This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Find out more here

FEATURE1 September 2009

Data distrust: lies, damned lies and statistics

In an era of spin, public trust in statistics has taken a knock. Add in fears about the handling of personal data, and you have the makings of a crisis of trust. Richard Young surveys the scene.

?Much of the growing uncertainty about privacy and data protection has only emerged in the past ten years, thanks to three key technologies.

First, the web. The very fact that most computers are now connected to each other means that it’s much harder to verify counterparties or to keep control of your own digital information. High-profile email scams have alerted the public to the dangers of using the web, and although hacking is rare it’s hard to find anyone who hasn’t been troubled by a virus or trojan contracted via the web.

Second, portable storage – in the form of laptops, smart phones, memory cards and flash drives. Not so long ago it was actually quite hard to lose a large amount of data. Now, with terabyte-sized portable drives and dozens of gigabytes fitting on a keychain, it’s as easy as forgetting your umbrella on the train. Just one example, in October 2007 two CDs containing the child benefit data for 25 million people were lost in transit between HM Revenue & Customs and the National Audit Office. Third, massive and interconnected databases. This is becoming even more pertinent with the growth of cloud computing (when applications and databases are stored on centralised servers). Although many argue that specialist hosting companies are better able to manage data security and disaster recovery, it’s still a case of surrendering control, and a single breach could result in huge amounts of data being compromised.

But the biggest backlash against technology is faced by the state. The UK National Health Service’s massively delayed and over-budget IT project, for example, has convinced many people that the government “can’t do databases”. But there’s also huge resistance to the Conservative Party’s mooted alternative – handing over the project to Google or Microsoft. The government’s plans for a national identity database also face public hostility, rooted partly in fear of a surveillance state, partly in expectation that such a system would be costly and unworkable, and partly in concern over what could happen in the event of a security breach.

Getting personal
At the same time as the public are worrying about how others handle our data, the way they deal with their own has been transformed. People seem only too happy to share more and more information about themselves on the web, a situation that throws up huge opportunities and challenges for online businesses. The most high profile problems have centred on social networking sites, which have to maintain a delicate balance between privacy, community and commercial reality. As early as 2007 Facebook was forced to re-think its search settings which allowed users to filter supposedly private profiles by ethnicity, religion and sexual preference. The same year, its ad targeting system Beacon was withdrawn after it emerged it could track user activity on third-party sites.

More recently, Facebook had to rescind changes to its terms of service, interpreted by users as a contract giving the social network ownership over their data. “We’re at an interesting point in the development of the open online world where these issues are being worked out,” founder Mark Zuckerberg later explained. “It’s difficult terrain to navigate and we’re going to make some missteps… we take our responsibility to help resolve them very seriously.” In other words: please don’t leave – you can trust us.

It comes down to a trade-off for consumers between convenience, efficiency and social interactions. Whole industries need to ensure that their own reputations for providing online privacy and security are maintained. Market research has become part of the debate, with Facebook and LinkedIn attempting to grow survey and panel selection capabilities. The problem for social networks is that targeting surveys or pre-selecting panels – just like bombarding users with ‘relevant’ advertising – can be seen as an invasion of privacy.

In the US, regulators are starting to take an interest in this kind of profiling. David Vladeck, the new head of the Bureau of Consumer Protection, told the New York Times recently: “If the marketers are right, and the consumers like behavioural advertising, then it should be no big deal. If marketers are wrong, and consumers really want to protect their data, and have only acquiesced in the current situation because they don’t understand it, then we’re going to need to rethink this writ large.”

In a recent survey, Q Interactive found that while many consumers in the US are open to targeted ads, the range of data they’d be prepared to share to get them is limited to things like zip code, sex and age. And attitudes to sharing the more granular demographic information needed by researchers might harden if people feel that they are being turned into a commodity.

Keep calm and carry on
Despite these fears about data handling, evidence suggests people are still prepared to share remarkable levels of detail with the research community. In June, for example, Research reported that the ONS had met little resistance to its decision to factor sexual orientation into some of its surveys for the first time.

The remarkable growth of online methodologies suggests that, far from scaring away the public or bringing their brands into question, research firms are succeeding in using new methods to pioneer new ways of interacting with people. But there are other, more subtle threats.

Public cynicism about politicians and the media is rife. In a recent GfK survey journalists were seen as trustworthy by 41% of respondents and politicians by just 18%, while market researchers soared ahead on a healthy 55%.

A bigger threat is the use of research tools and methodologies by companies and individuals not properly qualified to deliver reliable results. How many headlines have been generated using basic online polls on newspaper websites? How many door-to-door salespeople from power companies use surveys as part of their pitch?

As dealers in data, research firms must prove they can handle these technologies and relationships more adeptly than the social networks, corporations and governments whose activities have aroused the public’s suspicions.

Recovering a reputation

?Ever heard of truthiness? The term was coined by talk show host Stephen Colbert in 2005 to describe “things that a person claims to know intuitively or ‘from the gut’ without regard to evidence, logic, intellectual examination or facts”. Truthiness allows decisions to be made for the wrong reasons, and it feeds off mistrust in official sources of facts.

It’s yet another reason why government statistics have to be reliable. They’re the foundation for policy decisions, they determine budget allocations of billions of pounds and they inform the public about the performance of their politicians and civil servants.

But confidence in this data is shot. As early as 2004, Ipsos Mori had already uncovered massive lack of trust in official figures in research conducted for the UK’s Statistics Commission. In 2008, an ONS poll found a majority of the public thought official figures were manipulated for political purposes; just 36% of people believed official stats were accurate.

In response, the UK Statistics Authority was launched in April last year. Reporting to Parliament, in theory it’s free of departmental oversight, ensuring it can criticise ministers for using bad data or spinning the numbers. And while it has a budget of just £5m (triple that of the old Statistics Commission, admittedly), initial feedback seems positive. According to Royal Statistical Society vice president Jill Leyland: “There is still much work to be done. But… the Authority has passed its first tests with distinction. It is becoming a body that Parliament and ultimately the public can look to with confidence.”

Interviewed on NPR in the US in June, Statistics Authority chairman Sir Michael Scholar struck a realistic note about whether his team have regained the public’s faith: “In a way it might even be the opposite because the first impact on people is to think, ‘Yeah, well, this is what we’ve always expected. These figures are fixed.’ The second stage is people thinking that this authority exists and is a guarantor upon which they can depend. We’re beginning to see a bit of that. But I think it’ll take a long time.”

1 Comment

9 years ago

very good

Like Report