Census Background

**This information has been copied across from the SCROL website and some of it may be out-of-date. This will be updated in due course**

Census Background

Consultation

Potential users of the results from 2001 Census have been consulted since 1996, when plans for the first tests of procedures and questions were being planned. A White Paper 'The 2001 Census of Population' published in March 1999 set out the (pre-devolution) UK government's plans for the 2001 Census. Further major consultations involving roadshows and consultation papers took place in spring 1999 and autumn 2001 to identify user needs and to assist the design of the results to be produced from the collected data.

Legislation

The 1920 Census Act allows for the carrying out of a Census no sooner than five years after the previous Census. However, various other legislative requirements need to be fulfilled before a Census can be held. In the period up to publication of the White Paper, the Census Offices and users had been refining both the business cases for each Census topic and the wording of the question (or questions) to be asked in order to obtain the required information. Prior to every Census, a Census Order is required. This states the date of the Census, the people who are required to complete the form, those who are to be included on the form and the topics on which questions will be asked. The operational aspects of the Census also require legislative approval. This information is set out in Census Regulations, which contain details of how the Census is to be conducted. The Regulations include a copy of the Census form. The decision to include a question on religion in the 2001 Census, resulted in additional legislation, with the 1920 Census Act, Census Order 2000 and Census Regulations 2000 all requiring amendment to make provision for religion information to be asked.

Fieldwork

The Census was designed to collect information on the resident population on Census Day - 29 April 2001. Ahead of this day, enumerators delivered Census forms to every identified household space and communal establishment. Residents were asked to complete the forms with the information as correct on Census Day, and to return the completed forms by post. If a completed form was not received back, an enumerator visited the address in order to collect the form by hand or to issue a reminder to post back. Special arrangements were made to enumerate the Armed Forces and people sleeping rough. The overall response rate (that is, the proportion of people included on a returned Census form) is estimated as 96 per cent.

The Census was followed by the Census Coverage Survey (CCS), which took place between 24 May and 18 June 2001. This was an interview survey of some 36,000 households and 92 communal establishments in Scotland. Comparison of the results of the CCS with the Census was used to adjust the Census counts for under-enumeration (see the One Number Census). The CCS achieved a household response rate (the proportion of identified households, which were successfully interviewed) of 95 per cent.

Census content

The results are based on the information collected from the Census forms. Copies of the Census forms used are contained in the 2001 Census Definitions Volume or available on the Census pages of the NRS website (http://www.nrscotland.gov.uk). All questions included in the 1991 Census were included in the 2001 Census with the exception of questions relating to usual address and whereabouts on Census night (see Population Bases). However, the answer categories in some questions, such as ethnic group were updated. There were also new questions on general health, the provision of unpaid care, time since last paid employment, the size of work force at place of work, supervision of employees and two voluntary questions on religion. The Questions on travel were extended to cover travel to place of work or study instead of work only (as in 1991).

The Census questions asked of all people covered:

  • sex, age (date of birth) and marital status

  • relationship to others in the household (where applicable)

  • whether schoolchild/student

  • whether term-time address

  • country of birth

  • ethnic group

  • religion - current and upbringing

  • health

  • limiting long-term illness

  • provision of unpaid care

  • address one year ago

  • address of place of work or study

  • means of travel to work or study

  • knowledge of Gaelic

Whilst questions for those aged 16 to 74 covered:

  • qualifications

  • economic activity and employment status

  • number of employees at place of work

  • year since last employed

  • occupation and industry of employment

  • hours worked

In addition, the person filling in the form in each household was asked about:

  • type of accommodation and whether self-contained

  • number of rooms

  • availability of bath/shower and toilet

  • lowest floor level

  • presence of central heating

  • availability of cars or vans

  • tenure

  • landlord

  • whether or not accommodation was rented as furnished

Processing

Returned forms were fed through scanning machinery, which captured all the ticked responses, and stored most written answers in digital form. The latter were coded into categories either by automatic systems, which recognise terms given in response to questions, or by manual coding. This data was then edited to ensure that the data was consistent, and was followed by an imputation process to supply responses for questions, which had not been completed on the original form.

The Census was followed by the Census Coverage Survey (CCS), which took place between 24 May and 18 June 2001. This was an interview survey of some 36,000 households and 92 communal establishments in Scotland. Comparison of the results of the CCS with the Census was used to adjust the Census counts for under-enumeration (see the One Number Census). The CCS achieved a household response rate (the proportion of identified households, which were successfully interviewed) of 95 per cent.

Following the 1991 Census, it was acknowledged that the Census suffered from a degree of undercount, the extent and nature of which was not identified by the 1991 Census Validation Survey. As a result the detailed 1991 Census tables were not consistent with the final estimate of 1991 Census Day population. To avoid a similar situation following the 2001 Census, the Census Coverage Survey (CCS) and the One Number Census (ONC) projects were initiated.

The 2001 CCS was an intensive enumeration of a representative sample of postcodes in Scotland and was designed to be independent of the Census and provide the ONC Project with the required data to estimate under-enumeration. The ONC Project derives its name from the intention to eliminate different population counts so that all Census outputs add to one number - the national population estimate that has been adjusted for undercounting. The ONC Project ensured that the most appropriate estimation procedures were used to calculate the final Census data.

More detail on the conduct of the One Number Census is available on the Census pages of the GROS website (Census forms. Copies of the Census forms used are contained in the 2001 Census Definitions Volume or available on the Census pages of the NRS website (http://www.nrscotland.gov.uk/).

Comparison of Census coverage in the 1991 and 2001 Censuses Percentage
1991 2001
Enumerated in the Census 96.9 96.1
People identified as missing and included in Census results 1.2 3.9
Census coverage of the population 98.1 100
Missing but not included in Census results 1.9 0
Total 100 100

Material published before 27 February 2003 compared Census coverage between 2001 and 1991 using 1991 percentages based on the level of census underenumeration estimated at the time. However, evidence from the 2001 Census has indicated that the level of underenumeration in the 1991 Census was overestimated. Consequently the 1991 mid-year estimates and 1991 Census day counts have been revised (see NRS website (http://www.nrscotland.gov.uk) for more information).

As a result of these revisions, the proportion of population enumerated in the 1991 Census was greater than previously reported but remained fairly similar in 1991 and 2001 (97.3 in 1991 to 96.1 in 2001). As the table shows, the 2001 Census results cover the whole population compared with 98.6 per cent coverage in the 1991 Census. The difference in coverage lies in the approach to estimating under-enumeration in the two Censuses.

In 1991, the Census method allowed enumerators to identify many occupied households where the people within them had not returned a form and include an allowance for them in the Census tables. This gave an overall coverage of 98.1 per cent. The method did not enable enumerators to identify people missed from households where a form was returned, nor to make an allowance for any occupied households that they failed to locate. Using data from administrative sources and the 1991 Census Validation Survey it was estimated that these people accounted for about 1.9 per cent of the population. These were not added to the Census tables.

In the 2001 Census, the One Number Census methodology allowed for both types of missing people to be adjusted for in the Census tables.

Quality of the Results

The use of the One Number Census methodology means that the results of the 2001 Census cover the entire population of Scotland, and are believed to be the most reliable results obtained by any Census in Scotland. However, there are a number of sources of potential error in the results. These include

  • Incorrect or incomplete information provided on the forms.

  • Sampling error related to estimates derived through the One Number Census process.

  • Unidentified dependencies between the Census and the Census Coverage Survey.

  • Errors introduced during processing and imputation.

Some elements of incorrect information will have been corrected during the edit process (see Processing above). Other biases will have been corrected by the One Number Census process. Following this, the results have undergone an extensive quality assurance process, including checks against aggregated administrative information on particular groups such as students and the Armed Forces. Elements of dependency between the Census and Census Coverage Survey have been identified and corrected for, by cross-checking with alternative data sources.

Since the adjustments for under-enumeration through the One Number Census methodology are estimates based on a sample survey, sampling errors can be used as a guide in assessing the accuracy of the adjustments. The sampling error can be used to construct a 95 per cent confidence interval - that is a range in which we can be 95 per cent confident that the true value lies. For the population of Scotland, this confidence interval is ± 0.3 per cent of the estimated population.

Confidentiality

The Registrar General has a legal obligation not to reveal information collected in confidence in the Census about individual people and households. Protecting the Census data is of key importance and steps have been taken to safeguard confidentiality and protect against disclosure of personal information provided on the Census form. Further information on confidentiality is available here.