.......... and we've a lot of interesting and thought provoking responses to consider over the coming weeks / months.

Full responses, analysis and conclusions will be published on the National Statistics website in due course.  If you responded to the consultation and provided your current e-mail address with your response, we will let you know when this material is available.

As I mentioned in my last post, there is a geography seminar being arranged at the Royal Statistical Society on the 3rd May.  I hope to see some of you at this seminar, although I will actually no longer be working for ONS.  The ONS view is likely to be presented by Alistair Calder and Pete Benton.

Thank you to everyone who has taken part in this blog and sent back their responses to the consultation.  If you haven't replied yet, this is the final week to let us know your thoughts.

Please do carry on adding to the debate on this site this week.  There are several discussion threads that I have not commented on myself.  You might want to add to these, or to comment on something new, such as:

• whether SOAs should be named;
• whether Business (or Workplace) OAs would be a useful development in city centres;
• the relative importance of social homogeneity, population size range, or area shape in the design of OAs;
• whether there are any clear and unambiguous benefits to be gained from redrawing OAs and SOAs;
• what you use OAs and SOAs for, and why that might mean it's important that they remain as stable as possible.

We are now starting to analyse your responses and will continue to do so over the next two to three months.  Analysis will be presented on the main National Statistics website.  Keith Dugmore and David Martin are also putting together a geography seminar for the 3rd May at the RSS.

A hybrid solution to the question of stability or redrawing has been offered by one or two people.  This hybrid is arrived at from a viewpoint that sees the merits of keeping Super Output Areas stable, but that is less sure about Output Areas.  I'd like to investigate this a bit.

I'll try and summarise the thinking behind the solution offered.  Apologies if I don't do it justice.  Corrections welcomed!

Apart from the 2001 Census, few data sources have been made available to ONS for publication at the OA level.  Outputs and data definitions for the next Census have yet to be defined.  It is possible that Census indicators of change at a level as small as OAs could be rendered invalid as a result of any changes to data definitions.  Given these points, OAs could be entirely redrawn within stable Lower Layer SOA (LSOA) boundaries and be based on Census 2011 data.  SOA time series data would therefore be safeguarded.

Please do take the time to complete the surveys that are on this site.  The consultation has a survey designed to understand your views on all the issues raised.  It is important that you complete as much of it as is relevant to you, as well as taking part in the debates around the blog.

The Hansard Society, who have put this site together for us, are also committed to evaluating online forms of deliberation.  As such, this blog is part of a larger study.  They have two surveys available for you to complete.

Their first survey is designed to find out what you think of online consultation and political blogs.  Their second survey is designed to understand how online consultations are working and how they could be better organised in the future.  Any thoughts or comments would be very much appreciated.

OAs were specifically designed to have a strong relationship with postcode geography.  Census 2001 enumerated postcodes were the building blocks used in the construction of OAs.  Only where postcodes straddled ward, parish (England), and community (Wales) boundaries, were they split into more than one OA polygon.  This affected approximately 2.5% of all enumerated unit postcodes.

The boundary set that resulted was irregular, but the postcode foundations of the geography were considered important for data linkage purposes.

As with administrative geographies, postal geography moves around over time.  It is a convenient, rather than ideal, geography for statistical purposes.  Ideally data records would be geo-referenced to addresses or grid references, and lookups from these would be easy to produce and distribute.  But further progress in this direction is still needed.  The desire has not yet proved strong enough to facilitate necessary data developments.

The consultation asks whether OA boundaries should be neatened to real world features.  This could be achieved by snapping them to, for example, OS MasterMap.

OA boundaries are largely abstract.  They were built up from artificial Thiessen polygons drawn around 2001 Census addresses.  As a result they cut across all other real world features, as do SOA boundaries further up the hierarchy.

Justin Martin picked up on this earlier in the blog and posted a comment.  He's keen to see a more coherent hierarchy of geographies down to individual features (or TOIDS).  There's clearly a potential data licensing issue here, but is this something worth investigating further?  What might the benefits be?

Thank you to everyone who has posted a comment so far on this blog.  I've been really pleased to see the thoughts of people I've not previously talked to, as well as those of some familiar faces. 

The response to the consultation survey has also been very encouraging.  Please make sure you take the time to complete it so that your views are included on all topics.

You may feel the need to escape the fun for a few minutes over Christmas and the New Year?!  If so, please do carry on posting your thoughts to me.  Although I'm not in the office, I will still be checking the site and will continue to publish your comments.

I want to concentrate this post on one of the topics in the survey: whether there should be an Upper Layer of SOAs, or a layer in any future geography that contains 25,000 - 40,000 people.

I've picked up on a potential area of confusion here from early survey responses.  Some of you have pointed out that you can create your own customised areas from the layers below.  This is true.  But you can only aggregate the data that's already available to you.  

The benefit of having an 'Upper Layer' is that it may enable the release of data that you can't currently get for areas smaller than local authority districts.  I'm thinking about Neighbourhood Statistics now, not the Census.  Data that is too low incidence, or potentially too disclosive to release for Lower or Middle Layer SOAs, or wards, or any other small area geography.  Datasets related to topics like illness and mortality, or ethnicity etc.

The key issue to be decided is whether we should aim for stability with OAs and SOAs through the next Census and beyond.  The consultation paper describes the issues surrounding this decision.  I know that there are many different views held.  Here are a selection of the comments made to me in the last couple of weeks, as I've been trailing the consultation:

• redraw the geography completely using the same algorithm but based on 2011 Census data and 2011 postcodes;
• keep SOAs, but redraw OAs within fixed SOA boundaries using 2011 Census data;
• the fundamental building blocks need to be real not abstract: redraw the whole lot and build it up from street blocks;
• we have to have ward, community, and parish level data too;
• now it's established you must keep it fixed, that's what it's for;
• as long as you've got grid referenced data you can give us any geography we want.  The disclosure risk from overlapping geographies is not so great.

To me, the message that has always come through most consistently, from members of all sectors of our user community, has been the stability one.  

This consultation is a first for ONS as we play our part in developing the concept of 'e-democracy'.  I am delighted that our consultation has been selected for inclusion and thank you for your interest in investigating this far.

The topic of local statistical geographies, for the Census, for Neighbourhood Statistics and for data analysis in general, is one that has generated significant interest and comment in the four years that I have been part of it.  Many times I have wished that I could get all of you with dramatically conflicting views together in the same room.  This blog gives us an opportunity to start to make that happen.  I hope that the following 12 weeks will allow us to see some real benefits in opening up the debate in this way to everyone's desktop.