Extending the NILS: Making a Full Link to the 1991 Census
Lead Research Organisation:
Queen's University Belfast
Department Name: Sch of Education
Abstract
The Northern Ireland Longitudinal Study (NILS) is a powerful resource for the research and policy community to research and understand the social processes changing NI. Because of it history of intercommunal strife it is one of the most studied societies in the world and, more recently, has become an exemplar for conflict resolution in the transition from violence to peace. Consequently, there is great value in understanding pre-existing and on-going social, economic, and demographic processes in NI and this can only increase over time as more data sources become available. As noted, the NILS is ideally placed to fill this need: it is already established as a robust, wide-ranging and source; it is linked to the 2001 Census and to a variety of administrative sources; and is currently being extended to include linkage of the 2011 Census (planned for 2013).
However, in comparison, with the other UK longitudinal studies, it is weaker in data linkages backward through time. This is the rationale for the proposal for a link to the 1991 Census, the primary focus of this application. Such a linkage will allow analyses of the crucial period of rapid change in Northern Ireland (from 1991 to 2011), expecially of the discrete changes marked by social and demographic transitions measurable using census data (1991, 2001 & 2011). This may be of especial importance given the nature if the 'Peace Process' and the implementation and monitoring of policy interventions in fair employment and community relations. In addition, therefore, the proposal will also seek to scope the feasibility and value of a link to the 1981 Census to extend coverage further back through time so as to better understand Northern Ireland society during a very crucial period for its development. This is not a trivial problem - there may well be technical difficulties to overcome in terms of matching from the 1981 Census and methodological problems arising from the troubled history of the 1981 Northern Ireland Census which saw under-enumeration and selective response to some questions.
However, in comparison, with the other UK longitudinal studies, it is weaker in data linkages backward through time. This is the rationale for the proposal for a link to the 1991 Census, the primary focus of this application. Such a linkage will allow analyses of the crucial period of rapid change in Northern Ireland (from 1991 to 2011), expecially of the discrete changes marked by social and demographic transitions measurable using census data (1991, 2001 & 2011). This may be of especial importance given the nature if the 'Peace Process' and the implementation and monitoring of policy interventions in fair employment and community relations. In addition, therefore, the proposal will also seek to scope the feasibility and value of a link to the 1981 Census to extend coverage further back through time so as to better understand Northern Ireland society during a very crucial period for its development. This is not a trivial problem - there may well be technical difficulties to overcome in terms of matching from the 1981 Census and methodological problems arising from the troubled history of the 1981 Northern Ireland Census which saw under-enumeration and selective response to some questions.
Planned Impact
The full 1991 Link will be incorporated into existing NILS-RSU plans for impact and will help to expand and to strengthen these. There has always been a strong emphasis on policy implications of NILS research and outreach and dissemination strategies are a major part of any research proposal. The NILS Research Approvals Group (RAG) takes full account of these aspects when assessing all applications. This recognises the NI Executive as a source of NILS funding (they also support an RSU support officer) and this ensures an emphasis on policy and the NILS relationship with government.
Where possible, projects are planned with impact for policymakers and the public in mind, and examples include: suicide in NI; religion and health outcomes (both O'Reilly & Rosato); breast screening (Kinnear et al); fertility (McGregor & McKee); dental activity (Telford et al); entry to care-homes (McCann et al); antibiotic prescribing patterns (Johnston & Rosato); accuracy of address information from the healthcare system (Barr & Shuttleworth); and migration and residential segregation (Shuttleworth, Barr & Gould). Based on already published research the RSU routinely produces summary Research Briefs, helping maintain the relationship with government. The forthcoming challenge is to deepen this impact by increasing the range and quality of NILS projects and reaching out to new audiences. The 1991 Link will help this in two ways.
Firstly, plans for a programme of research around the 2001-2011 Census Link which are being developed will be extended backwards to 1991 to engage broadly with the theme of 'Northern Ireland in Transition' over the whole period. As part of this, it is planned (with support from funders) to engage with experienced and early-career researchers to produce a landmark edited book on 'NI in Transition'. This will generate impact, broaden the user base, provide a forum for young-researcher interchange, and focus current work with future research in a single reference volume.
Secondly, the 1991 Full Link will be used a basis to develop a programme of research with government and other researchers focussing on the transition out of conflict (in divided societies). This is ambitious and NILS involvement in such an enterprise is predicated on this link. The development of such ideas will involve cooperation with the Northern Ireland Equality Commission, the Community Relations Council, the Northern Ireland Housing Executive, and the Office of the First and Deputy First Ministers. It is planned that such a research programme could target UK and non-UK researchers with a track record of research in conflict studies, especially with a Northern Ireland focus, with approaches to foster research projects.
Finally, the 1991 Full Link will be used to increase the impact of the NILS by helping to make a case for further data linkage. The addition of data from the 1990s with its longer-term perspective might, for instance, help neighbourhood effects and exposure to environment to be better specified and thus help to make a stronger case for linking current administrative data to the NILS.
Where possible, projects are planned with impact for policymakers and the public in mind, and examples include: suicide in NI; religion and health outcomes (both O'Reilly & Rosato); breast screening (Kinnear et al); fertility (McGregor & McKee); dental activity (Telford et al); entry to care-homes (McCann et al); antibiotic prescribing patterns (Johnston & Rosato); accuracy of address information from the healthcare system (Barr & Shuttleworth); and migration and residential segregation (Shuttleworth, Barr & Gould). Based on already published research the RSU routinely produces summary Research Briefs, helping maintain the relationship with government. The forthcoming challenge is to deepen this impact by increasing the range and quality of NILS projects and reaching out to new audiences. The 1991 Link will help this in two ways.
Firstly, plans for a programme of research around the 2001-2011 Census Link which are being developed will be extended backwards to 1991 to engage broadly with the theme of 'Northern Ireland in Transition' over the whole period. As part of this, it is planned (with support from funders) to engage with experienced and early-career researchers to produce a landmark edited book on 'NI in Transition'. This will generate impact, broaden the user base, provide a forum for young-researcher interchange, and focus current work with future research in a single reference volume.
Secondly, the 1991 Full Link will be used a basis to develop a programme of research with government and other researchers focussing on the transition out of conflict (in divided societies). This is ambitious and NILS involvement in such an enterprise is predicated on this link. The development of such ideas will involve cooperation with the Northern Ireland Equality Commission, the Community Relations Council, the Northern Ireland Housing Executive, and the Office of the First and Deputy First Ministers. It is planned that such a research programme could target UK and non-UK researchers with a track record of research in conflict studies, especially with a Northern Ireland focus, with approaches to foster research projects.
Finally, the 1991 Full Link will be used to increase the impact of the NILS by helping to make a case for further data linkage. The addition of data from the 1990s with its longer-term perspective might, for instance, help neighbourhood effects and exposure to environment to be better specified and thus help to make a stronger case for linking current administrative data to the NILS.
Organisations
Publications
Brad Campbell
(2019)
Northern Ireland on the Move 1981-2015, completed PhD thesis
Ferry F
(2022)
Multimorbidity among persons aged 25-64 years: a population-based study of social determinants and all-cause mortality
in Journal of Public Health
Ferry F
(2018)
Social-determinants of self-reported multi-morbidity among adults in Northern Ireland: a census-based administrative data study
in International Journal of Population Data Science
Foley B
(2021)
How Does Administrative Data Compare with Census Data for the Measurement of Internal Migration? The Case of Northern Ireland, 2001-2011
in Applied Spatial Analysis and Policy
Northern Ireland Statistics And Research Agency
(2015)
Northern Ireland Longitudinal Study: Linking the Northern Ireland Longitudinal Study to the 1991 Census
| Description | Introduction The work reported in this document is not an orthodox research project but instead summarises the output of an ESRC-funded project to create linked Census data to enhance the research potential of the Northern Ireland Longitudinal Study (NILS). Its genesis lies in the successful bid to secure funding for the Northern Ireland Longitudinal Study Research Support Unit (NILS-RSU) from 2012 to 2017. In response to this, the Methods and Infrastructure Committee (MIC) of the ESRC suggested that a full link (rather than the then-existing partial link made using fuzzy matching of non-identifiable attributes such as postcode, gender and country of birth) to the 1991 Census should be made to extend the temporal depth of the NILS. The proposal offers three clear benefits. Firstly, it places the NILS on a more closely-comparable basis to the Scottish Longitudinal Study (SLS) which begins with 1991 Census data linkage and the ONS LS in England & Wales which commenced in 1971. Secondly, it makes cross-UK analysis using the LSs possible from a common start date of 1991. Thirdly, it covers the 1990s - a decade which saw the start of the transition away from violence in Northern Ireland - and is therefore a period of interest to social and political scientists as well as other scholars. Finally, it increases the power of the NILS since, when the 2011 Census data are linked, it will include the three Censuses of 1991, 2001 and 2011 rather than just the two of 2001 and 2011 (or the 'two and a half' given the partial nature of the 1991 Link before the project). The objectives of the project reflect this context. The overall intention was to create a full and robust linkage of the 1991 Census to the NILS that was comparable to the SLS and the ONS LS but beyond this the other objectives were to: 1. Provide full documentation on the 1991 Census and the linkage process for the benefit of users. 2. Promote the benefits of the 1991 Link to existing academic and policy users. 3. Use the 1991 Link to attract more users to the NILS. 4. Exploit the benefits from the 1991 Link to encourage linkage from other sources. 5. Use the full run of NILS data, 1991-2011, to promote research across the UK using the LSs for this time period. The next section of this key findings report reviews how far these objectives were met. Following this, the project's most significant achievements are discussed, before concluding comments on how the work may be taken forward and developed are made. To what extent were the objectives met? Nearly all the objectives were met to some extent, others were met in full, and yet others were exceeded. The brief account below justifies this statement. The success of the overall objective to create a full and robust linkage of the NILS to the 1991 Census comparable to the ONS LS and the SLS is documented in NILS Working Paper 5.0 (http://www.qub.ac.uk/research-centres/NILSResearchSupportUnit/GuidesResources/FurtherReading/). The headline statistics from this show that 78.5% of 2001 NILS members were identified in the 1991 Census compared with 75.6% in the SLS and 77.6% in the ONS LS, and that the NILS population with a 1991 Census link closely match the age and geographical distributions of the Northern Ireland population enumerated in the 1991 Census. The 1991 Link was completed by the planned date of September 2014 with 2001 occupation data being backcoded to SOC90. The specific objectives set out above will now be discussed in turn. That to provide full documentation on the 1991 Census and the linkage process for the benefit of users has been met with the development of the NILS Data Dictionary, (in coordination with the Census & Administrative data LongitudinaL Studies Hub (CALLS-Hub)), and the provision of Working Paper 5.0 by NISRA. A further objective was to promote the benefits of the 1991 Link to existing academic and policy users and this has been done at a variety of venues, most recently in July 2015 at the Understanding Society Conference held at the University of Essex. It was also hoped to use the 1991 Link to attract more users to the NILS. This has been met: to date, since the release of the 1991 Link, there are 5 current projects using the data, and other proposals which plan to use it. The objective to exploit the benefits from the 1991 Link to encourage linkage from other sources has been met partially through the additional linkage of registration data - see the next section on Significant Achievements. Finally, the objective to promote research across the UK using the LSs for this time period has not been realised yet in full given the time elapsed between the 1991 data release and the production of this report but it has been partially met by project proposals and ideas by postgraduate research students and others for cross-LS (and cross-UK) comparative analyses of topics like commuting, labour market transitions, and migration. The original objectives have, however, been substantially exceeded since within the project budget it has also been possible to complete a NILS link to the 1981 Census, link all deaths 1991-2001 to the 1991 Census to extend the Northern Ireland Mortality Study (NIMS), and to extend the linkage of birth registration data to the NILS from 1974-1991 - and thereby create a data series running from 1974 until the present. Significant achievements The significant achievements of the grant lie in the creation of additional Census and administrative datasets linked to the NILS. The first achievement is the production of a documented, robust and full linkage of the 1991 Census data to the NILS. Because of economies and technical improvements that were made by NISRA during the matching and linkage of the 1991 Census, there was a substantial underspend of the original budget. With the agreement of the ESRC it was possible to re-allocate this resource and use it for extra data linkage. The second major achievement was therefore to link the 1981 Census to the NILS. In effect, this has given two Censuses for the price of one and means that the NILS now extends from 1981 to 2011 over four Censuses and is second only to the ONS LS in its temporal scope. With the additional future linkage of the 2021 Census, the NILS will eventually cover five Censuses and this will greatly improve in its power to investigate events over the life course. The final significant achievement is the creation of a 1991 NIMS - a 100% linkage of all deaths in Northern Ireland 1991-2001 to the 1991 Census - and the linkage of birth registration data 1974-1991. These are important developments in studying mortality in three different decades (with the 2001 and 2011 NIMS) and also in extending the temporal reach of the NILS data. Using the births data from 1974, for instance, it will be possible to trace people with NILS birthdays from that year, pick them up in the 1981 Census, again in 1991, and then in 2001 and 2011, and to have background information on them that is collected via through the registration of vital events. It is therefore already possible to describe and measure the life events of NILS members from 1974 all the way to their current age of 40-41 years. This is highly important and if used imaginatively by researchers and the user community. Full documentation on the 1981 Link is being prepared by NISRA; NILS metadata have been created (and modified) to accommodate the 1981 Census Link, the 1991 NIMS and the historical births. Narrative impact It is too soon for the creation of these data resources to have measurable impact. In some ways, the impact is in their existence and the provision of a NILS that now runs from 1981 to 2011 and, with the addition of registration data, from 1974 to 2015. This is a unique resource that will be available only in the NILS as it is impossible to link the 1981 and 1991 Censuses to other data. This resource will benefit Northern Ireland-based researchers and NILS users but also stakeholders far beyond Northern Ireland. There is already measurable emergent impact in the 5 projects that use (and others that seek to use) 1991 Census data. We have developed plans to maximise the impact and uptake of the new data resources that were funded by the ESRC. These include an Impact Workshop to be held in Belfast in October 2015 with invitees from Northern Ireland and other parts of the UK and, in 2016, a day-long event to celebrate the first ten years of the NILS where the full range of data available in the NILS will be discussed. This is in addition to the normal promotional work undertaken by the RSU in academic conferences such as Understanding Society, BSPS, and at the Society for Longitudinal and Lifecourse Studies. |
| Exploitation Route | There are significant shortcomings with the 1981 Northern Ireland Census. They arise from the political context which saw the hunger strikes of Republican prisoners in the Spring of 1981 when the Census fieldwork was conducted. There was a campaign of non-cooperation with the Census (as an element of the British state) which led to non-response to the voluntary question on religion and, more seriously after the murder of a census enumerator, to under enumeration with missing households and people. It is assumed that these problems were concentrated in Irish nationalist and Republican areas. To develop the NILS further and to explore the value of the 1981 Census link it will therefore be necessary to estimate the size of these gaps and to use statistical methodologies to add missing data, whether on religion, or for missing people and households. The use of the NILS itself may help this process as will possibly missing data and population synthesis methods. We intend to explore these avenues. |
| Sectors | Communities and Social Services/Policy,Education,Healthcare |
| Description | It is too soon for the creation of these data resources to have measurable impact. In some ways, the impact is in their existence and the provision of a NILS that now runs from 1981 to 2011 and, with the addition of registration data, from 1974 to 2015. This is a unique resource that will be available only in the NILS as it is impossible to link the 1981 and 1991 Censuses to other data. This resource will benefit Northern Ireland-based researchers and NILS users but also stakeholders far beyond Northern Ireland. There is already measurable emergent impact in the 5 projects that use (and others that seek to use) 1991 Census data. We have developed plans to maximise the impact and uptake of the new data resources that were funded by the ESRC. These include an Impact Workshop to be held in Belfast in October 2015 with invitees from Northern Ireland and other parts of the UK and, in 2016, a day-long event to celebrate the first ten years of the NILS where the full range of data available in the NILS will be discussed. This is in addition to the normal promotional work undertaken by the RSU in academic conferences such as Understanding Society, BSPS, and at the Society for Longitudinal and Lifecourse Studies. |
| Impact Types | Societal |
| Description | AQMeN International Conference: Rediscovering inequalities: exploring the interconnections between crime, education and urban segregation |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | This post-graduate presentation made use of 1991 NI Census data linked to the NILS to discuss labour market inequalities |
| Year(s) Of Engagement Activity | 2016 |
| Description | Big Data and the Information Age |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Undergraduate students |
| Results and Impact | The NILS dataset and linkage template is used within a lecture for first year students within Digital Society module to demonstrate examples of data linkage. This lecture is hosted by one of the CI's (John Moriarty) Students then use the linkage diagram as a template to create their own proposal for a linked dataset using open data resources, including NINIS, Open Data NI and ARK electoral archive. This introduces them to many different datasets containing information relating to NI and raised awareness of the NILS amongst students. |
| Year(s) Of Engagement Activity | 2021 |
| Description | British Gerontology Conference at the University of Southampton (Presentation) |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | The NILS was also promoted with poster and paper presentations at the British Gerontology Conference at the University of Southampton from 1st-3rd July 2015. The aim was to promote the NILS to a new research constituency. |
| Year(s) Of Engagement Activity | 2015 |
| Description | ESRC Longitudinal Roadshow - Aberdeen |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Professional Practitioners |
| Results and Impact | The ESRC suggested a series of promotional roadshows following a successful event in Belfast. This longitudinal studies were therefore taken 'on the road' to Scotland to reach regional audiences that had not been reached by national events in the capitals of the constituent countries of the UK. |
| Year(s) Of Engagement Activity | 2015 |
| Description | How to write a successful NILS application |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Third sector organisations |
| Results and Impact | This was a webinar to discuss how to get started writing a NILS application and tips on how to ensure it's successful. The event attracted approx. 20 attendees, primarily academic but some government and council representatives. Many attendees had a research idea in mind but hadn't completed an application before or maybe was a long time since their previous application so were looking for a reminder. Made the application process more transparent for new and returning users, encouraged more NILS applications. |
| Year(s) Of Engagement Activity | 2020 |
| Description | LS Roadshow - Glasgow |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Professional Practitioners |
| Results and Impact | This was part of a series of roadshows to take the longitudinal studies 'on the road' to reach audiences that had not been reached by earlier promotional events. |
| Year(s) Of Engagement Activity | 2015 |
| Description | NI Science Festival - What Big Datasets Can Tell Us About Population Change & Health in Northern Ireland |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Public/other audiences |
| Results and Impact | This event is part of the NI Festival of Science. It was held on February 26th and its aim was to open the possibilities offered by administrative and linked data to wider audiences. |
| Year(s) Of Engagement Activity | 2016 |
| Description | NILS Impact Workshop |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | The NILS Impact Workshop 'Exploring Impact with the Northern Ireland Longitudinal Study (NILS)' was on 15th October 2015 with invited speakers from data providers (such as NISRA), funders (such as the ESRC and the HSC R&D) and academics and other researchers. Its aim was to engage critically with the concept of impact, bringing together researchers, policymakers, data providers, and research funders. |
| Year(s) Of Engagement Activity | 2015 |
| Description | NILS Promotion - BSPS |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Policymakers/politicians |
| Results and Impact | The NILS was promoted at the British Society for Population Studies (BSPS) conference in Leeds from 7th-9th September 2015. This was aimed to reach out to the membership of this organisation which encompasses academics and official statisticians. |
| Year(s) Of Engagement Activity | 2015 |
| Description | NILS Promotion - Understanding Society Conference |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | The NILS was promoted at the Understanding Society conference at the University of Essex from 21st-23rd July 2015. The purpose of participation was to engage with other ESRC longitudinal data resources. |
| Year(s) Of Engagement Activity | 2015 |
| Description | Policy Briefs for Researchers |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Study participants or study members |
| Results and Impact | NILS received ESRC Impact Acceleration Funding to host a workshop delivered by Bulletin, a strategic communications company. This was designed to support researchers in writing policy briefs from their NILS research findings. This was a small workshop designed to provide personalised feedback and support in writing policy briefs from research findings. The company delivering the workshop were very knowledgeable in translating research to knowledge and provided information and templates to write policy briefs during a dedicated writing session during the day event. Following this, the host provided personal feedback and tips to each person on the course. The event supported NILS users to maximise their research impact via the production of policy briefs. |
| Year(s) Of Engagement Activity | 2021 |
| Description | Presentation to British Society for Population Studies, Winchester, September 2016 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Policymakers/politicians |
| Results and Impact | This was a post-graduate talk which used 1991 NI Census data to discuss changing patterns of labour market inequality |
| Year(s) Of Engagement Activity | 2016 |
| Description | Section 75 - a signposting guide |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Policymakers/politicians |
| Results and Impact | The Equality Commission have produced a new guide which provides signposts to Section 75 information sources, to assist those working in this area to ensure that their policy making is evidence based and complies with their Section 75 requirements in this regard. NILS have contributed a section to this guide and the webinar was to launch and explain how to use the new guidance document. There were 450 people registered for the event with approx. 410 logging on to the live event. The event was also recorded and sent to all those who registered and will be made available for anyone interested. Specifically in relation to NILS, we were able to raise awareness of section 75 related information available within the NILS to users who may not be familiar with or even heard of the dataset before. The event raised NILS awareness amongst potential new users, showcased previous research conducted relating to section 75, raised awareness of the NILS as a potential source of information for policymakers. |
| Year(s) Of Engagement Activity | 2021 |
| Description | Society for Longitudinal and Life Course Studies (SLLS) conference (promotion) |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | The purpose of this presentation was to promote the NILS to a wider user constituency, and to engage with other longitudinal researchers. The conference was held in Dublin from 18th-21st October 2015 |
| Year(s) Of Engagement Activity | 2015 |