NCRM Autumn School 2017 - New Data Horizons
When: 22-24 November 2017
Where: Highfield House Hotel, 119 Highfield Lane, Highfield, Southampton, S017 1AQ
Fully-funded places for researchers. Applications from PhD students and early career researchers welcome.
In recent years the data landscape has seen significant changes and new developments, featuring innovations in data collection and usage of surveys, the rise of linked data sources, such as the linkage of biosocial data to surveys, and new forms of data such as big data and social media data. Knowledge from social scientists about how to properly use new forms of data to answer social science research questions and statistical tools to accommodate social science problems remains underdeveloped. Such data innovations raise questions on data quality and measurement errors. Yet new data will allow social scientists to examine fundamentally new research questions and creates the potential for substantial breakthroughs.
The ESRC National Centre for Research Methods (NCRM) will be holding its 12th residential autumn school in Southampton in November 2017. The topic of this year’s autumn school is New Data Horizons and will explore innovations and changes in the data landscape. The autumn school will feature sessions on innovations in surveys, including online survey data collection, data linkage and use of (linked) administrative and census data, social media data and big data, and biosocial data. It will address the potential for answering substantive questions in social science research as well as methodological aspects such as data linkage, modelling of complex data and issues concerning data quality.
Sessions will be presented by experts from academia, government and market research organisations including:
Patrick Sturgis (University of Southampton)
Joel Williams (Kantar Public)
Curtis Jessop (NatCen)
Jack Britton (Institute for Fiscal Studies)
James Doidge (University College London)
Dave Martin (University of Southampton)
Matthew Williams (Cardiff University)
Susan Banducci (University of Exeter)
Iulia Cioroianu (University of Exeter)
Michaela Benzeval (University of Essex)
Melinda Mills (University of Oxford)
Mike Daly (DWP)
Sessions will see a mixture of presentations from experts as well as interactive and discussion sessions.
The aim of the autumn school is to enable social scientists to learn about new, cutting-edge developments in both survey research as well as new forms of (linked) data. Key data sources will be discussed and examples from survey and data practice with applications to key social science questions will be given.
Who is it for?
Sessions will not assume expert knowledge, but some familiarity with the topic will be an advantage. Priority will be given to early career researchers from across the UK social science community who are employed on research-led contracts. But other researchers, including PhD students, are very welcome.
NCRM autumn schools are fully funded (travel, accommodation) so all reasonable costs of attendance for the successful applicants will be covered.
The autumn school will start on day 1 (Wednesday 22 November) with registration from 11:30 followed by lunch at 12:00. Presentations will start at 13.00. The autumn school will finish on day 3 (Friday 24 November) at ca. 13.00.
Please note that presentation details and session timings may still change.
Day 1 – Wednesday 22 November 2017
11.30 Registration opens
13.00 Introduction to the autumn school
Patrick Sturgis and Gabriele Durrant (NCRM, University of Southampton)
Session 1: Changing nature of and innovations in surveys
Patrick Sturgis (University of Southampton) Changing patterns of Social Science data usage
Abstract: There has been a great deal of excitement in recent years about the potential for new forms of data to transform empirical social science. Administrative, transactional, and other forms of ‘big data’ are, we are told, set to replace more conventional forms of ‘planned’ data collection such as censuses and surveys. To what extent are such narratives reflected in the actual practice of social science? In this talk, I shall draw on recent research which uses surveys of social scientists and content analysis of journal articles to assess the extent to which social scientists data usage has changed over the past fifty years, with a particular focus on the last decade.
Joel Williams (Kantar Public) Innovations in survey research
Abstract: Throughout this decade the commissioners of social science research have been under pressure to reduce expenditure on surveys. This has led to many studies of the value-for-money of each component of expenditure. In particular, lots of work has been done to test whether lower cost methods can replicate the results obtained from higher cost methods. Kantar Public’s contribution has been to develop ‘address based online surveying’ (ABOS) – a flexible random sample method that may be used as an alternative to both higher-cost interview methods and lower-cost (non-random sample) online panels. This session introduces the method and the many studies that have been carried out to assess how well it works.
Curtis Jessop (NatCen) Developing the NatCen panel
Abstract: The NatCen Panel is the first open probability-based research panel in Great Britain. It is designed produce robust estimates for the British population in a shorter time-frame and at a lower cost than the ‘traditional’ probability-based approaches currently available. This paper outlines the development of the NatCen Panel over the last two years, explaining the rationale for setting it up, and providing technical information on its methodology and key metrics on sample quality. It also summarises some initial findings from methodological experiments, and provides examples of how the Panel has been used over the last two years.
Evening Dinner (18.30)
Day 2 – Thursday 23 November 2017
Session 2: Data linkage, administrative and census data
James Doidge (University College London) Practical aspects and methodological challenges in research using linked data
Abstract: This session will start by covering some basic questions such as: What is linked data? Where does it come from? (where can you get it and how do you apply for it?) What is it good for? (examples of research questions that can be addressed using linked data). We will then consider some of the methodological challenges that are specific to research using administrative or linked data, focusing on the concepts of dynamic populations and linkage error.
Jack Britton (Institute for Fiscal Studies) Using administrative data to investigate graduate earnings and beyond
Abstract: In this session I will discuss my project with Anna Vignoles, Lorraine Dearden and Neil Shephard that linked administrative tax records to Student Loan Company records. I will discuss the challenges associated with the linkage, the pros and cons of the dataset and show results from our project focussing on earnings differences by subject, institution, socio-economic background and gender. I will highlight administrative datasets that will come available in the future and talk about some of the possibilities for future education research.
Dave Martin (University of Southampton) Transforming the Census
Abstract: The 2020-21 round of international censuses represent a major break with traditional practice, with a huge reorientation taking place towards internet data collection and the linkage of traditional census enumeration with administrative data sources.Census data provide key baseline sociodemographic and denominator data for a vast array of research and policy applications via a variety of different data products, and the new methods being adopted will have major implications for users. Taking the England and Wales Census Transformation programme as a model, this session will explore some of the major opportunities and risks for researchers.
Session 3: Social media and big data
Matthew Williams (Cardiff University) Towards an Ethical Framework for using Social Media Data in Social Research
Abstract: New and emerging forms of data, including posts harvested from social media sites such as Twitter, have become part of the social scientist's data diet. In particular, some researchers see an advantage in the perceived ‘public’ nature of social media posts, representing them in publications without seeking informed consent. While such practice may not be at odds with terms of service, I argue there is a need to interpret these through the lens of social science research methods that imply a more reflexive ethical approach than provided in ‘legal’ accounts of the permissible use of these data in research. To challenge some existing practice in social media research, this presentation brings to the fore: (1) views of Twitter users through analysis of online survey data; (2) the effect of context collapse and online disinhibition on the behaviours of users; and (3) the publication of identifiable sensitive classifications derived from algorithms. Examples are provided from research on online hate speech, funded under the ESRC’s New and Emerging Forms of Data Policy Demonstrator call.
Susan Banducci and Iulia Cioroianu (EXPONet, University of Exeter) Online Data Sources: Linking Old and New, Big and Small
Abstract: In this presentation we cover two areas: 1) tools to harvest online unstructured data (e.g. Twitter, comments, clickstream data) and 2) linking unstructured data to traditional data sources (e.g. surveys). The diversity of online media (mixed text/image/video content in heterogeneous formats) creates challenges in data collection and analysis. Using social media and online data also adds new challenges such as representativeness, measurement error and causal inference. We explore how to address some of these challenges by discussing tools to harvest data, data management and data linkage to traditional sources of social data. We illustrate some of these tools with examples from EXPONet --a project designed to develop tools toward understanding information exposure in dynamic and dependent social networks.
Evening Dinner (18.30)
Day 3 – Friday 24th November 2017
Session 4: Biosocial data
Michaela Benzeval (University of Essex) Integrating biosocial and social science data
Abstract: Understanding the interaction between people’s social and economic circumstances and their health across the life span is essential to develop policies not only to improve the nation’s health but also its social and economic capacities. To investigate these complex links, studies are required that bring together rich data on people’s social and economic lives with accurate measures of health and health functioning. Such data are available in Understanding Society: the UK Household Longitudinal Study, a major infrastructure investment for UK social science, which in recent years has been enhanced with the addition of biomarker, genetics and epigenetics information. This session will outline the biological information available in Understanding Society, some of the issues that need to be considered in analysing them and illustrate the ways in which they might add value to social science research.
Melinda Mills (University of Oxford) Combining Social Science and Molecular Genetic Research
Abstract: Within the last decade there has been an explosion of data that includes both social science and molecular genetic data. The UK is a frontrunner in collecting this type of data (e.g. ALSPAC, 1958 Birth Cohort, ELSA, Understanding Society), including recent large samples such as the UK Biobank. This talk provides an overview of recent social science applications in this areas of fertility, education and wellbeing. Key techniques and applications will be covered including using polygenic scores from large-scale GWAS (genome wide association studies) and GREML (genomic-relationship-matrix based restricted maximum likelihood) techniques. The talk will also reflect on the relevance and possibilities of using molecular genetic data for social science research.
Concluding discussion session
Finish: ca 13.00
Applying for a place
Applications should be made by email to Jacqui Thorp (email@example.com) by 12:00 Friday 6th October.
Please use the following headings in your application:
- Full name
- Email address
- Current Job Title/Role, Department/School/Faculty/Institution
- Year your PhD was awarded or PhD year (if currently doing a PhD)
- Title of your PhD
- Department/School/Faculty/Institution your PhD was based in
- Was your PhD ESRC funded?
- Please provide details of your research history, including current position and research (200 word limit)
- Please explain how the Autumn School will be of value in your research in the Social Sciences (200 word limit)
Once you have sent your application to Jacqui Thorp, the NCRM team will consider the applications and inform applicants of the outcome by 13th October. Any questions about the autumn school (including the application process) should be sent to Jacqui Thorp.
Previous Autumn Schools:
- NCRM, UKHLS and ICLS Autumn School: The use of biomarkers in social science research Limited places still available (2016)
- Early Career Researchers: Radical Interdisciplinarity in the social sciences (2015)
- International and comparative research (2013)
- Structural equation modelling - principles and practice (2012)
- Methods Crossing Borders - Journeys of methodological innovation and evolution (2011)
- Five years of qualitative innovation (2010)
- Hybrid and crossover methods (2009)
- Exploring new data sources in the social sciences (2008)
- Presentations from 2007 summer school:
- Challenges of conducting collaborative research across disciplinary and methodological boundaries (2006)
- Data Generation, Complexity and Synthesis (2005)