Current research programme
We run a programme of methodological research, both ‘in-house’ and through externally commissioned projects. The research is focused in strategically important areas, enabling social scientists to address key substantive and policy-relevant research questions.
Our in-house research programme aims to advance (i) methods for the analysis of complex data with linked and time-dependent structures, and (ii) understanding of the pedagogy of methodological learning.
Analysing hierarchical time-dependent paradata to reduce nonresponse in large-scale surveys (WP1)
We develop and apply techniques for the analysis of complex linked datasets with time-dependent and hierarchical properties. Analysis techniques are developed and applied in the substantive context of using survey paradata to model unit nonresponse. Specifically, we analyse call record and other field process data linked to 2011 Census records and to survey outcomes for the same households. These data are characterised by having measurements made across several time points and datasets, as well as possessing hierarchical structures through the nesting of individuals within households, interviewers and areas.
Data collection agencies have recently started to collect paradata routinely, as their potential for reducing survey error and improving fieldwork efficiency have become increasingly apparent. However, little is currently known about how to model these types of datasets in ways that provide substantive insight, while also adequately accounting for their complex structures.
The key methodological challenges are to properly incorporate information about the non-independence of observations, while also accounting for linkage errors between datasets. We employ a range of modelling approaches but focus primarily on sequence analysis and multi-level modelling.
Synchronic and diachronic scaling up in qualitative longitudinal data analysis (WP2)
The UK has been at the cutting edge of recent developments in qualitative longitudinal research (QLR) following substantial ESRC investment. The goal of QLR is to enable the exploration of complex processes, contexts and trajectories of social change and continuity in primary qualitative longitudinal data collection, and archived qualitative longitudinal data for secondary use.
We extend and develop secondary analytic practice in working with complex qualitative longitudinal data. We develop new procedures for working with multiple sets of in-depth temporal qualitative data to produce analyses that scale up horizontally and vertically, and that explore and compare processes of change and continuity.
Accounting for informative item nonresponse in biomarkers collected in longitudinal surveys (WP3)
A growing number of longitudinal surveys now incorporate the collection of biological data such as blood, saliva, and anthropometric measurements. While these types of bio-markers offer enormous potential for addressing important public health and social policy questions, many survey respondents decline to provide this kind of data for a variety of different reasons. This type of item non-response is likely to be ‘informative’, in the sense that measurements are missing for reasons that are related to the unobserved values.
For instance, obese individuals may decline a request to provide a weight measurement because they feel embarrassed about their weight. Existing techniques to take account of this type of informative missingness include propensity weighting, multiple imputation, selection modelling, and sensitivity analysis. However, it is difficult to evaluate the effectiveness of these approaches without some sort of external criterion against which they can be compared. External criterion data allow the possibility of linking data on missing (and nonmissing) respondents in the survey. Linking to administrative databases requires permission from respondents and therefore an additional (possibly correlated) source of informative missingness.
Utilising linked administrative data to account for missingness in social surveys with biomarkers poses a number of methodological challenges, including linkage error; analytical methods (how the extra information from linked administrative data should be incorporated into missing data models and weighting methods); and choice of simulation strategy to assess the performance of different approaches.
The anatomy of disclosure risk in a world of linked population data (WP4)
Data Stewardship Organisations (DSOs) are charged with producing useful data products whilst minimising the risk of disclosure of personal information. Work in this area has traditionally focused on the disclosure risk associated with census and survey outputs in the context of other information that an ‘intruder’ might use to identify persons and households.
We address how such disclosure risks must be reconsidered in the rapidly changing linked data environment, in which extensive geographically and temporally referenced data about persons, addresses and households can be combined in analytically powerful ways. Despite current excitement over the potential of data linkage, for example as a replacement for a traditional census from 2021, remarkably little is understood about the anatomy of this new universe of disclosure risks, particularly which sources and linkages present the greatest risks and how they might best be mitigated.
Addressing these questions involve a fusing of data environment analysis with traditional statistical disclosure control techniques. We investigate both new analytical methods for such data, particularly for understanding spatial associations, and the development of relevant privacy analytics.
The pedagogy of methodological learning (WP5)
The provision of courses in advanced social science research methods is unlikely, in itself, to be sufficient to ensure that capacity is developed in areas of strategic need. Researching the pedagogy of social science research methods learning is therefore key to providing an evidence base to inform a shift from merely providing ‘more training’, to developing capacity building strategies which will facilitate methodological innovation and skill development suited to complex social research problems.
We aim to: (i) advance an emerging pedagogical culture and content knowledge for social science research methods teaching; (ii) create a typology of pedagogical approaches for social research methods to inform policy and practice in NCRM and related investments; and (iii) develop a coherent theoretical framework for methods teaching to inform national practice.
Key questions include: How is the subject matter of advanced and innovative research methods best taught and learned? When, and how, does methods learning produce practical benefits? How can methods teachers’ methodological and pedagogical craft be most powerfully articulated?
Quantitative methods pedagogy (WP6)
The challenge of training future cohorts of quantitative social science researchers appropriately, and securing the ‘pipeline’ from school and undergraduate study through postgraduate to postdoctoral research, depends not only on resources but on effective pedagogy. While there is research evidence on statistics teaching, both at school and university levels, most of this literature is US based, and much of it prescriptive in its approach. There is little evidence-based work that addresses the social sciences specifically, where the challenge of poor maths skills and of a lack of confidence in applying them is particularly acute. It is also the case that other social science disciplines have not been studied as thoroughly as Sociology has.
In the context of the Q-Step programme, the evolution of the ESRC Doctoral Training Centres, and the on-going success of the AQMeN initiative to build quantitative methods capacity, the next phase of NCRM has a unique opportunity to work with quantitative methods trainers and students to learn more about ‘what works’ in pedagogical and career development terms.
We focus on learning modes and achievements, on what motivates students to understand the value of quantitative methods, what influences student recruitment and retention, and on what determines whether high performing UG students continue to postgraduate study using a quantitative approach.