social science.  I believe that this is a matter of ethical as well epistemological significance, and needs to come to the fore in the discussion on the ethics of e-research.

The point is to understand the extent to which computability and digitisation act on that sphere of values, how they do so, and how, in turn, an ethically engaged social science can play its role in the development and use of these technologies.