ESRC Data Archive Bulletin:
The Northern Ireland Continuous Household Survey

The Continuous Household Survey (CHS) is a large probability sample survey of households in Northern Ireland conducted by the Policy Planning Research Unit (PPRU) of the Department of Finance and Personnel (NI). The major aim of the survey is to provide government departments with continuing information on topics such as population, housing, employment, education and health. The CHS is comparable to the General Household Survey (GHS) (Archive study number 33090) carried out by OPCS in Great Britain. The two surveys are similar in nature and aims and in the essential core of information collected. For the CHS in each year since 1983, some 6,500 adults aged 16 and over approximately 3,000 households have been questioned on a wide range of topics including employment, educational experience, health, family formation, welfare provision and housing conditions; with additional topics such as leisure, attitudes to the police and smoking and drinking habits being covered in some years. Yet, aside from government publications,[2] only restricted use has been made of this resource outside of government itself.

This report briefly reviews the content and structure of the CHS. The CHS is an analytical resource for Northern Ireland as valuable as the GHS (General Household Survey) (Archive study number 33090) is for Britain and can support a similar range of enquiries. Furthermore, in tandem the GHS, analyses of the United kingdom as a unit, rather than Britain or Northern Ireland alone, are possible.

The Content of the CHS

The CHS is a continuous survey in which interviewing is carried out throughout the year by government-trained staff in the respondents' homes. The source of the sample is the Rating Valuation List for Northern Ireland. In 1983 and 1984 a two stage sample design was used; electoral wards were selected and then addresses within wards were sampled. From 1985 onwards a simple random sample of 4,500 domestic properties on the list was selected. The survey has always been stratified by region within the province. In 1983 the regional strata were: Belfast, urban (non-Belfast), east and west of the province. After 1983 this was simplified to: Belfast (Belfast District Council only), east of the province and west of the province.

The interview schedule consists of two main parts: a household schedule and an individual schedule. The household schedule, common to all years of the CHS, deals with topics such as: personal details of all family members; accommodation; tenure; heating and fuel; household theft; and mobility. It is completed by either the household head or their spouse. All adults are then normally interviewed individually about their own occupation, education, health and other issues (only where an individual interview proves impossible is a proxy interview through another adult permitted). Because these latter questions are answered by the people themselves, rather than relying on an informant as the Labour Force Survey (Archive study number 33132) does. the responses are likely to have high reliability. The overall response rate averages about 73% of the effective sample. (For individuals within cooperating households, the non-proxy interview response rate has varied between 83% and 86%. If an individual cannot be interviewed directly after several retries, an limited amount of "proxy" information is collected from another household member). An innovation since April 1991 has been the direct collection of information through computer-assisted interviewing.

The content of the schedule changes somewhat from year to year, but always includes a constant core. The core topics are:

• Household composition, including the age, sex, relationship to head of household and economic activity of all those in the household;
• Household accommodation (eg type and age of building, number of rooms etc);
• Household tenure, cost of rent or mortgage, rates;
• Heating and fuel costs, including possession of a central heating system;
• Possession of a range of consumer durables (eg number of cars, possession of a refrigerator, television, freezer);
  

Individual Schedule

• General variables (eg age, sex, marital status, religion);
• Employment (coded using OPCS Classification of Occupation, Socio-economic group, Social class, Standard Industrial Classification Groups) and employment status. These data include details of second jobs, the last occupation of those currently not employed, hours worked and father's occupation);
• Education, age left school or college and educational qualifications, including CSE examinations and apprenticeships, and data on current part-time education;
• Income (detailed information on income was collected 198385, but this has produced problems of reliability; from 1986 on, more reliable but less detailed income band is available), national insurance and state benefits (all years);
• Family information on the number, sex and date of birth of children born to all married women along with anticipated fertility of all women under the age of 45;
• Health and use of health services, including chronic and acute sickness, consultations with Gps and hospital visits, children's health.
  
  

The Structure of the CHS

The fact that the CHS is a household survey is reflected in the structure of the data held by PPRU and is the source of one of its strengths - allowing analyses that are not necessarily individually-based - and one of the major technical difficulties. The data are organised on a year-by-year basis in SIR files with a "case" being a household containing one or more individuals. The result is that there is a variable amount of data for each household "case" with, furthermore, a varying amount of data for each individual in the household (eg on the number of visits a person made to the doctor). Moreover, individuals are also categorised into "family units", adding a further level of complexity. SIR provides sophisticated data management facilities and is now widely available to academics. Because it is a more powerful tool, however, it requires knowledge of data retrieval and is generally less straightforward to use than a package that assumes a 'flat" data structure.

Using the CHS Data

Some years ago, the 1983 CHS was documented for secondary analysts by an ESRC project (Grant H00 232082). The 1983 dataset was designed to be compatible with the GHS datasets produced by the University of Surrey (Gilbert. et al., 1981-6, 1988) (Archive study number 33124) and includes a number of additional "derived" variables such as ten- and seven-category approximations to the Nuffield "class" schema and the CASMIN educational scale (a "sociological" schema of educational categories developed for comparative mobility analysis by Muller et al., (See W. Konig, P. Luttinger and W. Muller. 1988). ESRC and PPRU are discussing the release of subsequent years of the CHS to the Archive for which the 1983 dataset will act as a model. It is anticipated that these will be available from the Archive in late 1992.

The scale and coverage of the CHS means that it may be used for a great variety of studies. A useful property of the survey is that the sample can be aggregated over several years to permit analyses to be carried out on sub-groups so rare that collecting quantitative data about them by normal means would be impossible or very costly. Clearly, the potential research uses of the data are legion.

References

Gilbert, G.N., Dale, A. and Arber, S. (1983) "The General Household Survey as a Source of Secondary Analysis", Sociology, 17(2).
Konig, W., Luttinger, P. and Muller, W. (1988) A Comparative Analysis of the Development and Structure of Educational Systems, CASMIN Working Paper, No.12, Institut fur Sozialwissenschaften, University of Mannheim.

Notes
[1] This description makes use of the "general notes on the survey" prepared by Jane Campbell, PPRU. Parallels in the GHS to some of the issues discussed here for the CHS can be found in G. Nigel Gilbert, Angela Dale and Sara Arber (1983).
[2] PPRU has published information from the CHS in the form of series of CHS Monitors and occasional papers on specific topics.