Last November I highlighted a report from the Commons political and constitutional reform committee which contained some pretty negative views about the use of 'data matching' to improve the electoral register.
"The evidence we have received... suggests that data matching will be of limited effectiveness, especially in identifying potential electors."
So it was slightly surprising to read the government's latest electoral registration proposals which suggest that checking the register against other official databases will "confirm the accuracy of the majority of entries on the electoral register".
Ministers have rethought their original proposals so that anyone on the existing household-based register whose existence can be confirmed by a separate official database will now be automatically transferred onto the individual electoral register.
The government says:
"Over the past year we have carried out a series of data matching pilots, comparing electoral registers in twenty-two areas with a range of data from public authorities. While the final evaluation is still being concluded, the evidence so far suggests that comparing entries on an electoral register with information held by the Department for Work and Pensions (DWP) allows us to confirm as accurate a significant majority (an average of two thirds for that data set alone in the pilot areas) of entries on the registers concerned.
"Subject to the results of the full evaluation, and further testing this year, we are therefore minded to build on this to simplify the transition to [Individual Electoral Registration] for the majority of electors. It is now our intention that the names and addresses of all individuals currently on an electoral register will be matched against the data held by public bodies such as the DWP and local authorities themselves. If an elector's information can be matched, the individual will be automatically placed onto the new IER register and would not need to take any further action to be registered under IER. Only those people who cannot be confirmed automatically will be invited to provide identifying information to be verified. This should simplify the transition process for the majority of electors, reducing the number of people required to provide personal identifiers and will also allow [electoral registration officers] to free up resource to target the smaller group of people whose information cannot be matched and those who are currently missing from the register."
The new Cabinet Office publication also includes the first hard data on how well those data-matching pilots are going.
The table below shows matches between the DWP customer information system (CIS) and the electoral register in different areas.
Its stands out just how badly the data matching has performed in some parts of London, and raises the prospect of differential transfers between old and new registers in different parts of the country.
Potentially, parts of London (and perhaps other inner city areas) could see far more people drop off the register than other areas of the UK if this table is any guide.
Although there will still be other work to encourage people to re-register, such as advertising campaigns.
The government says:
"It is our view that this finding, if confirmed, opens up a significant opportunity to simplify the process of transition to IER. Our proposal in the White Paper was that all electors will be asked to apply individually in 2014-15 and provide identifiers which will then be verified against the DWP CIS. Instead, our view is that if data matching could be used to verify the majority of entries currently on the electoral register as these results suggest, a majority of people can be transitioned into IER without having to take any further action. Each entry would still be individually verified but without requiring an individual to provide any extra data. We believe that this will simplify the transition for the majority of people, reducing the data burden (so less new personal data is required) while providing assurance about the accuracy of the register. It will also enable more focus on electors missing from the register."
The document also suggests that some problems with attempts at data matching occurred because the DWP supplied old records to the councils involved in the pilots.
"We have learned lessons from the pilots in terms of making the process of data matching more effective and efficient and have put some of these lessons into practice to provide nine of the twenty-two pilots with more up to date DWP records."
The Electoral Commission is expected to publish its full review of the data-matching pilots in the next few weeks.
Policy information and political accountability
February 4, 2012
No comments
Neil Williams has written an excellent post on the Government Digital Service blog about what constitutes a government policy and how the single gov.uk website (now in beta) should present information about it to citizens.
The post sets out how the gov.uk team is using as a working definition of policy "statements of the government's position, intent or action".
This includes mandatory information on the issue and actions being taken in response to it, plus optional information on the policy background, who is engaged with it, who is being affected by it, the legal framework, partner organisations and related news and publications.
It highlights just how much unpacking the simple word 'policy' seems to require.
Occasionally the government talks in a language that implies there are policies and meta-policies.
The original motivation behind the development of departmental business plans was not so much for Whitehall to achieve something itself, but for it to put in place the frameworks, systems and incentives for others to achieve it.
In which case the 'policies' may become more diffuse, being developed and implemented by a variety of local providers and getting blurred with the day-to-day decisions and delivery, operations and implementation.
Anyway, this post aims to suggest a couple of ways in which the presentation of policy information online could be used to significantly enhance political accountability, in line with my personal definition of eDemocracy.
There are two classes of information, open data and freedom of information releases, which might implicitly be covered by the phrases "statements about actions" or "related publications" but which would benefit from being explicitly mentioned given their potential importance.
They might not be relevant in every scenario, but as well as the statements and speeches about what the government says it is doing, policy pages should also include the datasets which might provide some kind of evidence about what it is actually achieving.
Given that some of the most significant policies (those in the departmental business plans at least) have targets or intended outcomes associated with them, and deadlines, it should be possible to pull out the data from the information strategies which is being used as an indicator for delivery success.
Progress on each of the business plan objectives is already being tracked in monthly updates, but more could be made of this information than is currently the case.
Some data visualisations of this information might also be a massive step forward for visibility and accountability, certainly on the headline commitments if not on every last detailed policy.
Further down the road, gov.uk could also go further on some of the other open data that's out there and relate spending figures to policies so everyone can see how much a policy costs.
Another significant step would be to publicly assign the policy to people or bodies in the departmental organograms which are available now, so it is also clear who is responsible for it.
Adding in this kind of information (gov.uk might be planning some of this already for all I know) – and making it available for re-use and publication anywhere else – could significantly transform the quality of information available to citizens about what their government is both trying to do and actually delivering.
Tech Policy Scorecard
All edemocracy and politics posts
