@@ MOT failure rates released ... BBC @@
The government agency which oversees the MOT system has backed down after 18 months and released data which shows how often different makes and models of cars and small vans fail MOTs.
This means that car and van buyers will now have access to the detailed MOT records of individual models, including reasons for failures. The figures show wide variation between different models, even when comparing vehicles of the same age.
The Vehicle and Operator Services Agency (VOSA), an arm of the Department for Transport, yesterday revealed 1,200 pages of detailed statistics on MOT failures following a freedom of information request made by the BBC in July 2008.
VOSA initially declined to supply the material, but last month the information commissioner ruled that disclosure is in the public interest and overturned VOSA's refusal.
The agency had maintained that publication could mislead the public and would breach commercial confidentiality. It argued: "The release of information relating to specific make and model would be likely to be commercially damaging to vehicle manufacturers whose failure rates appear higher, and therefore less favourable, than other manufacturers...this information would be likely to be used by some manufacturers to gain a competitive advantage, for example by publicising that their failure rate is lower than another manufacturer's failure rate for a comparable vehicle model."
When I wrote about the commissioner's decision to reject VOSA's case, prior to the material being issued, there was some disagreement in the comments as to how useful it might be. Now you can judge for yourself.
The data has now been placed on the VOSA ...site, although only in the form of a pdf file of 1,200 pages, which is a format difficult to analyse.
The agency insisted, for reasons that are unclear, that it would be several more days before it could supply the material in a more useful format such as a spreadsheet.
However I have converted the pdf into an Excel spreadsheet which I am posting here, so that you can analyse it yourself if you want to do so.
The information relates to MOTs in 2007, since that was the latest annual data when the BBC's FOI request was made in July 2008. VOSA now says it will issue the 2008 and 2009 statistics later this year, and will publish this material routinely in future on an annual basis.
The figures give average MOT failure rates for individual models of cars and small vans, as well as breaking them down by year of manufacture.
Not surprisingly older vehicles fail more often, and older models therefore tend to have higher average failure rates. However I have used the spreadsheet to examine vehicles first used in 2004, which were therefore all three years old in 2007 at the time generally of their first MOT.
Limiting this to the most popular models - those with over 20,000 MOTs in 2007 - produces the following table for these vehicles made in 2004:
http://www.bbc.co.uk/blogs/opensecre...04vehicles.jpg
The Society of Motor Manufacturers and Traders stresses that the roadworthiness of vehicles is influenced by how their drivers treat them. A spokesman said:
"Pass rates will be determined by a wide range of factors, but particularly the level of use, ownership profile and the frequency of servicing or regular maintenance."
A Vauxhall Motors spokesman commented: "Many of the failed items highlighted in this report, such as brakes, driver's view of the road, reg plates and tyres, are directly attributable to vehicles that have covered above average mileages. Since Vauxhall was one of the largest suppliers of fleet vehicles during the period surveyed (2004-2007), and fleet vehicles typically cover higher mileages than most, it comes as no suprise that some of our cars appear in this list."
I have also looked at vehicles first used in 1997, ten years old at the time of their 2007 MOT. Limiting this to the popular models with over 10,000 MOTs gives this table:
http://www.bbc.co.uk/blogs/opensecre...97vehicles.jpg
One small point to note: it's rare for a dataset as large as this not to contain some errors and you can see this one isn't perfect. For example, one MOT testing station seems to think there's an Audi 80 out there which was first used in 1883. There is clearly an enormous wealth of information in the 1,200 pages now available. It's possible to examine all sorts of angles - how less popular models fare, the records of individual manufacturers, how failure rates for various models change over time, the different kinds of reasons (brakes, lights, steering, and so on) which different models are most likely to have problems with.
Update 12:30, 15 January 2010: Thanks very much to everyone for your comments below so far, many of which help to illuminate some wider aspects of this.
What has struck me most in the response to my post is the way others have taken the data, now that it is publicly available, and presented it in various more accessible and useful forms. For those both within and outside government who want the large quantities of state-held data to be opened up, this is exactly what they hope and argue would happen.
A couple of readers have produced searchable databases - see comments 78 and 79. And comments 23, 24, 58, 60 and 71 are also relevant.
I have also been sent this table [69KB PDF] by Tony Hindle of Lancaster University Management School, which compares the overall records of manufacturers for cars first used in 2004 (whereas I compared popular individual models).
And Simon Marsh, a business analyst, has sent me a tidied-up spreadsheet with analysis by brand, model and age.
I should stress that I haven't examined any of these pieces of work to check their calculations are valid. The raw data remains available for those who are interested.
Various points have been made about reasons for MOT failure. The question when comparing models is whether these involve systematic differences between models. Mileage can clearly be a relevant factor in some cases (as Vauxhall and Renault have pointed out), and I hope we will be able to do some more work on this point.
A number of you have pointed out that there are plainly some errors in the VOSA data (these may well be errors in the records made at MOT testing stations which have then been supplied to and compiled by VOSA). This is indeed a point I made myself in the final paragraph of this blog entry. Errors of the kind identified do not invalidate the comparisons I have made between the most popular models, where a few incorrect dates won't matter. Such errors could conceivably undermine the robustness of the failure rates for much less common models (which is one of the reasons why I decided to limit my analysis to popular models).
الخبر منقوووووووول من BBC
===================
متهيألي ده بيوضح العربيات الأقل في الاعتمادية والمعروف عنها المشاكل الحقيقية وليست الأشاعات
وبين العربيات الأكثر اعتمادية
وده ليس كلامي مع انه كلام سبق قولته ولكن كان راي شخصي
وده دليل عليه مش اكتر ولا اقل
