Another CVPI fire death in Florida this time

[ifcar]

Those numbers were developed by Ford.  That alone makes it significant, statistically or otherwise.

Who compiled the numbers has no bearing on statistical significance. Data are not statistically significant if there isn't enough of it, so figures about a handful of very different accidents say nothing about vehicle safety.

[hounddog]

While I would like to say "thanks" for the un-needed lecture on data collection and use, I would like to point out that the interested party, in this case Ford, has a vested interest in the matter.  With over 24 years at the time of the study, I am more than satisfied that they (whomever they were) able to find sufficient crashes with sufficient similarities. 

Two decades of data was studied.  How do I know that?  The Crown Victoria is now nearly 30 years old. 

Who compiled the numbers has no bearing on statistical significance. Data are not statistically significant if there isn't enough of it, so figures about a handful of very different accidents say nothing about vehicle safety.

[hounddog]

Even with their vetted interest in the matter, Ford was still able to collect data which shows that their car is far more likely to be involved in these types of incidents than vehicles with specifically similar design.  It supports Randys causes for concern. 

Of course, it does not provide concrete evidence to support his concerns only that they are warranted.

[ifcar]

The Crown Victoria's platform dates to 1979, but it was still the LTD then.

Two decades of data doesn't necessarily mean statistically valid data, particularly when the safety complaints are about the current design. If the data is 10 accidents over two decades, the data says nothing.

[hounddog]

Well, we both know that is not even close to the case, so your point is moot.  There were hundreds of rear end collision crashes studied, as simple laws of probability would suggest. 

[ifcar]

Well, we both know that is not even close to the case, so your point is moot.  There were hundreds of rear end collision crashes studied, as simple laws of probability would suggest. 

Yet when I asked how many accidents were included in the statistics you provided, you couldn't/didn't answer. Is it hundreds? And if it is, is a few hundred over 20 years statistically significant?

That's impossible to know from the pretty chart you provided.

[sparkplug]

It is an older design. I wonder if the police would consider the new Ford Taurus. I know it's front wheel drive but it's also but roomy and has good power.

[Raza]

It is an older design. I wonder if the police would consider the new Ford Taurus. I know it's front wheel drive but it's also but roomy and has good power.

I think I've heard of some departments getting the Taurus back when it was the 500, but it never seemed to catch on.  I see more Chargers and Impalas than anything nowadays, although the town I'm currently in is nearly 100% Crown Victoria and my hometown is nearly 100% Tahoe.

[hounddog]

Yet when I asked how many accidents were included in the statistics you provided, you couldn't/didn't answer. Is it hundreds? And if it is, is a few hundred over 20 years statistically significant?

That's impossible to know from the pretty chart you provided.

Well, since the chart, which someone other than me provided, does not give that information how would I know?  Every crash is statistically significant, that is the entire point of the existance of statistics; to document and examine each incidence of occurance.

Otherwise, what exactly are you saying?  That the dozens and dozens of people burned, or burned to death, are also statistically insignificant and therefore irrelevant because they only number xxx within 20 years as opposed to xxx? 

[ifcar]

No, I'm saying that you can't necessarily determine a trend from too few data. You don't know how much data were used to create the graph you posted, so you therefore don't know if the theory that the Crown Victoria is more prone to explode is statistically valid.

[hounddog]

Again, I did not post it. 

Secondly, and most obviously, Ford seemed to think so to the point that they put out a news release saying it was.

Ford has admitted the car is flawed, on several different occasions including Federal Court in a 42million dollar lawsuit they lost.   The car has the potential to be very dangerous and deadly, they admitted it.  Why is that so hard for you to deal with?

[hounddog]

Ok, since you want to get technical...

Statistical significance simply means that a result is unlikely to have occurred by chance.  The test for this is typically hypothesis testing using the p-value.  This test basically looks at the probability of a false positive or a false negative. 

The population in this study would have been "all crown victorias" (and all Impalas and Ford Tauruses as well) as evidenced by the data displayed in the chart.  The data is the likelihood of fatal impacts where rear-end fires occured (converted to a rate to compare apples to apples across models).  THIS IS THE DATA - NOT THE SAMPLE!! In this instance, there is no sample, the entire population is being examined as is evidenced on the rates that were calculated.  Hypothesis testing is commonly performed on a small sample, so sample size is a moot point by that standard as well. 

The hypothesis testing in this case would most likely look at the rate of fatalities in the crown victoria (police) vs. the other models and ask what the probability that the higher rate on the crown victorias happened by chance.  With such a large difference in rates, I would imagine that the p-value would be fairly small for this study.

[ifcar]

Again, it depends on how much data is involved. To use an extreme example, if the entire population of Crown Victorias in fires Ford decided to count is 3 and the number of Caprices is 1, that looks bad for Ford but isn't statistically significant.

There are too many factors at play to necessarily draw a conclusion from a set of data about relatively few accidents.

[Minpin]

Again, it depends on how much data is involved. To use an extreme example, if the entire population of Crown Victorias in fires Ford decided to count is 3 and the number of Caprices is 1, that looks bad for Ford but isn't statistically significant.

There are too many factors at play to necessarily draw a conclusion from a set of data about relatively few accidents.

I'm guessing here, so take it easy but...


Look beyond the numbers for a second. Obviously the resident LEOs on this board have a inherent fear of CVs and feel they are a threat to themselves and their fellow co-workers. So, this fear cam from somewhere? Where? Real world situations I suppose, right, i mean they are LEOs and all. So, even if they aren't statistically mathematically bannaphonnetically proven to be more dangerous I would imagine every time a LEO in this nation gets into their CV to do their job they have a fear in the back of their head knowing if they get in a wreck it's a good chance it could end VERY badly. So, this would in turn affect the ability for them to do their jobs properly, and maybe they wouldn't pursue whatever as forcefully or be as brash as they need to be for fear of a inept vehicle (in their heads).

Honestly, if I was a LEO, and word was CVs wee unsafe, how could you possibly think to yourself, "ohhh but the data is to homosexual, obviously it's as safe as a charger".

Not everything can be solved with maths and sciences.

[ifcar]

People being afraid of them is very different from a design flaw. Both are Ford's problem if departments choose to buy their cars based on that, but in only one of those scenarios is it Ford's fault.

[JWC]

Look beyond the numbers for a second. Obviously the resident LEOs on this board have a inherent fear of CVs and feel they are a threat to themselves and their fellow co-workers. So, this fear cam from somewhere? Where? Real world situations I suppose, right, i mean they are LEOs and all. So, even if they aren't statistically...I would imagine every time a LEO in this nation gets into their CV to do their job they have a fear in the back of their head knowing if they get in a wreck it's a good chance it could end VERY badly.

Honestly, if I was a LEO, and word was CVs wee unsafe, how could you possibly think to yourself, "ohhh but the data is to homosexual, obviously it's as safe as a charger".

Not everything can be solved with maths and sciences.

The Crown's problem isn't with just getting into a wreck.  It is almost inherit to the task it performs.  Namely, sitting along a interstate and being struck from behind by another vehicle traveling 80mph.   

In the case of the OP, if the vehicle struck an object broadside, then the analysis doesn't apply.

You could almost say the fault lies mostly with the general public, since they are not paying attention and ramming LEO vehicles.  This and the number of officers struck while standing along the roadway is what led NC to impose heavy fines on drivers who do not change lanes when approaching a LEO alongside the road.

[TurboDan]

It is an older design. I wonder if the police would consider the new Ford Taurus. I know it's front wheel drive but it's also but roomy and has good power.

The only place I've ever seen the 500/Taurus used is in NYC by the NYPD. 

I definitely have noticed many departments locally switching from CVs to other vehicle, though.  Generally, they've been buying the Chargers, though a couple departments have gone with the Impala.  Really, though, I've seen a good number of departments switching to SUVs around here - namely the Dodge Durango.  Since we're in a coastal area, the departments often are responding to incidents on the beach, which has necessitated this.

The Manasquan, NJ police recently retired their fleet of Jeep Wrangler police vehicles and switched to Jeep Liberties.

[GoCougs]

IMO it points to the simple fact that the Panther platform is an ancient, inherently less-safe body-on-frame design that should have stopped production at least decade ago.

[hounddog]

Again, it depends on how much data is involved. To use an extreme example, if the entire population of Crown Victorias in fires Ford decided to count is 3 and the number of Caprices is 1, that looks bad for Ford but isn't statistically significant.

There are too many factors at play to necessarily draw a conclusion from a set of data about relatively few accidents.

Two things;
I strongly suggest you revisit the definitions of "population" and "sample."

"Population" means "all" or "everything."
"Sample" is a representative group of a population.  You cannot chose to have a population, only a sample.

The Central Limit Theorem basically says that at some point the statistics for the sample approximate the statics of the population. 

Generally speaking, statisticians will say that 30 data points is usually a good sample.  However, in order to be precise a sample size calculation should be done. 

Also, generally speaking, typically when they run rate statistics, such as the ones used in this study, the rates are run off the population and cannot be run off the sample.  It makes no sense to run rate statistics on a sample.

If you would like me to recommend a basic statistical text for your enjoyment, I have several in my library that would meet your needs. 

[ifcar]

Two things;
I strongly suggest you revisit the definitions of "population" and "sample."

"Population" means "all" or "everything."
"Sample" is a representative group of a population.  You cannot chose to have a population, only a sample.

The Central Limit Theorem basically says that at some point the statistics for the sample approximate the statics of the population. 

Generally speaking, statisticians will say that 30 data points is usually a good sample.  However, in order to be precise a sample size calculation should be done. 

Also, generally speaking, typically when they run rate statistics, such as the ones used in this study, the rates are run off the population and cannot be run off the sample.  It makes no sense to run rate statistics on a sample.

If you would like me to recommend a basic statistical text for your enjoyment, I have several in my library that would meet your needs. 

Statistics are only useful when they're a collection of identical or extremely similar events. A handful of collisions at various speeds would not be the example your statistics textbooks would use to illustrate anything but a potential error.

[hounddog]

Well, you can say what you want about it, but if the numbers were not statistically significant Ford would bring in experts to make that point evident.  They would also not release this information, and they would not be paying out financial rewards to survivors based on that fact.  In short, your point is invalid based on what we know of Ford and their past practices and statitics.

[ifcar]

Well, you can say what you want about it, but if the numbers were not statistically significant Ford would bring in experts to make that point evident.  They would also not release this information, and they would not be paying out financial rewards to survivors based on that fact.  In short, your point is invalid based on what we know of Ford and their past practices and statitics.

Because we know that all legal outcomes are infallible, right?

[hounddog]

You are assuming facts not in evidence regarding the crashes.  You are wrong here, and I have shown you how and  why,  move on.

[rohan]

Yes there are but that should end.  The Panther is about 28 years old- and many departments aren't about to risk their officers lives in them anymore.  Mine included.  The guy who's position I got had ordered 8 new ones for this January's budget------- I cancelled the order. 
What did I order you asked?  Chargers.

Ha!  Little did I know the Captain cancelled that order and orded 8 Imapalas.  We received confirmation from the dealership 3 of them are in.  I just found out a little while ago when the dealer called and now I'm going to have to ask the captain why he did that and if he bothered to cancel or change the car specific equipment we ordered. 

[rohan]

Yet another person killed in CVPI fire- this time a drunk woman under arrest in the back seat.

By ARTHUR H. ROTSTEIN
Associated Press Writer


TUCSON, Ariz. --

A woman who had been arrested and placed in the back of a police cruiser was fatally burned Wednesday when another car slammed into the parked cruiser, igniting both vehicles and another cruiser in front, authorities said.

It was the latest death in the past decade attributed to a fuel tank rupture in a Ford Motor Co.-made Crown Victoria during a rear-end collision.

Lt. James Warriner, an Arizona Department of Public Safety spokesman, said investigators are looking into why the gas tank exploded this time since both Ford Crown Victoria Police Interceptors at the scene were equipped with fire suppression systems that deployed.

An officer stopped the 45-year-old motorist at about 1:15 a.m., arrested her and put her in the back of a cruiser along Interstate 10 northwest of Tucson, Warriner said.

Another westbound car traveling at least 65 mph then drifted into the emergency lane and plowed into the back of that vehicle, Warriner said.

He said the cruiser ignited instantly, and the officers were unable to rescue the woman inside because of the intensity and heat of the fire.

The impact pushed the first cruiser's trunk into its back seat, and rammed the car into the patrol cruiser in front, severely damaging its rear end as well. The second cruiser continued the chain reaction, striking the woman's parked vehicle.

The officers were able to remove the driver of the moving car, which vaulted over a guard rail and sustained fire and crash damage, Warriner said.

He said the 28-year-old driver was taken to a hospital with undetermined injuries.

There have been no charges, with the investigation continuing, Warriner said.

Over the past decade, Crown Victorias have been the subject of significant controversy and lawsuits over numerous rear-end collisions that ruptured fuel tanks and caused horrific fires, killing or disfiguring a number of police officers and others. Since 2005, Ford has installed a powder-emitting device designed to suppress fire in its Crown Victorias.



--------------------------------------------------------------------------------

http://officer.com/web/online/Top-News-Stories/Car-Hits-Arizona-Cruiser--Kills-Suspect/1$46973

[TurboDan]

So with Chrysler seemingly going down the tubes, I guess a lot of departments won't be so keen on buying more Chargers. Several departments around here have added a number of them to the fleet. I guess the CVPI will continue to be the vehicle of choice - albeit the only choice. Also, most of the police SUVs around here are Durangos, so perhaps Explorers will get more of that market.

[Madman]

So with Chrysler seemingly going down the tubes, I guess a lot of departments won't be so keen on buying more Chargers. Several departments around here have added a number of them to the fleet. I guess the CVPI will continue to be the vehicle of choice - albeit the only choice. Also, most of the police SUVs around here are Durangos, so perhaps Explorers will get more of that market.

Most of the local police cars I see in my town are Chevrolet Impalas.  It's only the State Troopers I see in CVPIs.

Cheers,
Madman of the People

[TurboDan]

Most of the local police cars I see in my town are Chevrolet Impalas.  It's only the State Troopers I see in CVPIs.

Cheers,
Madman of the People

Wow, really?  A couple of departments used the previous generation Impalas, but I haven't seen a single one using the current model. In fact, the only town I know that had an entirely Impala fleet has switched to entirely CVPI in the last year. Honestly, I didn't even know Chevy was still outfitting them for police work.

[hounddog]

In fairness, the Impala gets substantially better gas mileage in town than does a Crown Vic.

Too bad about the lady.  Simple drunk driving should never end in a death sentence.

I wish Ford would just man up and deal with the issue.  Either delete the car line and find a suitable replacement, or move the damn tank. 

[ifcar]

Wow, really?  A couple of departments used the previous generation Impalas, but I haven't seen a single one using the current model. In fact, the only town I know that had an entirely Impala fleet has switched to entirely CVPI in the last year. Honestly, I didn't even know Chevy was still outfitting them for police work.

There are a lot of them in DC and in many of the surrounding jurisdictions.