500 cockroaches on a bus – or are there?

A recent expose of dodgy statistics in the UK about pests on public transport shows just how hard it can be to, firstly, get to the truth about unreliable or fabricated statistics that are uncritically reported, and, secondly, how hard it can be to get corrections made.

The London Evening Standard reported Rentokil’s press release:

The average commuter train contains up to 1,000 cockroaches, while seats can contain 200 bedbugs and 200 fleas, it was claimed today.

Rentokil say they also discovered that a bus was home to 500 cockroaches, along with dozens of fleas and bedbugs. Both claims were disputed by Transport for London.

Staff at Rentokil sprayed insecticide throughout the carriages of a train and a bus and then counted the bodies of insects.

Rentokil warned that infestations on public transport were are at an all-time high. “The average commuter will always be close to cockroaches, bedbugs and fleas,” said Savvas Othon of Rentokil.

“People eat on the move, and there is a lot of food left on seats. Pests are thriving.
“Although we looked at a train not running in London, we believe that London trains, both underground and overground, will have a similar number of infestations.

“The bus we studied was within the M25, and we are already in talks with bus and Tube operators about a new cleaning system we’ve developed, which heats the vehicles to kill the insects, and their eggs.”

A TfL spokesman said: “Rentokil has provided us with no evidence to support these claims and have not been in touch with us.

Ben Goldacre, on the blog Bad Science, was suspicious:

But Transport for London say they’ve had no contact with Rentokil, and that no such study has been done on their vehicles. I asked Rentokil for more details on what vehicles they had studied, where, and how, what was counted, how the bugs were collected, and so on.

After a bit of prodding, their PR company Brands2Life explained how these bugs were counted. No buses were studied, and no trains were studied either. How did people get the wrong end of the stick? What was that about with Savvas Othon? I have no way of knowing. Brands2Life and Rentokil both declined to show me what they sent to journalists, but in any case, contrary to what was said earlier, wherever it came from, these numbers did not come from measurements and counts, they are actually based on a “theoretical model”.

Models are handy. They’re a simulation of reality, based on a series of assumptions. Rentokil’s model for the number of bugs on trains and buses made some interesting assumptions, and you will have your own view on whether they make for a reasonable approximation to the real world.

They assumed, for example, that the railway carriage or bus was left alone, by itself, in isolation. They assumed this isolated carriage was helpfully furnished with a plentiful food supply. They assumed that the ratio of male and female bugs was perfectly optimal for breeding.

They assumed – surprisingly for anyone involved in modelling populations, surprisingly for anyone, really – that the population of bugs would be left entirely unchecked, with no external factors to control the mortality rate. They assumed that the siding or garage was controlled at a constant temperature all day and night, with no extremes, they assumed there were no trampling commuters, no cruel vaccum cleaners, no anything. In fact they assumed there was no cleaning, ever, and no passengers, ever. This was their model of insect populations on commuter vehicles.

“On the above basis” Rentokil’s PR explained to me: “it is possible that the stated numbers of cockroaches and bed bugs/fleas could live on a train carriage or bus.”

You will have your own view on whether you could trust an organisation that makes assumptions like these in estimating the “average” population of a bug. But it’s somehow unseemly that Rentokil, a company with £2,356m revenue, a 54% increase in profits in 2009 to £166m, and poised to pay £90m in bonuses to its top 3 executives, feels the need to make everyone afraid of public transport on a PR whim. There is also the ugly thought that Rentokil will do more business if they can make everyone scared of bugs on the bus.

And on March 2nd, the day before the cockroaches press release, Rentokil announced the single biggest ever contract in the history of their business: £200m over 5 years with London Underground.

Finally, after a twitterstorm, Rentokil issued the following apology.

Some debate around a story doing the rounds over last couple of weeks, when our PR Agency released numbers calculated on a hypothetical worst case scenario, which were presented as “average” or “typical”.

I thought it might be helpful to explain how we arrived at the numbers and where things went so wrong.

We had tested our new Entotherm technology on a bus and various buildings during the last three years of its development. So, when asked for a worst case scenario situation, we based our hypothesis on:

  • A bus, or anything else, being left by itself in an isolated place
  • With no external factors to affect the mortality rate (so the population would be left unchecked)
  • Then we assumed that there is a perfect male to female ratio that allows optimal breeding numbers
  • That the environment would be controlled to a constant temperature, with no extremes
  • Finally, there would be a plentiful food supply to support the numbers of insects

On the above totally theoretical basis it’s possible that very high numbers of cockroaches and bed bugs/fleas could survive, although it’s clearly a worst case scenario.

Now, obviously real life is not a hypothetical model. There are loads of contributing factors that would affect any insect infestation. For example regular cleaning, people unwittingly stepping on insects and, as in real life, there is not a perfect male to female ratio! All of which means that, in our experience, it is very rare to find heavily infested buses, trains or other forms of transport in the UK. Standards will vary around the world but UK standards are very high.

The point of the story was about a new process we launched – and are very excited about – to combat bed bugs and cockroaches.

We’re really sorry that the numbers that appeared in the media were wrong and misleading and we’ve put in place a number of measures to ensure it doesn’t happen again.

Adam Banks commented on the Rentokil blog:

So, let’s be clear. Your PRs seeded a story about how many bugs existed on public transport in London – a story guaranteed to send shivers up the spine of every commuter. Turns out not only was the data not based on any research into public transport in London, it didn’t refer to any real world scenario at all, only a set of obviously invalid assumptions. It was, in fact, made up.

And your idea of “an apology” for this is to dance around Ben Goldacre all day on Twitter, hoping the whole issue will just magically go away, and finally post a blog entry that tries to make it sound as if there was something reasonable about releasing junk data to journalists, then links to a press release that doesn’t even contain the claims referred to, which, as Goldacre has already established, were made in separate communications that you decline to make public.

All this, you say, was to publicise “a new process we launched – and are very excited about – to combat bed bugs and cockroaches”. Those same bed bugs and cockroaches that, we now know, may exist only on “a totally theoretical basis”.

March 12, 2010, 11:33 pm

Comments are closed.