Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
And I'm sure you realize that, similar to Albuquerque in the Allstate report, that puts New Mexico in the lower 50 percentile. But what's interesting about the Car Insurance Comparison report is that they included several other factors in their evaluation -- even something like "careless driving," which New Mexico scored kind of high for. I don't know how they defined "careless driving."
There's actually an argument in favor of NOT controlling for the annual miles per household for each city, and if I got this right, it's because you would be removing an important quality of the driving conditions of a particular region. If in a city like Albuquerque people drive more per year because things are spread out, despite making it more likely a person will have an accident, by statistically controlling for the miles you actually remove an important part of the actual driving conditions in this region. It's also not fair in comparison with cities where drivers do not need to drive as much per year because, conversely, the driving condition of these regions often requires less driving. When you rank the cities without controlling for mileage, you are allowing for the different conditions of the amount of driving needed for a region to be a possible reason why one city may have more accidents.
*But you have to remember that interpreting the results of a study that doesn't control for miles per household in this manner has an important limitation: the results are not necessarily a reflection of the skill of drivers.
However, if you want to control for annual miles driven per year per city, you need to create a multiplier for the number of years between accidents, and it's different for each city based on two things: 1) the average number of miles driven per year per household for all the cities in the study, and 2) the average number of miles driven per year per household for each city. You can then use this forumula:
Average annual miles driven for the city per household
_____________________________________________ = Multiplier for the city
Average annual miles per household for the study
Then, you can use this multiplier as follows:
Multiplier x (number of years between accidents reported for the city) = new number that can be used to re-rank the city. You need to do this for all 200 cities to re-rank them.
But I still say if you statistically control for mileage like this you are removing an actual driving condition of the region of the driver, which makes external validity of the study more questionable. The Allstate report has a degree of internal validity, which is why I believe the ranking is somewhat valid, but it's erroneously called "America's Best Drivers Report," and this title should have been something like "America's Best Cities for Driving," because given it's only based on accidents, with no conditions controlled for in the field (impossible to do), what they are measuring are the driving conditions, which includes the skill of drivers but also includes plenty of other factors.
No analysis of how to rank cities for driving safety can incorporate all possible factors. For example, there are different types of vehicles being driven (small cars, motorcycles, buses, SUVs, RVs, small trucks, large freight trucks), some are for personal use others are for commercial use. Commercial vehicles will have more miles driven. The quality of all these vehicles varies. The "average" quality of the vehicle, and in areas where more people have a low socio-economic status the cause of accidents can be higher because of something like mechanical failures (e.g., bad brakes). The number of cars on the road is in and of itself a factor (recall my point above about how KRQE said there are more and more cars each year). Even a factor like how much sleep someone gets can have serious consequences for driving performance.
In reference to the comment Fran made above about Albuquerque having the "worst" drivers, I think she didn't mean statistically the worst. She meant "horrible," but part of the reason drivers are horrible and accidents occur at a particular rate in the city are because of more factors than just driving skill.
The driving skill of an individual can really only be determined from an experimental design that actually controls for confounding variables.
There's actually an argument in favor of NOT controlling for the annual miles per household for each city, and if I got this right, it's because you would be removing an important quality of the driving conditions of a particular region. If in a city like Albuquerque people drive more per year because things are spread out, despite making it more likely a person will have an accident, by statistically controlling for the miles you actually remove an important part of the actual driving conditions in this region. It's also not fair in comparison with cities where drivers do not need to drive as much per year because, conversely, the driving condition of these regions often requires less driving. When you rank the cities without controlling for mileage, you are allowing for the different conditions of the amount of driving needed for a region to be a possible reason why one city may have more accidents.
Well in that case you're not measuring the quality of the drivers, which is what the "report" and the KRQE articles were talking about. You're measuring something, it may or may not be a useful metric, but it's not driver skill.
Quote:
The Allstate report has a degree of internal validity, which is why I believe the ranking is somewhat valid, but it's erroneously called "America's Best Drivers Report," and this title should have been something like "America's Best Cities for Driving," because given it's only based on accidents, with no conditions controlled for in the field (impossible to do), what they are measuring are the driving conditions, which includes the skill of drivers but also includes plenty of other factors.
Still doesn't consider the idea that people who use Allstate insurance might be different from people who use other insurance companies or have no insurance.
Quote:
No analysis of how to rank cities for driving safety can incorporate all possible factors. For example, there are different types of vehicles being driven (small cars, motorcycles, buses, SUVs, RVs, small trucks, large freight trucks), some are for personal use others are for commercial use. Commercial vehicles will have more miles driven.
Yup. So why so defensive of the Allstate report which clearly has limited validity/usefulness?
Quote:
In reference to the comment Fran made above about Albuquerque having the "worst" drivers, I think she didn't mean statistically the worst. She meant "horrible," but part of the reason drivers are horrible and accidents occur at a particular rate in the city are because of more factors than just driving skill.
I'll leave that up to Fran to clarify.
EDIT: and thank you for agreeing that driver skill is only one of many factors that influences your mean time between accidents.
Last edited by funkymonkey; 09-12-2014 at 11:01 PM..
And I'm sure you realize that, similar to Albuquerque in the Allstate report, that puts New Mexico in the lower 50 percentile. But what's interesting about the Car Insurance Comparison report is that they included several other factors in their evaluation -- even something like "careless driving," which New Mexico scored kind of high for. I don't know how they defined "careless driving."
60th percentile, middle quintile. Makes us about typical in my estimation. The rank of 30th on the worst drivers list means we're better than 29 states by their methods. So we're in the "lower 50th percentile" for "worst drivers" which is a good thing.
The rank of 30th on the worst drivers list means we're better than 29 states by their methods. So we're in the "lower 50th percentile" for "worst drivers" which is a good thing.
I read the table you see for the Car Insurance Comparison’s (CIC) data as if it was laid out for the Allstate report. They have the ranks in opposite order from worst (number 1) to best (number 51). The state of New Mexico scores better here than what the city of Albuquerque scored on the Allstate report.
*But "we" are not the whole state of New Mexico, and it confounds the issue presented here to evaluate Albuquerque drivers based on the CIC report. The Allstate report and KRQE news article are in reference to Albuquerque, not the whole state. If you include the whole state of New Mexico, you'll likely get results about drivers from cities that do not have nearly as many drivers on the road compared to Albuquerque. So this report has little to do with evaluating Albuquerque as a city to rank with 200 other cities.
So why so defensive of the Allstate report which clearly has limited validity/usefulness?
I'm not. I reacted to your calling the report "absurd." I said above in post #7 "I agree the Allstate report does have limitations in interpreting results, but it's not meaningless." It took more explanation to show if you look past the title of the report and consider it's about general driving conditions, not only driver skill, Albuquerque is approaching an "unsafe" city compared to the others in the study. I'm suggesting not to throw the baby out with the bathwater.
What's with your "There's actually an argument in favor of NOT controlling for the annual miles per household for each city"
That is absurd... No one but you brought that up. Do you think anyone cares about your calculations in the following post?
Your following post has nothing to do with Albuquerque.
Quote:
Originally Posted by Aero
There's actually an argument in favor of NOT controlling for the annual miles per household for each city, and if I got this right, it's because you would be removing an important quality of the driving conditions of a particular region. If in a city like Albuquerque people drive more per year because things are spread out, despite making it more likely a person will have an accident, by statistically controlling for the miles you actually remove an important part of the actual driving conditions in this region. It's also not fair in comparison with cities where drivers do not need to drive as much per year because, conversely, the driving condition of these regions often requires less driving. When you rank the cities without controlling for mileage, you are allowing for the different conditions of the amount of driving needed for a region to be a possible reason why one city may have more accidents.
*But you have to remember that interpreting the results of a study that doesn't control for miles per household in this manner has an important limitation: the results are not necessarily a reflection of the skill of drivers.
However, if you want to control for annual miles driven per year per city, you need to create a multiplier for the number of years between accidents, and it's different for each city based on two things: 1) the average number of miles driven per year per household for all the cities in the study, and 2) the average number of miles driven per year per household for each city. You can then use this forumula:
Average annual miles driven for the city per household
_____________________________________________ = Multiplier for the city
Average annual miles per household for the study
Then, you can use this multiplier as follows:
Multiplier x (number of years between accidents reported for the city) = new number that can be used to re-rank the city. You need to do this for all 200 cities to re-rank them.
But I still say if you statistically control for mileage like this you are removing an actual driving condition of the region of the driver, which makes external validity of the study more questionable. The Allstate report has a degree of internal validity, which is why I believe the ranking is somewhat valid, but it's erroneously called "America's Best Drivers Report," and this title should have been something like "America's Best Cities for Driving," because given it's only based on accidents, with no conditions controlled for in the field (impossible to do), what they are measuring are the driving conditions, which includes the skill of drivers but also includes plenty of other factors.
No analysis of how to rank cities for driving safety can incorporate all possible factors. For example, there are different types of vehicles being driven (small cars, motorcycles, buses, SUVs, RVs, small trucks, large freight trucks), some are for personal use others are for commercial use. Commercial vehicles will have more miles driven. The quality of all these vehicles varies. The "average" quality of the vehicle, and in areas where more people have a low socio-economic status the cause of accidents can be higher because of something like mechanical failures (e.g., bad brakes). The number of cars on the road is in and of itself a factor (recall my point above about how KRQE said there are more and more cars each year). Even a factor like how much sleep someone gets can have serious consequences for driving performance.
In reference to the comment Fran made above about Albuquerque having the "worst" drivers, I think she didn't mean statistically the worst. She meant "horrible," but part of the reason drivers are horrible and accidents occur at a particular rate in the city are because of more factors than just driving skill.
The driving skill of an individual can really only be determined from an experimental design that actually controls for confounding variables.
if you look past the title of the report and consider it's about general driving conditions, not only driver skill, Albuquerque is approaching an "unsafe" city compared to the others in the study. I'm suggesting not to throw the baby out with the bathwater.
Except Allstate Drivers are not a random sample of drivers in Albuquerque. In no way, shape or form can you say anything about driving conditions in Albuquerque vs other cities by looking at customers of one particular insurance company. Customers of specific insurance companies are not random. They are self selecting. It is a snapshot of people who are drawn to Allstate's marketing and pricing scheme.
Would you take a survey of Albuquerqueans who shop at Walmart and try to apply their responses to everyone in Albuquerque? Would you take a survey of Albuquerqueans who buy all their kitchen gadgets at William Sonoma and conclude that the average Albuquerquean is willing to spend $600 on a blender and $200 on a frying pan?
Different insurance companies serve different market segments. You can't use data from one company's customers to make conclusions about general driving conditions in Albuquerque or anywhere else.
I think it's a good idea. The more bicyclists the better.
Cycling is a healthier and cheaper way of getting around. Bicycles cost markedly less than cars, and use NO gasoline. Plus, one will burn fat, provided one eats right.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
