Series details: CRDA/30

The Survey of Rural Services (RSS) was created as a survey of rural parishes and areas in England, to "provide information on the availability of a wide range of services to people living in rural communities . . . what access people living and working in rural areas have to public, private and voluntary services and whether the availability of these services is increasing or in decline".1 The RSS was conducted by the former Rural Development Commission (RDC) in 1991, 1994 and 1997. In 1999 it became the responsibility of the RDC's successor, the Countryside Agency. The Countryside Agency conducted the RSS in 2000, extending it to incorporate not only a survey of rural parish councils (as in previous sweeps), but also a geographical analysis of the distribution of rural services based on pre-existing datasets on the location of service outlets. For further information on how data was gathered in the various RSS sweeps, see How data was originally captured and validated.

The launch of the RSS in 1991 was intended to follow up recommendations in a report by D.M. Clark and S. Woollett, English Village Services in the Eighties, which was published by the RDC in 1990. The report concluded that a standard model survey should be established to allow data on rural communities to be gathered consistently and in a way which would contribute to a national picture of change. The first survey was preceded by extensive consultation in 1990-1991 between the RDC and a variety of outside bodies: Rural Community Councils, Action with Communities in Rural England (ACRE), the National Association of Local Councils, the Association of District Councils and the Association of County Councils. The RSS was initially known as the "Village Services Survey".2 "Survey of Rural Services" was adopted as the title of the report on the 1991 survey, and was subsequently used in 1994 and 1997. "Rural Services Survey" was used as the title of the 2000 Survey.

Data gathered in the various sweeps of the RSS was used to trace changes in the provision of rural services, and to inform government policy towards rural communities. The 1991 and 1994 surveys were a source of information for the White Paper Rural England (published in 1995), while the 1991-1997 surveys provided background data to the 2000 White Paper Our Countryside: The Future - A Fair Deal for Rural England. The latter placed a renewed emphasis on the RSS by setting out minimum standards and targets for the delivery of public services in rural areas, which would be measured against data gathered from future editions of the survey. Following publication of the results of the 2000 Survey, the Countryside Agency announced that it would use data from the RSS in its annual State of the Countryside report, which would report on progress towards meeting the government's targets. RSS data would also be used to produce elements of the Agency's regional and national indicators of rural disadvantage, and as one of the indicators measuring the UK's progress towards sustainable development.3

The results of the 1991, 1994 and 1997 surveys were published by the RDC, while the results of the 2000 survey were published in 2001-2002 by the Countryside Agency (see Publications produced by the originating department). Datasets from the first three surveys were distributed by the RDC to outside organisations, such as Rural Community Councils and parish councils. Downloadable data files for the 2000 survey were placed by the Countryside Agency on its website.4

The 1991, 1994 and 1997 Rural Services Surveys were conducted by the former Rural Development Commission, in association with Rural Community Councils and some county councils. RCCs were local charities supported by the RDC which provided advice and help to voluntary organisations and groups. They or the relevant county council sent out the survey questionnaires to the clerks of parish councils, and collected the completed forms for despatch to the RDC or to the research company. This procedure was not adopted in the 2000 survey, in which questionnaires were sent out directly to parish councils by one of the survey contractors, and were returned to the contractor.

The 1991, 1994, 1997 and 2000 surveys all employed outside contractors, listed below, acting on behalf of the RDC or its successor, the Countryside Agency:

• 1991: Research Surveys of Great Britain Limited
• 1994: BMRB International Limited
• 1997: Public Attitude Surveys Limited
• 2000: MVA Group, and the South Eastern Regional Research Laboratory (SERRL) of Birkbeck College, University of London.

The extent of the contractors' involvement varied from sweep to sweep of the RSS. In 1991, the research company was chiefly concerned with inputting the data, performing some validation checks and producing hard-copy and machine-readable tabulations. The design of the survey questionnaire, despatch and collection of the forms from RCCs and county councils, and authoring of the final report was undertaken by the RDC. In 1994 and 1997 the companies appear to have had a wider role which included not only data inputting and analysis but also authoring of the published report. It is also clear - at least for 1997 - that the research contractor played a major role in the design of the questionnaire, and was responsible for the distribution and collection of forms from RCCs and county councils.5 In the 2000 RSS, MVA Group dealt with the administration of the parish questionnaire component of the survey, including the design of the questionnaire (in conjunction with the Countryside Agency), despatch of the questionnaire to parish councils and the coding of responses. SERRL carried out all subsequent analysis, including the geographical analysis of data from external sources on the location of rural service outlets, which was the second component of the 2000 survey (see How data was originally captured and validated).6

For further information on the history of the Rural Development Commission and its successor, the Countryside Agency, see the Administrative History of the Countryside Agencies.

NDAD has received datasets for the 1991, 1994, 1997 and 2000 RSS surveys. There are key differences between the subject matter of the first three datasets and that of the dataset for the 2000 survey, resulting from the changes to the data gathering process which accompanied the 2000 RSS (see How data was originally captured and validated). The differences between the 1991-1997 datasets on the one hand, and the 2000 dataset on the other, fall into two main categories:

(1) The scope of the datasets. The 1991-1997 surveys were postal surveys of rural parish councils, in which parish clerks were asked to report on the availability of services in their parishes. The 2000 survey also included a parish survey component, along the lines of the previous surveys. However, SERRL (one of the research contractors: see Responsibility statement) also conducted a separate analysis of various datasets from external data suppliers, which recorded the location of certain types of services (e.g. cashpoints, schools) in relation to unit postcodes. This postcoded data could in turn could be grid referenced, allowing the data on the distribution of services to be analysed using GIS software. The analysis was used to produce tabulations about the geographical availability of services at various administrative levels, including ward and settlement level tabulations which form part of the dataset for the 2000 survey. The postal questionnaire data was also analysed using the GIS, and the resulting maps showing the distribution of services were included in the published survey results.7

The 2000 RSS dataset therefore comprises not only tables which record the data returned from the parish questionnaires, but also data on the geographical availability of services resulting from the analysis of external data sources. This second component of the 2000 survey is not present in the earlier datasets.

(2) The 1991-1997 datasets and the 2000 RSS dataset also differ in the granularity of the data and the way in which the data is grouped. The first three datasets consist of records for individual parishes, based on their returns, which are grouped into tables by county (see Logical structure and schema). By contrast, in the 2000 survey, the basic unit of data gathering for the parish survey component was the settlement - a unit within a parish (see How data was originally captured and validated). The results for the parish survey component are presented at settlement level (i.e. one record per settlement), and also aggregated to parish level (one record per parish). The data on the geographical availability of services arising from the GIS analysis is presented at settlement level and at ward level. Both the parish questionnaire data and the geographical analysis data are grouped into tables according to eight regions, rather than by county (see Logical structure and schema).

The remainder of this section looks at the subjects covered by the parish questionnaire component of the 1991-2000 surveys, and the subject matter of the tables in the 2000 dataset relating to geographical availability.

Parish survey (1991-2000)

The subject matter of the RSS remained broadly the same in the 1991, 1994 and 1997 sweeps, although there were changes to some of the questions and changes to the structure of the questionnaire. The main report on the 1997 survey includes a summary of the differences between the questions asked in 1997, and the questions asked in the two earlier sweeps of the RSS.8 The questions asked of parish clerks in 2000 were more focussed than those in previous surveys, as a result of an emphasis on gathering as much data as possible from existing data sources. Some areas which had been covered by the parish questionnaires in the previous surveys were not covered at all in 2000, including questions on banking facilities, child care groups, schools and colleges, adult education facilities, daycare groups, residential homes and sheltered housing, police and emergency services, and public utilities. Other topics which had featured in previous sweeps were dealt with in 2000 more narrowly. For example, in 1991-1997 parishes were asked to supply data on the numbers of doctors' practices, dental surgeries, clinics, etc in the parish. In 2000, by contrast, the questionnaire only asked for data on branch surgeries (i.e. arrangements for seeing medical staff away from the main surgery) in each settlement, as data on the location of main surgeries, health centres, dentists and pharmacies was already available to the Countryside Agency from other sources. At the same time, however, the 2000 questionnaire also covered some subjects which had not been dealt with previously, such as the availability of outlets selling locally produced food.

The following list summarises the topics covered in the data files for the parish survey component of the 1991-2000 datasets. As previously noted, in the 1991-1997 datasets this information is presented at parish level, whereas in 2000 it is presented in two forms (two tables per region): settlement level, and aggregated to parish level.

• Shops and other services: availability of post offices, garages and petrol stations (not 2000), shops, markets, banking facilities (not 2000), food and drink outlets (e.g. pubs, restaurants, take aways), and mobile shops and mobile services (e.g. milk delivery). The 2000 survey also asked about the sale of locally produced food by shops and delivery services.
• Childcare, education and other care facilities (not 2000): numbers (and numbers of enrolled pupils) of child groups (e.g. nurseries, pre-school playgroups); numbers (and numbers of enrolled pupils) of schools and colleges, and the age groups which they served; numbers of adult education classes; and numbers of daycare groups for the elderly or disabled.
• Medical facilities: availability of hospitals, GP surgeries, clinics, dental surgeries, opticians, pharmacies and prescription services, and residential homes and nursing homes. In the 2000 RSS, this area was limited to questions on the availability of doctors' branch surgeries, and of facilities for obtaining prescriptions away from main surgeries or chemists.
• Places of worship: numbers of churches, chapels and other places of worship. The 2000 survey also asked for data on the total size of congregations, numbers of resident ministers, numbers of places of worship open during the day, availability of Sunday schools, and non-religious uses of places of worship (e.g. for school assemblies).
• Information services: availability of libraries, newsletters, notice boards, Citizens Advice Bureaux (not 2000) and Job Centre facilities (not 2000). The 2000 survey also covered public Internet facilities, community legal services and counselling services
• Recreational facilities: numbers of meeting places (e.g. village halls, church halls, social clubs/Working Men's Clubs) and recreation areas (e.g. sports fields, swimming pools, children's play areas). The 2000 survey also asked about the frequency of sports events and community events within each settlement.
• Clubs and societies: availability of church groups, sports groups, women's groups, youth organisations, clubs for older people, etc.
• Transport: availability and frequency of rail services (not 2000), bus services and other transport services (e.g. dial-a-ride schemes).
• Crime prevention and emergency services: e.g. numbers of police stations, fire stations and ambulance stations, special constables and Neighbourhood Watch schemes. In 2000, this area was limited to questions on Community Safety Partnerships, dissemination of crime prevention advice, and crime prevention measures (e.g. CCTV).
• Public utilities: the proportion of households connected to main gas, water, sewerage and electricity, and data on public telephones and recycling facilities. In 2000, this area was limited to questions on the availability of recycling facilities provided by the parish and by voluntary groups.9

Copies of the parish survey questionnaires used in the 1991-2000 surveys are included in the main report for each survey: see Publications produced by the originating department.

Geographical availability of services (2000 survey)

In the 2000 survey, data supplied from various external suppliers (e.g. CACI Ltd, LINK Network plc) on the geographical availability of services was initially analysed to calculate the distance from each unit postcode to the nearest outlet for each service. This distance - expressed in size bands - was used to produce the national maps in the main report on the 2000 survey. However, data at postcode level does not form part of the 2000 dataset. Instead, the dataset contains the results of aggregating the unit postcode data to two higher levels: settlement and ward. The settlements covered by the geographical analysis are those appearing in the 1998 Rural Settlement Gazetteer - a smaller set than the population of settlements in the parish component (see How data was originally captured and validated). The wards used in the ward-level tables on geographical availability are taken from the 2000 Index of Multiple Deprivation, and cover both urban and rural wards.

The settlement files on geographical availability (referred to as 'settlement summaries') record the distance from the centroid of the settlement to the nearest service outlet, expressed in eleven 2 kilometre bands (0-2 km up to 11-20+ km). In the ward level files, the same bands are used to record the number of household postal delivery points (taken as a measure of the number of households) within each band, in relation to a particular type of service outlet: e.g. the number of the delivery points within 0-2 km of a cashpoint. Both the settlement and ward level tables on geographical availability measure proximity to the following 'service outlets': primary schools, secondary schools, Job Centres, doctors' surgeries, cashpoints, banks and building societies, petrol stations and supermarkets. In addition, the settlement level tables record proximity to post offices, Benefits Agency offices, and hospitals, while the ward level tables also record proximity to libraries. The ward-level tables held by NDAD appears to represent an update of the data issued in 2001 (see Dynamic or closed).10

While SERRL also used GIS software to plot the results obtained from the parish questionnaire component of the 2000 survey, GIS data does not form part of the 2000 dataset (although maps resulting from the GIS analysis are included in the regional reports).11 The ward and settlement tabulations on geographical availability are based on the analysis of the external datasets only.12

No further transfers of data are expected in this series as the Rural Services Survey was discontinued after the 2000 survey and has been replaced by the Rural Services Series.

The Survey of Rural Services datasets and related dataset documentation are public records under the Public Records Acts 1958 and 1967. The National Archives has assigned these datasets and documents to class D 16, except for the main reports on the surveys (Dataset Documentation Catalogue, references CRDA/30/DD/1) which have been assigned to class D 10.

The Survey of Rural Services datasets and related dataset documentation are open without restriction. Data is available for browsing on demand by users of NDAD and does not have to be booked in advance.

The copyright of the Survey of Rural Services datasets and related dataset documentation belongs to the Countryside Agency. Copies may be made for private study and research purposes only.

The Survey of Rural Services datasets are not subject to registration under the Data Protection Act.

The language of the materials is English.

Documents and publications relating to the Survey of Rural Services have been transferred to NDAD and can be consulted via the Dataset Documentation Catalogue.

Datasets for the 1991 and 1994 Surveys have also been deposited at the UK Data Archive, study number 3476.

Reports for the 1991, 1994 and 1997 surveys were published by the RDC in 1992, 1995 and 1998 respectively as part of the RDC's "Rural Information Series". The 1991 report was authored by the RDC itself whereas the 1994 and 1997 reports were written by the research companies involved with the survey (see Aim and purpose). The main national report on the 2000 survey was written by the Countryside Agency and was published in 2001. Reports setting out results for each of the eight regions of the 2000 survey were published in 2002, as was a technical report on the survey authored by SERRL. Copies of these reports have been transferred to NDAD and can be consulted via the Dataset Documentation Catalogue.

No information is available regarding the hardware which was originally used to input and analyse data from the 1991-2000 surveys. It is known that the version of the 1991 survey dataset which was distributed to data users (e.g. RCCs) was meant to be compatible with IBM-type PCs or Apricot type computers. In 2000 (when they were transferred to NDAD), the 1991-1997 datasets were held by the Countryside Agency on a single stand-alone PC in the Agency's research department.

No information is available regarding the operating systems which were used when the data from the 1991-2000 surveys was originally input and analysed. In 2000, Microsoft Windows NT was the operating system on the PC on which the 1991-1997 datasets were held at the Countryside Agency.

No information is available regarding the software which was originally used to input data from the Rural Services Survey in 1991, 1994 and 1997. It is known that the dataset for the 1991 Survey was distributed to data users outside the RDC as DataEase files, which were meant to be compatible with several versions of DataEase (a database system manufactured at that time by Sapphire Systems Ltd.: later Sapphire International). The 1997 dataset was distributed to external data users as comma-separated variables (csv) files. When they were transferred to NDAD in 2000, the 1991-1997 datasets were held by the Countryside Agency in Microsoft Excel 2000 (it is thought that the data was converted to Excel 2000 from earlier versions of Excel), and in Microsoft Access. 13

The data from the parish questionnaire component of the 2000 RSS was input into the SPSS statistical analysis package (version not known), which was used for subsequent analysis of the questionnaire data. ArcView GIS software was employed for the geospatial component of the survey. The published version of the dataset for the 2000 survey was presented on the Countryside Agency's website as a series of Microsoft Excel files, which form the basis of the data preserved by NDAD.14

NDAD has received a user manual relating to the dataset of 1991 RSS data which was sent to external data users (see the Dataset Documentation Catalogue, reference CRDA/30/DD/6/1). This outlines the functions which were available to users of the DataEase dataset, including the ability to generate three types of standard reports. External data users were allowed to modify their copy of the database and were encouraged to add data to it (e.g. from locally conducted surveys).

There are fundamental differences between the structure of the 1991-1997 datasets on the one hand, and the structure of the 2000 dataset on the other, which reflect the changes in the RSS that accompanied the 2000 survey (see Scope and content).

1991-1997 surveys

The datasets for the first three RSS surveys are divided into tables by county. Each table contains the parish-level data for a single county or former county (there are tables for Avon, Cleveland and Humberside in the 1997 dataset, although these counties were abolished in local government re-organisation in 1996; presumably they were retained by the Survey to allow comparison of the 1997 data with earlier years). Not all of the "counties" correspond to the area of a county council: e.g. Sussex and Yorkshire are represented in each of the datasets by a single table covering more than one county council. The tables are in the form of flat files with a single record per parish. Each dataset also contains a lookup table which translates the codes used for parishes in the data tables.

This structure corresponds to the way in which the datasets were transferred to NDAD, except for the fact that in the 1997 dataset, the data for each county had been divided into two files. The dataset was distributed to external data users in this way because single files per county were judged to be "too big for most analysis packages to handle in one go".15 Additional details of the structure of the individual datasets are given in the dataset catalogues: see Links to dataset catalogues.

2000 survey

The data for the 2000 survey is structured by region and according to the component of the survey to which it relates. The eight regions (East of England, East Midlands, North East, North West, South East, South West and Yorkshire and Humberside) correspond to Government Office Regions. For each region, the dataset has four tables: a 'full settlement' table containing the results of the parish survey component, at settlement level; a 'parish aggregation' table, presenting the results of the parish survey component aggregated to parish level; a 'settlement services summary' table; and a 'ward level services' table. The last two are based on the geospatial analysis of external data sources (the second component of the survey), and present results on the distribution of service outlets at settlement level and ward level. See Scope and content for further information.

The first three RSS datasets are closed, in the sense that once each dataset was created and validated, the data was preserved by the RDC and was not overwritten by data from subsequent sweeps of the RSS. However, this is true to only to a limited extent for the data from the 2000 survey. When the Excel files for the 2000 RSS were downloaded by NDAD from the Countryside Agency's website in November and December 2002, the files which presented the results of the parish questionnaire component aggregated to parish level were described as having been "enlarged as of 13/5/02 to include further variables".16 Similarly, the files on the geographical availability of services at ward level were described as "2001 updates". NDAD was unable to obtain versions of files that preceded these modifications/updates, and it is not known whether earlier versions had been kept by the Countryside Agency.

This section outlines how data was gathered in the parish survey component of the 1991-2000 surveys, and the geographical analysis component of the 2000 survey. Further information about the methodology used in each sweep of the RSS is provided in the main report on each survey, and in the technical report on the 2000 survey: see the Dataset Documentation Catalogue.

Parish survey (1991-2000)

In the first three RSS surveys, data was gathered from parishes in England which had been identified as having a population of under 10,000. For the purposes of the RSS, these parishes were defined as "rural". The selection of eligible parishes was done using data from the decennial Census. In 1991 data from the 1981 Census was used to develop a list of 9,852 parishes which were sent questionnaires. RCCs were meant to assist this process by identifying eligible parishes which had been created since 1991, and parishes whose populations now excluded them from the survey. By the time of the 1994 survey data from the 1991 Census was available, and was used to draw up a revised list of 9,902 parishes. This was refined in the course of the survey, on the basis of information from RCCs and parish clerks, to produce a reduced list of 9,823 eligible parishes. In 1997, Public Attitude Surveys Limited used the 1994 list as a starting point and refined it before the start of the survey by consulting with the Department of the Environment, local authorities and RCCs to produce a final list of 9,677 eligible parishes.17

The contractors involved in the 2000 survey found that they were unable to simply use the list of parishes which had been used in 1997. Instead, parishes were selected by starting with parishes identified in the Ordnance Survey Boundary Line 2000 data series; connecting these parishes to local authority areas identified in the 1991 Census; and then matching the 2000 list to the list of parishes used in 1997. Those parishes in the 2000 list which could not be matched were matched by hand using records of statutory changes to parishes between 1997 and 2000. A further aspect of the 2000 survey was the identification of "settlements" within parishes, as the settlement rather than the parish was the basic unit of data gathering in 2000 (see Scope and content). Settlements in the 2000 RSS fell into two categories:

• Urban Areas within the Ordnance Survey/Department for Transport, Local Government and the Regions definition of urban areas. Essentially, this meant contiguous built up areas that intersected with at least four enumeration districts in the 1991 Census.
• Smaller settlements identified with the aid of the 1998 Rural Settlement Gazetteer. The Gazetteer used postcode data to identify clusters of unit postcodes sharing a common locality name.

These sources were used to supply parish councils with a list of the settlements in their area for which information was sought in the 2000 RSS. Where an Urban Area only partly intersected the parish, councils were told to only provide data on the part of the settlement that fell within the parish. By contrast, they were asked to supply data on all settlements that were identified from the Rural Settlement Gazetteer, even if part of a settlement fell outside the parish. Although the questionnaire told parish clerks not to identify additional settlements, it appears that, in practice, they were allowed to do so, and that these settlements are included in the dataset.18 The 2000 survey adopted an upper limit of 12 settlements per parish.19

In all of the 1991-2000 sweeps of the RSS, data was gathered from parish councils via postal questionnaires. In 1991 the questionnaire was developed by the RDC's Research Branch in consultation with RCCs and county councils, and was piloted in four counties. The 1994 survey used a modified version of the 1991 questionnaire, developed in light of the findings of the first survey and comments during the survey. In 1997 the RDC drafted the survey questionnaire, which was refined using a "readers' panel" of eight rural services experts, and then piloted in 20 parishes in Buckinghamshire and Northamptonshire. For the 2000 RSS, the questionnaire was developed by an Advisory Panel of Countryside Agency staff and external representatives of other government departments and community groups, and was refined in consultation with county councils, RCCs and contractor responsible for the postal survey (MVA Group). It was piloted in 30 parishes in Dorset, Herefordshire and Berkshire.

Different methods were used in the 1991-1997 surveys on the one hand, and the 2000 survey on the other, for distributing the questionnaire. In the first three sweeps the finalised questionnaire and accompanying material (e.g. covering letters) were distributed to parish councils by RCCs and by a small number of county councils. These bodies were also responsible for collecting and checking the completed questionnaires, and sending them to the RDC or to the research company. In 1991 forms were sent to the RDC which forwarded them to the research contractor; it is thought that a similar procedure was followed in 1994; in 1997 they were sent directly to Public Attitude Surveys Limited. In 2000, all distribution of the questionnaires was handled by MVA Group, to which completed questionnaires were also returned by parish councils. In addition to posting out the questionnaire in paper form, MVA also placed the questionnaire on its website, and allowed councils to submit their return electronically by email if they chose to do.

In each of the 1991-2000 surveys, inputting and validation was performed by the research contractors, which also undertook the post-inputting tabulation and analysis of the data. In 1994, this not only involved analysis of the 1994 data, but also re-analysis of the 1991 data using information from the 1991 Census. Both the 1994 and 1997 surveys also saw comparative analyses of the data from the latest and the earlier sweeps to trace changes over time. In the 2000 RSS, MVA Group coded parish questionnaire data, but subsequent analysis of the data was apparently undertaken by SERRL (see Statement of responsibility). This including aggregating the settlement level data to parish level (to provide parish as well as settlement files), and adding population estimates to both the settlement and parish files.20

In 1991-1997 the questionnaires were sent out to parish clerks in April, and were meant to be returned to RCCs/county councils by 30 April 1991 (the 1991 survey), 20 June 1994 (the 1994 survey) and 30 June 1997 (the 1997 survey). The first RSS had a "census" date, in that parish clerks were asked to supply information which reflected the situation on 21 April 1991. The timing of the 1991 survey was intended to facilitate future comparisons with data from the 1991 Census of population. This approach was abandoned in 1994 and 1997 (in 1997, clerks were asked to ensure that information was correct on the date on which they completed the questionnaire; answers relating to the provision of scheduled public transport services were meant to reflect the situation in the week commencing 7 April 1997). Another area where the 1991 survey differed from the 1994-1997 sweeps lay in the fact that RCCs and county councils were given the option of surveying by village rather than by parish. Those who chose to do so were instructed to aggregate the village returns to parish level and to send only the aggregated parish returns to the RDC. In the 2000 RSS, parish councils were told to return their completed questionnaires by 8 December 2000, though in the event returns received up to 17 January 2001 were included in the database. The 2000 survey had a response rate of about 68 percent (6149 parishes), which was lower than the response rate of around 80 percent which had been achieved in the three earlier surveys (7893 parishes in 1991, 7822 in 1994 and 7789 in 1997).21

Geographical availability of services (2000 survey)

As previously noted (see Scope and content), the 2000 RSS included an element of geospatial analysis using GIS software which had not been present in earlier sweeps. This analysis was carried out by SERRL, and involved two components:

(1) The analysis of the data gathered through the postal survey questionnaire, to plot the availability of services. This was done for those settlements which were identified from the Rural Settlement Gazetteer, each of which had a grid reference which was its population-weighted centroid. Maps showing the results of this analysis were included in the regional reports (see the Dataset Documentation Catalogue, reference CRDA/30/DD/8).

(2) The analysis of datasets on the distribution of service outlets which were obtained from a number of external data suppliers. These suppliers included other government departments (the Department for Education and Skills, Benefits Agency, Department of Health and the Employment Service), and commercial sources such as the LINK Network plc and CACI Ltd. The datasets covered the location of schools, post offices, Benefits Agency offices, Job Centres, doctors' surgeries, cash machines, banks and building societies, hospitals, petrol stations and supermarkets. The datasets included the unit postcode of each service outlet. Where grid reference data was not included, unit postcodes were assigned grid references by SERRL using the Postcode Address File. SERRL was then able to calculate the straight line distance from each unit postcode to the nearest outlet for each service, using eleven 2 kilometre bands (0-2 km up to 11-20+ km). This data was analysed for the purposes of the 2000 survey at various administrative levels, including regions, counties, and wards (using wards defined in the 2000 Index of Multiple Deprivation). The postcoded data was linked to settlements in the Rural Settlement Gazetteer, by calculating the distance from each settlement's centre to the nearest unit postcode for each service outlet. As previously noted (see Scope and content), ward and settlement level tabulations on geographical availability arising from this analysis are included in the 2000 dataset.22

One of the reasons advanced by the Countryside Agency for its decision to supplement the parish postal survey with the analysis of external datasets was that the latter were updated at regular intervals (annually or more frequently), which would permit frequent re-analysis of the data. Thus the ward-level files on geographical availability which form part of dataset for the 2000 RSS in fact represent updates which were issued in 2001, based on 2001/2002 data from external data suppliers (see the Dataset Catalogue for the 2000 dataset for further information: Links to dataset catalogues). In 2003 the Countryside Agency announced that it would be carrying out a pilot study in the South West of England to determine whether it would be feasible, in future sweeps of the RSS, to replace the parish postal survey with an online survey of county councils and unitary authorities. This would focus on gathering the limited amount of information on rural services which could not be obtained by analysing external data sources.23

Details of the content and transformation validation checks performed by NDAD staff on each Survey of Rural Services dataset are contained in the catalogues of individual datasets: see Links to dataset catalogues.

Dataset catalogues provide more detailed information about individual datasets, and are currently available for the following dataset(s):

1. Dataset Documentation Catalogue, reference CRDA/30/DD/1/3, p. 1.

2. Dataset Documentation Catalogue, reference CRDA/30/DD/6/1.

3. Dataset Documentation Catalogue, reference CRDA/30/DD/1/3, p. 1; CRDA/30/DD/1/4, pp. 4-5, 36.

4. Dataset Documentation Catalogue, reference CRDA/30/DD/6/1; note of telephone conversation between NDAD and the Countryside Agency on 22 June 2000; extract from the "Rural97 Survey Manual" supplied to NDAD as non-archival documentation.

5. Web site of the Countryside Agency, page entitled "Help for RCCs and ACRE" (http://www.countryside.gov.uk/what/helprcc.htm) consulted on 4 July 2000; Dataset Documentation Catalogue, reference CRDA/30/DD/1/1, p. iv; CRDA/30/DD/1/2, "Acknowledgements"; CRDA/30/DD/1/3, pp. 2, 173-177; CRDA/30/DD/3/1/1; CRDA/30/DD/6/1; note of telephone conversation between NDAD and the Countryside Agency on 22 June 2000.

6. Dataset Documentation Catalogue, reference CRDA/30/DD/9/1, pp. 3, 6-7.

7. Dataset documentation catalogue, reference CRDA/30/DD/9/1, pp. 8-10.

8. Dataset Documentation Catalogue, reference CRDA/30/DD/1/3, p. 172

9. Dataset Documentation Catalogue, reference CRDA/30/DD/1/1, pp. 93-102; CRDA/30/DD/1/2, pp. 193-218; CRDA/30/DD/1/3, pp. 180-195; CRDA/30/DD/1/4, pp. 38-41 and CRDA/30/DD/4/2.

10. Dataset documentation catalogue, reference CRDA/30/DD/9/1, pp. 8-14.

11. Dataset documentation catalogue, reference CRDA/30/DD/9/1, p. 10; CRDA/30/DD/8/1/1, pp. 3-4.

12. This is indicated by the sources listed at the start of each of the original Excel files, in the 2000 dataset. See 'How data was originally captured and validated' in the Dataset Catalogue for the 2000 dataset for further details (Links to dataset catalogues).

13. Dataset Documentation Catalogue, reference CRDA/30/DD/6/1; note of telephone conversation between NDAD and the Countryside Agency on 22 June 2000; extract from the "Rural97 Survey Manual" supplied to NDAD as non-archival documentation.

14. Dataset Documentation Catalogue, reference CRDA/30/DD/1/4, p. 5; CRDA/30/DD/9/1, pp. 5, 8; note of telephone conversation between NDAD and the Countryside Agency on 22 November 2002.

15. Extract from the "Rural97 Survey Manual" supplied to NDAD as non-archival documentation.

16. Countryside Agency website (http://www.countryside.gov.uk/ruralservices/downloads.htm) consulted on 21 November 2002.

17. Dataset Documentation Catalogue, reference CRDA/30/DD/1/1, p. 1; CRDA/30/DD/1/2, pp. 3, 179; CRDA/30/DD/1/3, pp. 173-175; CRDA/30/DD/3/1/1.

18. The 2000 technical report (Dataset Documentation Catalogue, reference CRDA/30/DD/1/4, p. 8) states that if a parish clerk entered a settlement name which was not recorded in the Rural Settlement Gazetteer, no population estimate for that settlement was added to the data file. The 'full settlement' files in the 2000 dataset contain records for a number of settlements where the DELIVERY field (estimated resident population) is empty, but the settlement is not identified as an urban area (population estimates were also not added for urban areas).

19. Dataset Documentation Catalogue, reference CRDA/30/DD/1/4, p. 38; CRDA/30/DD/9/1, pp. 5-6.

20. Dataset Documentation Catalogue, reference CRDA/30/DD/1/1, p. 90; CRDA/30/DD/1/2, p. 3; CRDA/30/DD/1/3, pp. 1-3, 173-176; CRDA/30/DD/3/1/1; CRDA/30/DD/6/1; CRDA/30/DD/9/1, pp. 6-8; note of telephone conversation between NDAD and the Countryside Agency on 22 June 2000.

21. Dataset Documentation Catalogue, reference CRDA/30/DD/1/1, pp. v, 5, 93, 101; CRDA/30/DD/1/2, pp. 3, 193; CRDA/30/DD/1/3, pp. 3, 181, 195; CRDA/30/DD/3/1/1; CRDA/30/DD/9/1, pp. 7-8.

22. Dataset Documentation Catalogue, reference CRDA/30/DD/1/4, pp. 5-6; CRDA/30/DD/9/1, p. 9-14; CRDA/30/DD/8/1/1, pp. 3-4.

23. Dataset Documentation Catalogue, reference CRDA/30/DD/8/1/1, p. 4; CRDA/30/DD/10/2.