2011 UK Census

Please note - these initial findings were first presented on the pilot consultation website, and are reproduced here for reference.

Response to the survey

How many responded?

The questionnaire was completed by 529 people or groups. Of these

• 64 per cent were from individual census users
• 33 per cent were from single organisations
• 4 per cent were from multiple organisations.

Who do they represent?

view full size chart

The largest proportion of responses came from census users working for Local Authorities, but all other sectors of census users were represented

Are the people who responded experienced with using Census products?

How many had experience of using Census data?

• just over 95 per cent had experience of 2001 census products and services
• just under two-thirds had experience of 1991 census outputs
• just over one-third had used census data going back to at least 1981 outputs
• only 2 per cent had no experience of census data to date.

What 2001 Census products have they used?

Tables
The most frequently used 2001 census tables were

Value-added products
Over half the people used these, and the most popular were the area classifications and lookup tables.

The most commonly used products for analysing and/or presenting 2001 Census data were MS Excel and GIS packages (used by 81 per cent and 54 per cent respectively).

Metadata
The most commonly used (by 69 per cent) 2001 metadata product was the Definitions Volume.

About a third said they made use of other supporting data quality indicators such as overall response rates, question response rates, imputation rates and confidence intervals.

The majority of people thought the detail and technical level of the metadata supplied was about right.

What do people want from the 2011 Census?

How do they prioritise requirements for 2011 outputs?

The priorities for most people, in rank order, are for

1. timeliness of the outputs
2. accuracy, quality and reliability of the data
3. 2001/2011 comparative outputs
4. flexibility in outputs and geography
5. ease of access to the data

Types of census products - is commentary or analysis needed?

Around half think that it is important for the Census Offices to produce reports which provide analysis and commentary on census tables. The other think that just tables of counts will be sufficient to meet their needs.

What about 2001/2011 comparisons?

• the vast majority (83 per cent) will have a requirement to compare 2011 Census results with those from the 2001 Census.
• 38 per cent want to be able to compare results at all levels of geography including for the UK
• a majority (83 per cent) want comparisons at the County and LA (Local Authority) /UA (Unitary Authority)
• two-thirds need the flexibility to compare 2001/2011 results for their own bespoke areas of interest

How should data be accessible?

Of those who reponded, 72 per cent said that it was either important or very important to be able to access data from each Census Office from one access point (eg all England and Wales census data to be available via the same access point). 54 per cent also think it is either important or very important to have one access point to all 2011 Census data for the UK.

What media is important?

There is a strong demand (77 per cent) that 2011 Census output should be available either via web download or on physical media, in a machine readable format (such as Comma Separated Variable (CSV)) that can be easily imported into a users’ own systems.

2/3 of survey respondents would be content with electronic output only. Printed reports containing tables are still important to 1/3 and these are found particularly public and third sector groups of census users.

What about more flexible output?

The survey asked about different types of functionality that should be present and / or prioritised in a 2011 Census dissemination system.

For more flexible outputs, 69 per cent said that the facility to create their own user defined tables from given sets of variables (hypercubes) would be very useful. 54 per cent gave this a high priority in terms of functionality.

More than two-thirds would prefer to have a small set of pre-defined tables supported with a facility to make their own tables from data cubes. This was preferred having a large pre-defined set of tables equivalent in scope and quantity to 2001 census tables.

… and Metadata?

Tying up metadata with statistical output is important to the majority of respondents but not at the expense of any possible delay to publication of results.

… and data visualisation?

The facility to chart and map census output was not a high priority for a core Census dissemination system.

And the important issue of Statistical Disclosure Control (SDC)?

Most users fully understand the need for Statistical Disclosure Control to protect the confidentiality of 2011 Census outputs, however just under a quarter of people indicated they were not previously aware that SDC methodology was applied to Census outputs.

For the effects of SDC on the outputs, the most important issues for users were

• table additivity and consistency (i.e. that tables that add up internally to column and row totals and cell counts are consistent across different tables)
• that the data is as accurate as possible.

Accurate data is the most important feature. Only 25 per cent said they would be prepared to sacrifice some accuracy in order to have additive and consistent tables.

So what are the next steps?

The findings presented above are an initial overview. More detailed analysis is being carried out and further results and analysis will be posted as they become available. Although members of all census user communities responded to the survey, given that 44 per cent of respondents worked in local authorities, the requirements of these users will inevitably be dominant in the overall results. Results for the different user communities will be compared to identify similarities and differences in requirements and priorities.

User requirements emerging from the survey and other user engagement and consultation activities will feed into the development of the overall output strategy However it is apparent that meeting some emerging requirements would have significant resource implications. Also other factors such as potential conflicting user requirements across the different user communities, policy decisions and constraints, technical feasibility, disclosure control considerations as well as aforementioned resource constraints will also impact on the extent to which identified requirements may be fully met.

See other posts similar to this one: