Stats: questioning the ACC claims

[Ixion]

More detail here .

It appears that the figure is based on a survey of 354 motorcylists in 1998 !

Moreover at that time the fleet was estimated at 58000. Now it is nearly three times that size. But crashes (ACC figures) have only gone up 27% . So that halves the figure immediately!

I suspect the rest of it is just as shonky.

[bogan]

More detail here .

It appears that the figure is based on a survey of 354 motorcylists in 1998 !

Moreover at that time the fleet was estimated at 58000. Now it is nearly three times that size. But crashes (ACC figures) have only gone up 27% . So that halves the figure immediately!

I suspect the rest of it is just as shonky.

Hmm, very interesting information, I wonder if we can confirm whether these are the actual stats they are basing their figures from. Would leave them wide open to a good thrashing:

National, using statistics from last time we were in government, they're still current right? ... TUI

[Ixion]

Bloody hell

They included car passengers.

So if a car with four passgeners travelled 100km, that was considered as 400 km !

No wonder it made bikes look bad.

[dipshit]

And that the figures are crap can be seen at a glance by the "motorcyclist ride for only 44 hours per year".

Average.

It isn't that surprising. There must be a lot of bikes that don't get ridden much when you go looking at the age of the bike and mileage.

Compare this page's year and mileage http://www.trademe.co.nz/Trade-Me-Mo...1255-2509-.htm

against this page... http://www.trademe.co.nz/Trade-Me-Mo...0268-0334-.htm

[motorbyclist]

hadn't considered they might do it per kilometer travelled - you'd think that was a bit shonky if we don't to as many k's then on an annual basis (on which the rego levy is based) that 16 times would come down wouldn't it?

ie using numbers out of the air, bike rider is 16x/km more likely to crash, but only does 1/8 of the k's in a year, making him only twice as likely to crash in that year he pays his levies?

good ol' statistics eh?

[dipshit]

ie using numbers out of the air, bike rider is 16x/km more likely to crash, but only does 1/8 of the k's in a year, making him only twice as likely to crash in that year he pays his levies?

Nope, say your average bike only does 1/8 of the mileage your average car does in a year. This and the fact bikes only make up 3.5% of registered vehicles in NZ... makes it all the more worse that a third of road injurie claims are from motorcyclists.

[Mystic13]

Nope, say your average bike only does 1/8 of the mileage your average car does in a year. This and the fact bikes only make up 3.5% of registered vehicles in NZ... makes it all the more worse that a third of road injurie claims are from motorcyclists.

It doesn't help when we misquote facts. We're not 1/3rd from what I read. We are 1/3rd compared to cars. "Other vehicles", "bus", "truck" etc are excluded. In another set of stats the SUV's and vans were separated out and I have to wonder if they are the "other vehicles". But when you see these at about the same as cars it means effectively the motorcycle cost is halved.

What is the SUV only rate?

Moving on if the 16 or 18 x is a kilometre thing it's completely irrelevant. The fact is a rider using their bike compared to a driver in theirs. The ACC stats clearly show a much reduced figure.

Again this is still only a car and motorcycle comparison with SUV's and van's left out of the discussion.

ACC is being mischevious with their manipulation of stats.

[maxf]

OK guys, newbie here, but has anyone checked the ACC 2010/2011 full year, please officially submit on me, consultation PDF?... this one here: http://www.acc.co.nz/PRD_EXT_CSMP/gr...ctrb118085.pdf

Go to Pg28 - its got the relative risks & costs ACC has calculated vs a car - their "relativity" baselines. What boils my brake fluid is that the WORST rating for a bike (their 126-600cc band) is only 443% more likely than a car - eg 4 TIMES, not this crappy-math "16 times"... The only figure that approaches that is the 1222% relative likely cost (relative chance of accident x relative likely cost of injury) for the 600+cc class... and that's still not 1600%, even if they were quoting comparable numbers!

OK - EVEN BIGGER KICKER. Go to the table on Pg 29 - the "culpability factor" calculations for motorcyclists only being at fault 58% of the time (lets not debate the correctness of that yet!)... so they take the relativity factors from Pg 28 and mutliply by 0.58 - SO the figures they're quoting in all the press packs are EVEN MORE ERRONEOUS.

Now - remember this is the official consultation document for however many billion in levy income? Have a look at the incendiary little second table on Pg29 - the one that gives the cross-subsidy conclusions... ITS CALCULATED WRONG. They use the full figures, not the culpability adjustment figures - they're all 42% out (for the 600cc+, the $3770 quoted / car cost @ $312 = 1205%, not the 708% under the culpability corrections). This carries thru into the totals too ... the recalculated-from-their-stats cross-subsidy is around $100-115M, and the cost-per bike gets lower as the on-road stats increase - which they also seem to have under-quoted... see the second worksheet in the spreadsheet with MOT 2008 rego figures from http://www.landtransport.govt.nz/sta.../docs/2008.pdf, around pg 59 from memory...

I've put this in the attched spreadsheet - please check if you can see if I'm wrong anywhere. I've emailed Smith, Key & Brownlee (local MP) to raise the issue that they're glibly spouting incorrect facts, but have only had replies referring me to the consultation docs and/or Smiths' PR releases... Getting a definite impression of a do-not-care agenda here!

Max

Personal position: Christchurch, 25yrs & no ACC claims, getting more pissed off with the lying than the proposed charge now Wouldn't mind if their insurance premiums reflected some reality...!!

[maxf]

It's play-time analysis by the looks - playing with datasets on a computer and then drawing semi-connected but usually politically acceptable conclusions, without questioning the base data classifications - but they do quote it a lot - it seems to be the one with the stupid paragraph of

"Motorcyclists are considerably over-represented in both fatal and serious injry accidents in proportion to the distance travelled. The degree of this over-representation is confirmed by the latest LTSA travel survey [LTSA 2000a] which shows that injury risk per time travelled for motorcyclists including pillions is 18 times that for car van or ute occupants and four times that for cyclists. Motorcyclists in fact ride for far fewer hours than car drivers drive, on average about 44hrs per year compared to 280 hours for a car driver. Indeed, motorcyclists covered only 0.7% of the annual distance travelled by cars, vans & utes in 1997/98...." (pg 53)

Spot the errors?
- they're quoting stats 9 years old - by their own MOT stats, the average bike is 15yrs old, so we're getting back to covering 1985 era bikes with cable drum brakes & cross-plys here...
- they start off with throw-away comment re accidents per distance and then quote accidents per riding hours to justify it?
- would love to see where they get the 44hrs/yr figure, eg maybe one hour a weekend, not even every weekend? Even on a 5min commute (each way) you should be around 60hrs/yr
- they again use subjective comparisions - 0.7% of car travel distances, instead of comparing distance per vehicle on-road, 2.5M cars can rack up a heck more kms than 50000 bikes ...

Also quite depressing how the same people seemed to have carried some "street cred" from the DSIR and taken it private - they do credit an awful lot of their own prior reporting, also paid for by ACC!

Personal hate - on pg 56, "Motorcyclists are gregarious creatures and seem often to ride in groups."
Even sadder - some of your money went to pay for it.

Joy.

[motorbyclist]

http://www.kiwibiker.co.nz/forums/sh...d.php?t=110116

that there is the actual stats ACC should be using.

I'm told that Ixion found the study that ACC is basing their claims on: it's from 1998 and the sample size is only 365 bikes, out of 50k bikes rego'd at the time. Ignoring the insignificance of such a small sample, a decade has passed since then and rego'd bikes almost doubled

[Supermac Jr]

[url]sample size is only 365 bikes, out of 50k bikes rego'd at the time.

That sample is not too bad - 50k bikes, 5% margin of error and 95%confidence interval -> required sample size = 382

[Ixion]

It's play-time analysis by the looks - playing with datasets on a computer and then drawing semi-connected but usually politically acceptable conclusions, without questioning the base data classifications - but they do quote it a lot - it seems to be the one with the stupid paragraph of

"Motorcyclists are considerably over-represented in both fatal and serious injry accidents in proportion to the distance travelled. The degree of this over-representation is confirmed by the latest LTSA travel survey [LTSA 2000a] which shows that injury risk per time travelled for motorcyclists including pillions is 18 times that for car van or ute occupants and four times that for cyclists. Motorcyclists in fact ride for far fewer hours than car drivers drive, on average about 44hrs per year compared to 280 hours for a car driver. Indeed, motorcyclists covered only 0.7% of the annual distance travelled by cars, vans & utes in 1997/98...." (pg 53)

Spot the errors?
- they're quoting stats 9 years old - by their own MOT stats, the average bike is 15yrs old, so we're getting back to covering 1985 era bikes with cable drum brakes & cross-plys here...
- they start off with throw-away comment re accidents per distance and then quote accidents per riding hours to justify it?
- would love to see where they get the 44hrs/yr figure, eg maybe one hour a weekend, not even every weekend? Even on a 5min commute (each way) you should be around 60hrs/yr
- they again use subjective comparisions - 0.7% of car travel distances, instead of comparing distance per vehicle on-road, 2.5M cars can rack up a heck more kms than 50000 bikes ...

Also quite depressing how the same people seemed to have carried some "street cred" from the DSIR and taken it private - they do credit an awful lot of their own prior reporting, also paid for by ACC!

Personal hate - on pg 56, "Motorcyclists are gregarious creatures and seem often to ride in groups."
Even sadder - some of your money went to pay for it.

Joy.

The reason for the 44 hours is that the 354 bikes included ALL motorcycles owned by the sample group. Even if they weren't in use or were used by a different family member. They asked the respondents "list all vehicles in the household". So, Dad, who was filling the thing out, and recording HIS travel, dutifully entered his sons motorbike. Even though said motorbike wasn't in use . How many garages and sheds have an unused bike in them, even today. There were a lot more in 1998.

I'd like to get hold of the number of people who actaully recorded motorcycle journeys. I bet it was a LOT less than 354.

Their figure for average annual travel for motorcycles comes to about 800km per bike.

The Bailey report would be interesting.

[Ixion]

Does anyone (trade maybe ?) know of a data source that shows the average annual mileage of motorcycles.

The MoT survey, they one that produced the "bikes 18 x as dangerous" works out the average travel to be 800km per bike!

I'm thinking they must be including a lot of "back of the shed" unused bikes.

[Squiggles]

I dont fully remember but Phil Read who wrote "The New Zealand National Survey of Road Riders" may have some data on that

[sinfull]

Were all forgetting the fact that all off road motorcycles, quads etc are included in the figures they use !

Does anyone know where the stats are that can have them taken out of the equation ?
Bet that would make things look bloody good for an argument !