INCENSE - Influencing Policy and Practice by Stimulating Public Debate about Census Data

The UK census data will be released progressively from July 2012. There is an existing database of census responses from 1951 onwards. Census data are collected primarily to inform social policy - in particular to allow central and local government, health authorities (and others) to target resources effectively and to be proactive when planning housing, education, health and transport services. Here, we will focus on: the implications of demographic changes in ethnicity; religion; and health problems associated with an ageing population.

This proposal is a collaboration between the Data Visualisation Centre of the Office for National Statistics and the SMART Centre at Durham University. We propose to present census data in mash-ups - interactive displays that comprise multivariate data and commentaries (such as newspaper articles) using displays that have already been created by SMART and the ONS. A key feature of these displays is that users can explore complex data for themselves, and can form their own opinions about the commentaries that are offered, and can contribute insights, analyses and opinions to forums. We will promote the use of these interactive displays by journalists, politicians and citizens in a variety of ways - via the provision of free applications that can be embedded in websites, and circulated electronically, and via personal contacts, workshops, seminars, and press releases.

The research focuses on impact: the extent and nature of the use of these resources, and the consequences of this use. We will track resource use via applications such as Google analytics and hashtags on Twitter, and will analyse impact via counts (website hits, citations in policy documents and newspapers), and in terms of the interpretations made of the data (documenting conceptions, misconceptions, and the use of the same data set to support political arguments across the political spectrum). There is almost no research (outside our group) into ways that multivariate data can be understood by people who are statistically naive. Our research shows that naïve users at school and in colleges (aged 13-18 years) can draw sensible conclusions from multivariate data - and can articulate ideas such as interaction and effect size without formal training. This proposal will explore statistical reasoning by citizens and stakeholders in unconstrained environments.

We expect the research to have major political, theoretical and practical consequences. We anticipate having an impact on the political process by increasing the use of evidence in decision making, and by empowering some lobbyists (notably, the 'data driven journalism' movement). We will contribute to the statistical sophistication of journalists. We anticipate that many voters will be better informed, and better able to understand complex evidence (less directly, we hope to increase public perceptions of the value of social science). We anticipate theoretical insights into the processes of political influence; into the cognitive processes of drawing conclusions from complex evidence (both to understand, and to justify pre-existing beliefs). We anticipate practical consequences in terms of ways to communicate multivariate data better, lessons for some aspects of interface design, and in terms of developing well honed mechanisms to influence policy and practice.

It is almost certain that new media will emerge (c.f. Facebook) as will new ways of attempting to influence government policy (c.f. 38 degrees). We will respond flexibly. The direct involvement of ONS should ensure that research findings are useful beyond the life of the project; we will also work with Getstats, a major 10 year RSS initiative designed to improve the statistical literacy of citizens, and other groups which share the same mission.

There are a number of direct beneficiaries of the research outputs: policy makers and practitioners; offices of national statistics (and NGOs in general); the ONS itself; journalists; anyone concerned with the public use of evidence; politicians; and citizens.
It is difficult to overstate the importance of the work that is proposed here; increasing the accessibility and comprehensibility of census data, and the way it is used, should make a direct contribution to increasing the effectiveness of public services and policy. Better use of census data (especially the focus here on health and ageing) will enhance the quality of life, and health of many citizens. We expect the project to have an immediate impact that increases over the lifetime of the project, and one that continues well after it has finished.

Beneficiaries
A. Policy makers and practitioners will benefit from this work, because they will be able to interact with census data that is presented in far more detail than it has been in the past, and in ways that make it easy to access and interrogate.

B. National Statistics Offices (and NGOs such as eurostat, OECD and the World Bank) are directly concerned with influencing policy and practice. The next conference of the International Association for Official Statistics is entitled Getting Our Message Across-Strategies and Best Practice to Ensure the use of Statistics in Decision Making (Alan Smith will give an invited address). Professor Ridgway was a joint editor of the recent Statistical Journal of the IAOS issue devoted to statistical literacy). There is interest in the tools we have developed. Communication will be via conference presentations and via papers; JR is an associate editor of SJIAOS; JR and AS regularly present papers at OECD conferences.

C. The ONS will benefit directly; the research will stimulate interest in census data in ways that ONS is not allowed to do, and the study of media influences will be of long term use.

D. Journalists will benefit directly via access to materials, and the via a facility to incorporate displays directly into media. Research on reasoning with multivariate data will be useful in shaping communications (and will be useful for personal professional development). ONS are working directly with some influential journalists (e.g. Mark Easton at the BBC; Conrad Quilty-Harper (Telegraph)); we will extend this network.

E. A number of groups have an explicit brief to improve public understanding of data, and to improve the quality of presentation and debate in the media. The Royal Statistical Society has launched Getstats - a 10 year programme designed to increase statistical literacy - we have an agreement in principle to collaborate with Getstats. Other groups have similar ambitions, such as Straight Statistics and Sense about Science, and we will open discussions with them, once the project is underway.

F. There is a broad interest in stimulating public engagement with evidence (e.g. data.gov.uk and data.gov in the USA) which facilitate data access from a very large number of disparate databases. The research proposed here is directly relevant to this work.

G. Ultimately the public through better use of demographic data. The project is closely aligned with ESRC's aim to inform public policies and helps to make businesses, voluntary bodies and other organisations more effective through a more informed and critical use of census data.