Innovations in Digital Research Methods. Группа авторов
researchers the opportunity to use the data and test the generalizability of findings based on selected posts. Ideally, the data would be available to be used for social science research alongside and in combination with traditional data types. In social sciences, issues of transparency and open data are paramount.
News and media outlets now commonly refer to tweets to exemplify certain opinions or views on an issue. Though this falls a long way short of claiming inference to a population, this is not always made clear. There is a link here to good practices established for journalists when reporting social survey data and the need to include sample size numbers and response rates.75 As yet, such guidelines are not widely available in relation to the use of social media data.76 A key principle (which could be at risk in the trend towards data journalism and the pressure for instant analysis and commentary) is that reported results should fairly reflect the data collected.77 Implementing good practice regarding access and use of new sources of data would imply redesigning data management processes, including the development of reliable methods for coding highly unstructured data.78 Similarly, new research design and analysis methods will need to be developed and, where the research design requires it, researchers might draw on statistical analysis techniques for non-probability samples, including Bayesian analytical techniques.
Where the data about a population of interest are complete (Mayer-Schönberger and Cukier, 2013), then there is no need for methods based on sampling and inference. However, even in the future this is likely to be limited to certain areas of research. As such, it is important to link our thinking here to well-developed standards in social science research concerning: hypothesis testing, objectivity, metadata, archiving and replication; see Bryman (1998; 2013) for a good overview. To get robust and reliable use out of the new data we need to apply the same principles of rigour as we would to more traditional and established types of data sources: objective research design, data collection, analysis and reporting.
Despite our concerns here, it is clear that the new types of data provide a genuine opportunity for alternative narratives and perspectives to emerge in social science from the changing data environment, as evidence gaps are filled and new evidence is exploited.
2.5 Conclusions
In this chapter, we have highlighted the wide range of data – both orthodox forms and new data types – that are likely to be used, or considered for use, for social science research in the future. The new data types can track people’s daily lives in a more detailed and biographical way than ever before. The potential is that the data retains the strengths of more established forms of quantitative and qualitative data. The risk is that we are distracted by the scale and immediacy of the new data and may lose sight of the carefully constructed and tested rigour of traditional social science research methods.
Within each of the example policy areas we have explored, we note that similar patterns are emerging in the data sources: surveys, cohort studies, administrative data, commercial data, data deriving from new types of media, and trace data. We also note that different data types are now being combined and linked to address particular research questions. The potential of multiple data approaches presents an opportunity for social science research to tackle previously intractable social research questions and to facilitate a closer link to the policy making process by providing results that are grounded in real-world behaviour and delivered in almost real time.
The data identified across the example research areas cuts across all eight data types: orthodox intentional data, participative intentional data, consequential data, self-published data, social media data, trace data, found data and synthetic data. As such data becomes more accessible and the methods for exploiting the data mature, we would expect that selecting and combining data from different parts of the array will increasingly become a routine part of the research process and will transcend traditional divides such as those between qualitative and quantitative methodologies and primary and secondary research and data.79
As we have outlined, social science is moving from the idea of datasets to data streams and data arrays. Social science researchers may increasingly use near real time data systems as a tool and combine what, in the past, might have been seen as very different data types. This does not mean it is the end of theory, as has been debated (Anderson, 2007). Perhaps, more than ever, the testing of theories and hypotheses as a principal of social science research is paramount. Even with inductive techniques such as data mining, theory is still important. Without theory and hypothesis driven research, risks are posed by letting the data lead the research process.
The social and historical evolution of what has been termed the data environment has been, and will continue to be, characterized by a blurring of boundaries between data and subject, between researchers and researched, between research and its impact. In this new challenging context, there is a need to develop a new framework of ethics and good practice for accessing, analysing and archiving such data. This need cuts across society but the social science researchers should be at the forefront of developing and championing this new framework. In the next chapter we examine in more detail the methodological challenges and opportunities of using the new types of data and consider exemplar analyses.
2.6 Bibliography
Adkins, L. and Lury, C. (2011) ‘What is the empirical?’, European Journal of Social Theory, 12(1): 5–20.
Alasuutari, P., Brannen, J. and Bickman, L. (eds) (2008) Handbook of Social Research. London: Sage.
Anderson, C. (2007) ‘The end of theory’. Wired, Issue 7.
Association of Internet Researchers (AOIR) (2012) ‘Ethical Decision-Making and Internet Research: Version 2.0 – Recommendations from the Association of Internet Researchers Working Committee’. Available from http://aoir.org/reports/ethics2.pdf (accessed 06 Dec 2014).
Back, L. and Puwar, N. (2013) Live Methods. Oxford: Wiley-Blackwell.
Bates, L. (2014) Everyday Sexism. London: Simon and Schuster.
Berners-Lee, T. and Shadbolt, N. (2011) There’s Gold to be Mined from All our Data. University of Southampton. Available at http://eprints.soton.ac.uk/273090/1/Times%20OpEd%20TBL-NRS%20Final.pdf (accessed 6 Dec 2014).
Blasius, J. and Thiessen, V. (2012) Assessing the Quality of Survey Data. London: Sage.
Bowman-Grieve, L. and Conway, M. (2012) ‘Exploring the form and function of dissident Irish Republican online discourses’, Media, War and Conflict, 5(1): 71–85.
boyd, D. and Crawford, K. (2012) ‘Critical questions for big data’, Information, Communication and Society, 15(5): 662–79.
Boyle, P. (2012) Improving Access for Research and Policy, ESRC. Available from www.esrc.ac.uk/_images/ADT-Improving-Access-for-Research-and-Policy_tcm8-24462.pdf (accessed 6 Dec 2014).
Bryman, A. (1998) Quantity and Quality in Social Research. London: Routledge.
Bryman, A. (2013) Social Research Methods. Oxford: Oxford University Press.
Bucicovschi, O., Douglass, R.W., Meyer, D.A., Ram, M., Rideout, D. and Song, D. (2013) Analysing Social Divisions Using Cell Phone Data, in Proceedings of the 3rd International Conference on the Analysis of Mobile Phone Datasets (NetMob’13). Boston, MA. Abstract Available from www.orange.com/en/about/Group/our-features/2013/D4D/Folder/best-scientific