So, Wikipedia says the Ford Pinto sold 3,173,491 units. This article says there were 27 Ford Pinto fatalities. The article also says the Cybertruck sold 34,000 units and there have been 5 fatalities.
Your point 2 aside, you’re not trying to argue that (5/34,000) / (27/3,173,491) isn’t approximately 17, right?
(Again, point 2 aside) is your point that 5 deaths (or rather 4 deaths if you don’t count the guy from point 2) out of 34,000 units isn’t a big enough sample size to draw conclusions and that you think it’s likely that as more units are sold, the rate won’t stay that high and over time the data will average out to a fatality rate less than that of the Ford Pinto?
One more question if I may. Are you a fan of Tesla?
you think it’s likely that as more units are sold, the rate won’t stay that high
That would be a prediction of the future that I’m not prepared to make, but I do suspect that will end up being the case. Assuming they ever even sell a million of them before ending production. All I will comment on is today, and the idea that the first 34k Pintos probably had a much higher fatality rate than the last 3,139,491.
Are you a fan of Tesla?
I’m a fan of rational thought processes and statistics. Something I think very few people are capable of when it comes to anything having to do with Musk. The fact that Elon’s name is even in the headline leads me to believe that’s the case here as well. When was the last time you heard about “Farley’s Ford” or “Rawlinson’s Lucid”? If you’re a “journalist”, you just can’t go wrong with anything to do with Elon.
Assuming they ever even sell a million of them before ending production.
You don’t think there’s anything magical about the number “a million”, though, correct? Just “enough to form a sufficiently good sample size” (whatever your threshold might be.)
Are you a fan of Tesla?
I’m a fan of rational thought processes and statistics…
You don’t think there’s anything magical about the number “a million”, though, correct?
Sure.
I can’t help but notice you dodged my question.
I didn’t dodge it, I chose not to engage because you’re looking to attack my character and derail the conversation rather than discussing the merit of my statements.
For those spending this much time explaining that an N=4 being statistically conclusive (LOL), perhaps your time would have been better used if you looked into how badly the pinto data was cherry picked (1626 deaths in 2.5 years):
I get paid to teach stats. So not my job here. But let me use common sense and see if that works.
Is a comparison valid? Of course it is. Connecticut has a population of about 3.5 million. Torrington CT has a population of around 35,000. Are you telling me that you can’t compare death rates in Torrington vs. the rest of Connecticut because of “statistics”?
You can compare whatever you want. But there are good comparisons and there are bad comparisons, and this is the latter.
Also we’re talking about cars that roll off an assembly line, not people. If the death rate is higher in people, do you blame the people? Another bad comparison.
What are you even talking about? Failure rates in manufacturing are governed by the same statistics rules as human errors or deaths for sufficiently large n. And 35,000 is sufficiently large n .
Jesus Christ. You really need this spelled out, don’t you?
Machines coming off an assembly line are almost completely identical, which you cannot say for humans.
We can fix errors in vehicle manufacturing very easily, which you also cannot say for humans.
You’re comparing death rates in humans across locales, which is looking for environmental variables and not biological ones. When comparing death rates among different vehicles, you’re looking for manufacturing errors.
This is a bad comparison and statistically insignificant.
1 works against your point, 2 is irrelevant or you need to expand on what you mean, 3 is misunderstanding what is being compared when you compare samples of two different pops. a population in one environment being compared to a population in another, the difference is the environment. a car population of one make being compared with a car population of a different make, the thing being compared is the manufacturing and design.
you appear to be working from a conclusion backwards that this is an invalid comparison and grasping for why.
If machines coming off an assembly line are virtually identical, then a smaller sample size can be used due to reduced variation. Larger samples are required to control for variation.
I think you guys are just blowing smoke for kicks at this point. Your stats reasoning doesn’t display even a superficial understanding.
It’s reported as the rate per 100,000 units. So that’s accounted for.
Bad commenting is bad.
the total sample is between 0 and 34,000 across the past couple years. You are mixing up sample, 34,000 is actually a remarkably large sample size.
Maybe you dont understand clustering
do enlighten me then.
I’d suggest you familiarize yourself with the concept of sample sizes.
So, Wikipedia says the Ford Pinto sold 3,173,491 units. This article says there were 27 Ford Pinto fatalities. The article also says the Cybertruck sold 34,000 units and there have been 5 fatalities.
Your point 2 aside, you’re not trying to argue that
(5/34,000) / (27/3,173,491)isn’t approximately 17, right?(Again, point 2 aside) is your point that 5 deaths (or rather 4 deaths if you don’t count the guy from point 2) out of 34,000 units isn’t a big enough sample size to draw conclusions and that you think it’s likely that as more units are sold, the rate won’t stay that high and over time the data will average out to a fatality rate less than that of the Ford Pinto?
One more question if I may. Are you a fan of Tesla?
That would be a prediction of the future that I’m not prepared to make, but I do suspect that will end up being the case. Assuming they ever even sell a million of them before ending production. All I will comment on is today, and the idea that the first 34k Pintos probably had a much higher fatality rate than the last 3,139,491.
I’m a fan of rational thought processes and statistics. Something I think very few people are capable of when it comes to anything having to do with Musk. The fact that Elon’s name is even in the headline leads me to believe that’s the case here as well. When was the last time you heard about “Farley’s Ford” or “Rawlinson’s Lucid”? If you’re a “journalist”, you just can’t go wrong with anything to do with Elon.
You don’t think there’s anything magical about the number “a million”, though, correct? Just “enough to form a sufficiently good sample size” (whatever your threshold might be.)
I can’t help but notice you dodged my question.
Sure.
I didn’t dodge it, I chose not to engage because you’re looking to attack my character and derail the conversation rather than discussing the merit of my statements.
For those spending this much time explaining that an N=4 being statistically conclusive (LOL), perhaps your time would have been better used if you looked into how badly the pinto data was cherry picked (1626 deaths in 2.5 years):
(sauce: https://www.autosafety.org/wp-content/uploads/import/ODIPinto.pdf Page 9.)
Dude. I’ve taught statistics. I don’t think you understand what you’re arguing here. lol.
Why don’t you bring me up to speed instead of levying personal attacks and then disappearing?
I get paid to teach stats. So not my job here. But let me use common sense and see if that works.
Is a comparison valid? Of course it is. Connecticut has a population of about 3.5 million. Torrington CT has a population of around 35,000. Are you telling me that you can’t compare death rates in Torrington vs. the rest of Connecticut because of “statistics”?
You can compare whatever you want. But there are good comparisons and there are bad comparisons, and this is the latter.
Also we’re talking about cars that roll off an assembly line, not people. If the death rate is higher in people, do you blame the people? Another bad comparison.
What are you even talking about? Failure rates in manufacturing are governed by the same statistics rules as human errors or deaths for sufficiently large n. And 35,000 is sufficiently large n .
It’s a valid comparison and statistically sound.
Jesus Christ. You really need this spelled out, don’t you?
Machines coming off an assembly line are almost completely identical, which you cannot say for humans.
We can fix errors in vehicle manufacturing very easily, which you also cannot say for humans.
You’re comparing death rates in humans across locales, which is looking for environmental variables and not biological ones. When comparing death rates among different vehicles, you’re looking for manufacturing errors.
This is a bad comparison and statistically insignificant.
1 works against your point, 2 is irrelevant or you need to expand on what you mean, 3 is misunderstanding what is being compared when you compare samples of two different pops. a population in one environment being compared to a population in another, the difference is the environment. a car population of one make being compared with a car population of a different make, the thing being compared is the manufacturing and design.
you appear to be working from a conclusion backwards that this is an invalid comparison and grasping for why.
If machines coming off an assembly line are virtually identical, then a smaller sample size can be used due to reduced variation. Larger samples are required to control for variation.
I think you guys are just blowing smoke for kicks at this point. Your stats reasoning doesn’t display even a superficial understanding.