5. Scottish Government's Plan for Data Zones
2011 Census Output Areas (COAs) will be aggregated up to existing Data Zone boundaries on a best-fit basis. These aggregates will then form the first cut of 2011 Data Zones. Where the population of the Data Zone is outwith the accepted population range, that Data Zone will be split into 2 or more Data Zones or merged with a neighbouring Data Zone in order to maintain roughly standard populations. These splits will be made along Census Output Area boundaries and, where possible, any merges will be made within Intermediate Zone boundaries.
Census Output Areas
To understand the implications of this, it helps to understand how Census Output Areas will be updated following the 2011 Census. 2011 postcodes will be allocated to 2001 COA boundaries on a best-fit basis. These aggregations will form the first cut of 2011 COA. The 2011 COAs will then be aligned to Local Authority (LA) boundaries, as some postcodes cross LA boundaries and some LA boundaries have changed since 2001. COAs outwith the existing population ranges will then be split into 2 or more COAs or merged with a neighbouring COA to ensure a roughly standard population across all COAs. Where a split or merge is required, it will be made along a locally recognised boundary. A hierarchy of boundaries will be used to determine where the new boundaries should fall where a COA is being split or merged. As Data Zones are built up from 2001 COAs, 2011 COAs should have a good level of comparability with 2001 Data Zones.
The main points are:
- 2011 Data Zones will be built up from 2011 Census Output Areas, which in turn will be built up from 2011 Postcodes;
- Data Zones will align to Local Authority boundaries;
- Where the population of a Data Zone has grown or shrunk significantly, it will be split into two or more areas, or merged with a neighbouring Data Zone;
- Data Zones will not be both split and merged (i.e. part of a Data Zone will not be cut-off and merged with another Data Zone);
- Where possible, merges will occur within Intermediate Zones' boundaries;
- Intermediate Zones will not be split or merged, however they will be realigned to Local Authority and Postcode boundaries;
The reasons for taking this approach are:
- Exact 2011 Population Census results will be available on the 2011 Data Zones geography under this methodology;
- The methodology is consistent with that used when Data Zones were first created;
- Data Zones will once again have roughly standard populations;
- Aggregations of Data Zones can be used to produce a comparable area over time;
- Minimal changes will be made to Intermediate Zones.
The drawbacks to taking this approach are:
- Discontinuities in some statistical outputs (e.g. SIMD);
- Postcode drift will cause small changes to the boundary, but not necessarily a discontinuity in many statistical outputs;
- Small discontinuities may be forced if Census Output Areas are merged across Data Zone boundaries;
- Small discontinuities in both Data Zones and Intermediate Zones where Local Authority boundaries have changed.
One of the main reasons for this course of action is strong user demand for exact 2011 Census data on the Data Zone geography. If the link between Census Output Area and Data Zones were broken in order to ensure that no changes are made to Data Zone boundaries, then Census data would only be available at Data Zone level on a best-fit basis. On balance, it is thought that the benefits of more accurate data outweighs the problems caused by the discontinuities that this approach brings.
Using 2011 Census Output Areas as building blocks also sets the timetable for producing 2011 Data Zones and means that Data Zones will be realigned to adjust for postcode drift and Local Authority boundary changes.
Respondents to the consultation were equally split between those who want to make no changes to Data Zones, so that they can make accurate comparisons over time, and those who want Data Zones to be substantially redrawn so that they can get more accurate information on the communities that they are interested in. It is impossible to meet the demands of both groups of users and splitting or merging Data Zones can be viewed as a compromise solution to the problem of population drift.
Higher Level Geographies
Data Zones will nest into Intermediate Zones, which in turn will nest into Local Authority boundaries. A number of respondents requested that Data Zones nest into Multi-Member Wards (MMW). MMW are due for review in 2014-2015 and the geography used in the 2017 local elections may be considerably different to the geography currently used. Data Zones would require substantial changes in order to align to MMW. Given the relatively short period between 2011 Data Zones becoming available and the review of the MMW geography, it is unlikely that the benefits of alignment for four years would outweigh the problems caused by the relatively large number of discontinuities that this would cause.
Lochs and Lakes
Lochs and other inland bodies of water that have a Mean High Water level will not be included within Data Zone boundaries. Several respondents raised concerns around how this might affect population density, housing density or other statistics derived from the Data Zone area, so this proposal will not be implemented.
Coding and Naming
2011 Data Zones will continue to use the S01 standard code, however all Data Zones will be given new codes starting at S01006506. The codes will be grouped together so that all Data Zones within a Local Authority will have consecutive codes. A look-up will be provided so that users can map 2001 Data Zones to 2011 Data Zones.
Local Authorities will be encouraged to name Data Zones. There is substantial user demand for Data Zones names, however the high level of local knowledge required to name small areas is not available centrally and there are sensitivities around naming that cause difficulties in some areas. Existing names will be carried forward and Local Authorities will be consulted on Data Zone naming where boundaries have changed.
Data Zones without names will be named after the Intermediate Zone that they fall within, using the naming convention IZ name - 01, IZ name - 02, etc.
Population Data Used to Base Changes on
There was very little support to using population projections as a basis for Data Zone changes. It would have been difficult to do this without breaking the link between COAs and Data Zones. In the event, there was substantial support for maintaining the link between COAs and Data Zones and for using 2011 Census data in the production of 2011 Data Zones.
Data Zones will continue to be reviewed once every 10 years. They will remain static in-between these reviews, however small changes may be made to Data Zone boundaries if Local Authority boundaries are changed.
Island Data Zones
A number of respondents identified problem Data Zones in the islands. In particular, respondents identified issues with Data Zones covering a number of islands or a Data Zone combining an island with part of the mainland. This is a particularly tricky issue, partially due to the small populations of a number of islands and the difficult postcode geography in some of these areas.
Giving every island its own Data Zone would result in a large number of Data Zones with an unacceptably small population. This would cause disclosure and data protection problems when publishing statistics on these areas, which would be ultimately counterproductive as the statistics would have to be withheld. The only way to get around this problem without heavy disclosure control is to group islands and occasionally parts of the mainland together until the overall population is within acceptable population ranges.
The postcode problem is more of a practical issue. A number of postcodes cross islands or are shared across an island and part of the mainland. Since records tend to be allocated to Data Zones based on the location of the postcode centroid, splitting postcodes will result in misallocated records. In this case, if the postcodes were split, a number of records based on one island would be wrongly allocated to a different island or to the mainland. In order to ensure that data is accurately allocated to Data Zones, it is beneficial for the Data Zone boundaries to align to the postcode boundaries.
Impact of these Changes