2011 Census: how private?
How private will the results of the 2011 Census be? Officially, privacy is absolutely guaranteed under the Census Acts of 1920 and 1990.
But there are some grounds for fearing that the data to be submitted on March 27 next year by the UK’s 25 million households may not be quite as private as it has been in the past.
Jill Leyland, Vice-President of the Royal Statistical Society, has written to the Director General of the Office for National Statistics, Stephen Penneck, asking how the ONS would respond to requests for Census information from the police or the security services.
The source of her concern is a clause in the 2007 Statistics and Registration Act that might, in her words, “effectively oblige the ONS under certain circumstances to hand over individual data for non-statistical purposes to the police or security services.”
Clause 39, Confidentiality of Personal Information, is where the issue lies. The clause says that personal information must not be disclosed by ONS (referred to in the act as “the board”) or its employees and that breaches can carry a jail term of up to two years.
But the clause then goes on to say that this does not apply to a disclosure which:
- (a) is required or permitted by any enactment
- (b) is required by Community obligation
- (c) is necessary for the purpose of enabling or assisting the board to exercise any of its functions
- (d) has already lawfully been made available to the public
- (e) is made in pursuance of an order of the court
- (f) is made for the purposes of a criminal investigation or criminal proceedings (whether or not in the United Kingdom)
- (g) is made in the interests of national security, to an Intelligence Service
- (h) is made with the consent of the person to whom it relates or
- (i) is made to an approved researcher.
Subclauses (f) and (g) are those to which Jill Leyland refers to in her letter. In fact, subclause (g) relating to security services was repealed by a clause in the 2008 Counter-Terrorism Act which directs that Section 39 (4) (g) be omitted from the Statistics Act.
A further modification was made to the Statistics Act under the Disclosure of Pupil Information) (England) Regulation 2009, but that appears simply to relate to subclause (a). The regulation gives the Secretary of State for Children, Schools and Families (as it then was) the right to pass on pupil information to the ONS, and the modification to the Statistics Act 2007 relates to the circumstances in which the ONS can disclose that information to third parties.
So it appears that subclauses b, e, and f – the right to disclosure under Community obligations, after a Court Order, or to police pursuing a criminal investigation, still apply.
This is not quite the impression given by the ONS. On the Census website, under the heading How the Information is Kept Confidential, ONS says: “The confidentiality of personal information is of paramount importance. The census forms were collected and processed in secure conditions, and the Census Confidentiality Act 1991 gives legal protection by making the unauthorised disclosure of personal census information an offence"..
In fact, the 1991 Act provides a defence for anyone who does disclose information, by allowing the argument that at the time of the alleged offence the perpetrator “believed he was acting with lawful authority or believed that the information in question was not personal Census information, and that he had no reasonable cause to believe otherwise”.
This makes it particularly important that the rules be clear, and understood by everybody.
The 2011 Census will be more detailed than every before, with a form 32 pages long. Given that the permitted disclosure to the police refers to criminal investigations whether in the UK or not, many people (and not only criminals) may be reluctant to answer all of them. Among those of a suspicious nature, the fact that the Census' main contractor is a US company, Lockheed Martin, will hardly calm nerves.
The Northern Ireland Statistics and Research Agency conducted a lengthy Privacy Impact Assessment of the Census in Northern Ireland which says “no personal data will be shared with any other party, except organisations acting on behalf of NISRA” and “personal 2011 data will not be provided to any other government department or local authority.” There is no mention of the clauses in the 2007 Statistics Act that appear to contradict this claim.
Collecting Census data has become an increasingly controversial in recent years as the demand for more and more personal data collides with growing worries about intrusion and privacy. “Long-form” data collection - as in Britain - is especially controversial.
The head of Statistics Canada, Munir Sheikh, resigned this summer after the Conservative Government decided to abandon the long-form census and replace it with a voluntary survey. The Government claimed that the survey would provide results of sufficient accuracy, while minimising the invasion of Canadians’ privacy, but Sheikh disagreed. He argued that a voluntary survey could never replace the mandatory long-form census.
The US abandoned the long form years ago, and its 2010 Census consisted of just ten questions, which it claimed took only ten minutes to complete. It achieved a respectable 72 per cent participation rate, leaving the other 28 per cent who did not respond to mailed forms to be followed up individually.
The US Census Bureau backs up its basic count with the American Community Survey, a mandatory survey that has been sent to about 3 million addresses every year since 2005 in a rolling programme. The ACS asks the personal questions that are excluded from the Census itself.
But even the ACS has come under fire from the Republican National Committee, which says it is “overreaching and intimidating” and that in insisting that Americans answer the questions, the Census Bureau is behaving “exactly as a scam artist would.” The Census Bureau, it said, was spending billions of tax dollars to violate the right and invade the personal privacy of United States citizens.
Many of the same criticisms of the UK Census were made, in more moderate language, by Francis Maude before he became the minister responsible for it. He’s now swallowed his doubts and allowed it to go ahead.
But by sticking to a mandatory long-form Census while other nations are abandoning it, the UK is out on a limb. If too many people refuse to fill in the forms (risking a £1,000 fine) the Census will fail. That is why doubts about the privacy issue are so important and why clarification so urgent and necessary.
Anonymous (not verified) wrote,
Tue, 14/09/2010 - 09:15
One justification for Census per se touted by the government etc is that there needs to be a proper basis for forward planning of provision/services and for checking/confirming current provision. Quite simply there needs to be an evidenciary basis for government direction/focus in all its forms and at all levels. Few would disagree with this premis. Obviously, the true worth of Census results/outcome can become seriously questionable if doubts arise as to the 'honesty' of answers collected, when the answers provided by Census respondents themselves become distorted by individual fear of privacy and confidentiality. With full privacy and confidentiality, few would disagree that there is anything intriniscally wrong with "Long-form" data, indeed it is to preferred, as it presents perhaps the only way to get sufficient data properly identfied and established, and hopefully without institutional bias.
Additionally, there are many areas where the Census itself is flawed, by either asking the wrong questions, or by limiting the range of answer options. A very good example of the latter is the range of 'family' type options that are selectable, which so many of the outcomes of the Census are correlated to. It intrinsically ignores the many tens, if not hundreds, if not possibly millions of chidren (pre-18's), across the nation who have more than one place of residence, or spend not inconsiderable parts of their lives being cared for in more than one place. There is no open recognition whatsover of this class of children and therefore the follow on consequences/needs of that sector in terms of service provision/attitude is/will never be realised. With the Census proper only taking place every decade, then outomes/evidence once set from a Census, are almost fixed in stone for the next 10 years !
Thus overall we have a Census organisation, which is gradually vearing away form fuill confidentiality, that by it's very stance are endangering the 'honesty' of the Census data outcome(s) and one that appears not willing to alter the answer options in a sensible manner in order to elicit/allow/enable a full range of 'honest' answers/data to be collected and distilled. Thus many would conclude that neither the Census nor the Census organsiation/commission is any longer 'fit for purpose'.
Eric T (not verified) wrote,
Mon, 04/10/2010 - 16:50
It is worth remembering that Christopher Andrew's 2009 book 'The Defence of the Realm: the authorized history of MI5' records (page 48) that the security services were given access to the 1911 Census returns, which "were used to record the particulars of all male aliens aged eighteen and above of eight nationalities (in particular Germans and Austrians) living in areas which would be closed to aliens in wartime. Information on aliens taken from the Census was then circulated for checking to chief constables, who were also asked to take note of those on the Register [of Aliens] in their areas."
Shallow (not verified) wrote,
Thu, 10/03/2011 - 10:24
So did Jill Leyland ever get an answer from the ONS? I have always genuinely believed that census data could not be accessed by the police or MI5 - until now.