This second collaboratory, to be held on 3 – 4 February 2014 in London, will bring together social scientists and national statisticians to explore the range of meanings and implications of Big Data for national statistics by attending to the question, ‘what counts?’ While Big Data sources have the potential to introduce efficiency gains in the production of official statistics (i.e. time and cost savings), they also raise a series of methodological and legal questions about data quality, management and protection. In this collaboratory we will focus on how these concerns require us to rethink some of the fundamentals of what constitutes an official statistic. While this question can be thought of in relation to computational or mathematical principles, we will inquire about how Big Data remakes the substance and meaning of what is measured, captured and counted and in turn what comes to count as an official statistic. To ask “what counts” also raises the question of who does the counting: what organisation or authority organises and mediates the making of ‘official’ statistics and with what consequences? Finally, we seek to examine how new data providers and sources require us to rethink the fundamentals of statistics such as the relevance of sampling theory and the rise of model-based approaches.
By drawing on existing experiments with Big Data on the part of national statistical institutes (NSIs) (e.g., mobile phone data, social media sentiment analysis, Google trends and so on) the collaboratory will address these questions in relation to a number of provocations. The common starting questions will include:
- what can and can’t be measured using Big Data sources?
- does the use of Big Data sources call for new forms of statistical reasoning and tests?
- what can’t be counted or measured using Big Data sources and why not?
- what does the use of Big Data sources for official statistics mean for the role of NSIs?
Participants will include the Socialising Big Data project team and statisticians from England, Estonia, Ireland, Italy, Netherlands, Eurostat and the UNECE.