Most statistical information about online crime comes from vendors trying to flog products and governments eager for new police powers. We need something better. A lot better.
"New research shows identity theft affects one in six people," began last week's announcement by Australia's Attorney-General Robert McClelland. It's a typical example.
One media outlet -- I won't link, I'll save them the embarrassment -- carelessly reported this as "one in six people have fallen victim to identity theft." It sounds like a crime wave. I'm sure that was the government's intention.
"Nine in ten people are concerned or very concerned about identity theft and misuse," McLelland said. "This week, the Government introduced new legislation into Parliament to strengthen cyber security laws and enhance Australia’s ability to combat international cyber crime."
Right then. Strategy clear.
Online crime is certainly on the rise. But are things really that bad?
This research was a survey of 1201 people aged 18+ who use the internet, conducted by Di Marzio Research in May as part of an "omnibus survey" -- that is, people were asked questions on a range of topics, with customers paying by the question. In this survey, the Attorney-General's Department paid for six questions about identify theft.
Question 2 asked, "In the last six months or so, have you or someone you know had their identity information stolen or misused?" Five percent said it had happened to them, another 12 percent to someone they knew, a total of 17 percent. One in six.
But "affected by identity theft" is being used like "affected by cancer". If one member of a family is diagnosed with the Big C, it causes grief. I'm hardly "affected" if one of my Twitter followers had a few DVDs fraudulently charged to their credit card. I reckon the number actually "affected" by identity theft would be that five percent. One in twenty.
Identity theft included "to embarrass or misrepresent the person to whom the identity information belonged (either on the internet or offline)", some 8 percent of the incidents reported. That'd include such heinous crimes as posting an update in someone else's Facebook account. Technically a cybercrime, since it's accessing a computer without proper authorisation. Certainly wrong. But it's hardly in the same league as emptying their bank account.
Rather than the headline one in six, we're really looking at something more like 4 percent of people being victims of a crime. And given the sample size of 1201 out of an internet-using adult population of around 16 million, the margin of error is 2.8 percentage points.
In other words, the true number is likely to be anywhere between about 1 and 7 percent having been a victim of identity theft in a six-month period.
Or in other words again, we really have no idea.
This isn't a criticism of Di Marzio's research. This is just the (lack of) accuracy you can expect from these standard omnibus surveys. Is this really good enough for deciding government policy?
Sure, it seems safe to say that the vast majority of people are concerned. But how does this compare with the level of concern about other issues? Are their concerns well-founded? How does this level of identity theft amongst internet users compare with that of people who aren't online? How does the likelihood of being a victim of identity theft compare with the likelihood of being victim of other crimes?
If we're deciding how to allocate finite government resources, these are all important questions. Yet they can't be answered by this research. They can't be answered by the glossy scaremongering coming from security vendors. And they can't be answered by police.
Detective Superintendent Brian Hay of the Queensland Police told the AusCERT conference in May that he reckons less than one percent of online crime is reported to police. Recently the chief technology officer of AVG, Yuval Ben-Itzhak, called for mandatory reporting of all cybercrime so we could get accurate figures.
I'd go further. I'd make sure that online crime is properly researched by independent bodies like the NSW Bureau of Crime Statistics and Research and its equivalents -- although that'd require Victoria to create such an independent body.
Even the Australian Institute of Criminology (AIC) has little knowledge.
"As few police agencies identify cybercrimes separately, this section presents the results of Australian surveys of computer crime and security by AusCERT, Microsoft and the Australian Competition and Consumer Commission (ACCC)," says their Australian crime: Facts & figures 2010.
The AIC tells me that online crime continues to be a hidden crime. It's under-reported, under-researched and needs a lot more attention all round.
Sounds to me like we need more funding. Lots of it. And fast.