Ads |
#42
|
|||
|
|||
Bicycle statistics
On Monday, June 3, 2019 at 9:42:12 AM UTC-7, Frank Krygowski wrote:
On 6/3/2019 8:38 AM, Duane wrote: On 03/06/2019 7:05 a.m., John B. Slocomb wrote: On Mon, 3 Jun 2019 00:25:36 -0700, sms wrote: On 6/2/2019 8:56 PM, John B. Slocomb wrote: snip As I've said before, I think it's often forgotten that medical treatments have gotten much more effective. I suspect the drop in bike fatalities - and the _greater_* drop in pedestrian fatalities - is due in large part to better medical care. You might well be correct. Except of course that pedestrian and bicycle fatalities haven't dropped, at least not in the U.S.. So it's a bit difficult to attribute better medical care to something that didn't actually happen though I guess it's possible to claim that without better medical care the numbers would be even worse. "Pedestrian Deaths Reach Highest Level In Decades, Report Says" https://www.npr.org/2019/02/28/69919...each-hignearby surveillance camerahest-level-in-decades-report-says "Increased use of smartphones and the popularity of SUVs are among the likely factors to have caused pedestrian fatalities to jump 35 percent, the Governors Highway Safety Association says." Better medical treatment doesn't trump distracted driving or texting while walking. It's the same issue with bicycling. "According to the League of American Bicyclists, more cyclists died on U.S. roads in 2016 than at any other time in the past quarter-century. But that doesn't show the whole picture." https://www.outsideonline.com/2390525/bike-commuter-deaths Yes, that seems correct in that in 2016 some 840 cyclists died and in 1991 some 842 died, but what they don't say is that during that period from 1991 until 2016, the previous quarter century, in 24 of those years the death rate was lower than in 2016 and in 2017 the death rate was lower than in 2016. It is called "Cherry Picking" and the Wiki describes it as "the act of pointing to individual cases or data that seem to confirm a particular position while ignoring a significant portion of related cases or data that may contradict that position." The information regarding bicycle deaths is freely available on the Internet so I find it surprising that you didn't .research the facts, even a little bit, before trumpeting your cries of doom. -- Sure. But the more you look at "facts" the more you realize (or should realize) that cycling deaths are likely random.* Given that when dealing with statistical analysis of cycling accidents, deaths appear to be outliers, this is not surprising. We were talking specifically about fatalities, Duane. So what do you mean by "cycling deaths are likely random" or "deaths appear to be outliers"? Are you saying they're impervious to analysis, that we can't discuss them at all? It's true that biking deaths are rare. That does mean there's going to be very visible variation in the annual count. But there's clearly a long term downward trend over decades. It doesn't take advanced mathematics to spot it. See http://www.vehicularcyclist.com/fatals.html for example. Unfortunately, the data recording when the result isn't a trip to the morgue is less than adequate so people tend to use fatalities.* But this is at best statistically misleading.* You end up with nonsense like cycling is more dangerous than sky diving.* Or less dangerous than gardening. Damn, you really hate data, don't you? I think his complaint is the lack of data in non-fatality cases. I fractured my hand in a bicycle accident and went to an urgent care clinic operated by the same clinic that provides my primary medical care. I whacked my head, too, but I wasn't complaining of a scalp wound prevented by my helmet. And my treatment would not be part of the Oregon injury data set in any event since I was not hospitalized. https://www.oregon.gov/oha/PH/DISEAS...regon_v2.3.pdf. I would also not be in any of the ER data sets. Actually, all my bicycle-related injuries, including one that got me a CT scan and plastic surgery on my face probably would not be in any Oregon data set, but then again, I haven't done a comprehensive check of the reporting regulations. Without getting into the question of whether bicycle is safe or safer than gardening, one can argue about the completeness of the data -- at least in non-fatalities. And then one can argue about what the data means. -- Jay Beattie. |
#43
|
|||
|
|||
Bicycle statistics
On 6/3/2019 10:16 AM, jbeattie wrote:
snip Without getting into the question of whether bicycle is safe or safer than gardening, one can argue about the completeness of the data -- at least in non-fatalities. And then one can argue about what the data means. What?! You want to base decisions on actual data. That's so lawyer-like! Don't you know that using actual data is not acceptable when it conflicts with an agenda? It's so interesting when someone proclaims something that isn't true, without any supporting data, then gets upset when others provide the actual data. No one would think any worse of Frank if he simply admitted his mistake. "As cars have become safer for occupants (due to airbags, structural crashworthiness and other improvements) the percent of pedestrian fatalities as a percent of total motor vehicle fatalities steadily increased from 11% in 2004 to 15% in 2014 according to NHTSA data." "Inside Versus Outside the Vehicle The proportion of people killed “inside the vehicle” (passenger car, light truck, large truck, bus, and other vehicle occupants) declined from a high of 80 percent in 1996 to 67 percent in 2017, as seen in Figure 4. Conversely, the proportion of people killed “outside the vehicle” (motorcyclists, pedestrians, pedalcyclists, and other nonoccupants) increased from a low of 20 percent in 1996 to a high of 33 percent in 2017." The other issue is that whether you're a driver, cyclist, or pedestrian, the data doesn't tell the whole story because you can take steps to change your own personal odds, and governments can take steps to change the odds as well. A cyclist can take steps like increasing their conspicuousness, using protective equipment, learning how to ride in traffic, and by choosing routes that are less risky. Government can take steps by designing transportation networks that increase safety for all road users. |
#44
|
|||
|
|||
Bicycle statistics
On 03/06/2019 1:16 p.m., jbeattie wrote:
On Monday, June 3, 2019 at 9:42:12 AM UTC-7, Frank Krygowski wrote: On 6/3/2019 8:38 AM, Duane wrote: On 03/06/2019 7:05 a.m., John B. Slocomb wrote: On Mon, 3 Jun 2019 00:25:36 -0700, sms wrote: On 6/2/2019 8:56 PM, John B. Slocomb wrote: snip As I've said before, I think it's often forgotten that medical treatments have gotten much more effective. I suspect the drop in bike fatalities - and the _greater_* drop in pedestrian fatalities - is due in large part to better medical care. You might well be correct. Except of course that pedestrian and bicycle fatalities haven't dropped, at least not in the U.S.. So it's a bit difficult to attribute better medical care to something that didn't actually happen though I guess it's possible to claim that without better medical care the numbers would be even worse. "Pedestrian Deaths Reach Highest Level In Decades, Report Says" https://www.npr.org/2019/02/28/69919...each-hignearby surveillance camerahest-level-in-decades-report-says "Increased use of smartphones and the popularity of SUVs are among the likely factors to have caused pedestrian fatalities to jump 35 percent, the Governors Highway Safety Association says." Better medical treatment doesn't trump distracted driving or texting while walking. It's the same issue with bicycling. "According to the League of American Bicyclists, more cyclists died on U.S. roads in 2016 than at any other time in the past quarter-century. But that doesn't show the whole picture." https://www.outsideonline.com/2390525/bike-commuter-deaths Yes, that seems correct in that in 2016 some 840 cyclists died and in 1991 some 842 died, but what they don't say is that during that period from 1991 until 2016, the previous quarter century, in 24 of those years the death rate was lower than in 2016 and in 2017 the death rate was lower than in 2016. It is called "Cherry Picking" and the Wiki describes it as "the act of pointing to individual cases or data that seem to confirm a particular position while ignoring a significant portion of related cases or data that may contradict that position." The information regarding bicycle deaths is freely available on the Internet so I find it surprising that you didn't .research the facts, even a little bit, before trumpeting your cries of doom. -- Sure. But the more you look at "facts" the more you realize (or should realize) that cycling deaths are likely random.* Given that when dealing with statistical analysis of cycling accidents, deaths appear to be outliers, this is not surprising. We were talking specifically about fatalities, Duane. So what do you mean by "cycling deaths are likely random" or "deaths appear to be outliers"? Are you saying they're impervious to analysis, that we can't discuss them at all? It's true that biking deaths are rare. That does mean there's going to be very visible variation in the annual count. But there's clearly a long term downward trend over decades. It doesn't take advanced mathematics to spot it. See http://www.vehicularcyclist.com/fatals.html for example. Unfortunately, the data recording when the result isn't a trip to the morgue is less than adequate so people tend to use fatalities.* But this is at best statistically misleading.* You end up with nonsense like cycling is more dangerous than sky diving.* Or less dangerous than gardening. Damn, you really hate data, don't you? I think his complaint is the lack of data in non-fatality cases. I fractured my hand in a bicycle accident and went to an urgent care clinic operated by the same clinic that provides my primary medical care. I whacked my head, too, but I wasn't complaining of a scalp wound prevented by my helmet. And my treatment would not be part of the Oregon injury data set in any event since I was not hospitalized. https://www.oregon.gov/oha/PH/DISEAS...regon_v2.3.pdf. I would also not be in any of the ER data sets. Actually, all my bicycle-related injuries, including one that got me a CT scan and plastic surgery on my face probably would not be in any Oregon data set, but then again, I haven't done a comprehensive check of the reporting regulations. Without getting into the question of whether bicycle is safe or safer than gardening, one can argue about the completeness of the data -- at least in non-fatalities. And then one can argue about what the data means. My complaint is using incomplete data incorrectly and then justifying that by saying it's the only data we have so we have to make do. But yes, it's the lack of data in non fatal cases that make up the vast majority of samples. These are concepts taught in STATS 101. |
#45
|
|||
|
|||
Bicycle statistics
On Monday, June 3, 2019 at 2:15:09 PM UTC-4, sms wrote:
On 6/3/2019 10:16 AM, jbeattie wrote: snip Without getting into the question of whether bicycle is safe or safer than gardening, one can argue about the completeness of the data -- at least in non-fatalities. And then one can argue about what the data means. What?! You want to base decisions on actual data. That's so lawyer-like! Don't you know that using actual data is not acceptable when it conflicts with an agenda? It's so interesting when someone proclaims something that isn't true, without any supporting data, then gets upset when others provide the actual data. No one would think any worse of Frank if he simply admitted his mistake. Snipped Talk about the kettle calling the pot black. You consistently ignore data whenever it does not support your stance or agenda. Cheers |
#46
|
|||
|
|||
Bicycle statistics
On 6/3/2019 11:18 AM, Duane wrote:
snip My complaint is using incomplete data incorrectly and then justifying that by saying it's the only data we have so we have to make do.* But yes, it's the lack of data in non fatal cases that make up the vast majority of samples. These are concepts taught in STATS 101. It's not just using incomplete data incorrectly it's also using complete data incorrectly. For example, carefully choosing the time-span of data to mislead people about trends is something we've seen occur in r.b.t. on many occasions. Selectively choose your dates and you can "prove" that cycling rates per-capita went down, when they really have trended up. We just saw this sort of thing happen here--you get three guesses as who did this, and the first two don't count! You can also design charts and graphs in a way that is highly misleading. Also look at the source of the data. Is it from an organization or company with a specific agenda, or is the data reliable? An anti-helmet organization is going to carefully pick and choose their data to try to advance their position, we've seen this happening in r.b.t. for many years by "he who must not be named." And of course anyone is free to make statements not supported by any data at all. We just saw this: "I suspect the drop in bike fatalities - and the _greater_ drop in pedestrian fatalities - is due in large part to better medical care." Of course the reality is that neither bike nor pedestrian fatalities actually fell. So someone postulates a reason for something that didn't actually happen, but phrases it in a way that is intended to mislead the reader into accepting that the premise is actually true. |
#47
|
|||
|
|||
Bicycle statistics
On 03/06/2019 3:03 p.m., sms wrote:
On 6/3/2019 11:18 AM, Duane wrote: snip My complaint is using incomplete data incorrectly and then justifying that by saying it's the only data we have so we have to make do.* But yes, it's the lack of data in non fatal cases that make up the vast majority of samples. These are concepts taught in STATS 101. It's not just using incomplete data incorrectly it's also using complete data incorrectly. For example, carefully choosing the time-span of data to mislead people about trends is something we've seen occur in r.b.t. on many occasions. Selectively choose your dates and you can "prove" Most people would refer to that as incomplete data. that cycling rates per-capita went down, when they really have trended up. We just saw this sort of thing happen here--you get three guesses as who did this, and the first two don't count! You can also design charts and graphs in a way that is highly misleading. Also look at the source of the data. Is it from an organization or company with a specific agenda, or is the data reliable? An anti-helmet organization is going to carefully pick and choose their data to try to advance their position, we've seen this happening in r.b.t. for many years by "he who must not be named." And of course anyone is free to make statements not supported by any data at all. We just saw this: "I suspect the drop in bike fatalities - and the _greater_* drop in pedestrian fatalities - is due in large part to better medical care." Of course the reality is that neither bike nor pedestrian fatalities actually fell. So someone postulates a reason for something that didn't actually happen, but phrases it in a way that is intended to mislead the reader into accepting that the premise is actually true. |
#48
|
|||
|
|||
Bicycle statistics
On 6/3/2019 2:15 PM, sms wrote:
On 6/3/2019 10:16 AM, jbeattie wrote: snip Without getting into the question of whether bicycle is safe or safer than gardening, one can argue about the completeness of the data -- at least in non-fatalities. And then one can argue about what the data means. What?! You want to base decisions on actual data. That's so lawyer-like! Don't you know that using actual data is not acceptable when it conflicts with an agenda? Such bull****. I'm the one who repeatedly posts data, links, etc. As I did just a few hours ago. Here it is again: http://www.ohiobike.org/images/pdfs/...gIsSafeTLK.pdf -- - Frank Krygowski |
#49
|
|||
|
|||
Bicycle statistics
On 6/3/2019 12:07 PM, Duane wrote:
snip Most people would refer to that as incomplete data. Perhaps, but I would distinguish between making conclusions based on data that is presented as incomplete, or is obviously incomplete when published, and making conclusions based on data that is complete when published but where false conclusions are drawn based on using that data in a misleading way. Of course in r.b.t. we see both of these happening. I also see this on an almost daily basis as an elected official, but fortunately I live in a city with a highly educated citizenry who are unlikely to be taken in by this sort of misuse of data. In some cases perfect data isn't available and never will be. There is simply not going to be a double blind study on every possible subject in the world--in some cases it's not possible and in some cases when it is possible there will be no one interested enough to fund such a study. If a thousand ER doctors tell you that helmeted cyclists fare better in head-impact crashes than unhelmeted cyclists then you're probably going to believe them over someone who insists that helmets are worthless. In both cases there is incomplete data, but in one case there is credibility of those making the statements. If a police captain explains to you that you're better off making yourself more conspicuous while bicycling then you're probably going to believe him or her versus someone that insists that being more conspicuous is of no value, even though the data to prove this is incomplete. |
#50
|
|||
|
|||
Bicycle statistics
sms writes:
On 6/3/2019 10:16 AM, jbeattie wrote: snip Without getting into the question of whether bicycle is safe or safer than gardening, one can argue about the completeness of the data -- at least in non-fatalities. And then one can argue about what the data means. What?! You want to base decisions on actual data. That's so lawyer-like! Don't you know that using actual data is not acceptable when it conflicts with an agenda? It's so interesting when someone proclaims something that isn't true, without any supporting data, then gets upset when others provide the actual data. No one would think any worse of Frank if he simply admitted his mistake. "As cars have become safer for occupants (due to airbags, structural crashworthiness and other improvements) the percent of pedestrian fatalities as a percent of total motor vehicle fatalities steadily increased from 11% in 2004 to 15% in 2014 according to NHTSA data." It seems that from 2004 to 2014 total motor vehicle fatalities declined from 14.59 per 100K population to 10.28 per 100K population, or almost 30%. So I don't see the evidence that Frank is wrong. It's entirely possible for total pedestrian fatalities to decrease at the same time as their proportion of total motor vehicle fatalities increases. The trend in all motor vehicle fatalities over the past 20 years or so is down, perhaps largely due to better emergency treatment. In the past 4-5 years there has been an uptick, which many attribute to cell phone use. Standards for motor vehicle occupant protection have improved over the same period, at some expense in outward visibility; one might expect this to increase the proportion of "outside the vehicle" fatalities. Wikipedia has a nice collection of data: https://en.wikipedia.org/wiki/Motor_...n_U.S._by_year "Inside Versus Outside the Vehicle The proportion of people killed “inside the vehicle” (passenger car, light truck, large truck, bus, and other vehicle occupants) declined from a high of 80 percent in 1996 to 67 percent in 2017, as seen in Figure 4. Conversely, the proportion of people killed “outside the vehicle” (motorcyclists, pedestrians, pedalcyclists, and other nonoccupants) increased from a low of 20 percent in 1996 to a high of 33 percent in 2017." The other issue is that whether you're a driver, cyclist, or pedestrian, the data doesn't tell the whole story because you can take steps to change your own personal odds, and governments can take steps to change the odds as well. A cyclist can take steps like increasing their conspicuousness, using protective equipment, learning how to ride in traffic, and by choosing routes that are less risky. Government can take steps by designing transportation networks that increase safety for all road users. -- |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
accident statistics: car vs motorcycle vs bicycle per mile travelled? | [email protected] | General | 15 | June 11th 08 03:27 AM |
Bridge Statistics | _[_2_] | UK | 7 | September 10th 07 02:47 PM |
Bridge Statistics | _[_2_] | UK | 4 | September 4th 07 11:01 PM |
Where are those statistics? | bob | UK | 15 | August 30th 07 12:31 PM |
Bicycle Injury Statistics | [email protected] | General | 8 | August 1st 06 07:33 AM |