First, I want to thank LittleBigMan for starting this thread, and this discussion. It got me to read the above reference, Adult Bicyclists in the United States, Characteristics and Riding Experience in 1996
by William E. Moritz. I think that I was a part of that study, actually, as it was a time when I was a member of the League of American Bicyclists. But let me make a few comments on this survey.
First, and foremost, it is a survey, and not a "study." This survey was very limited in scope, in that it replicated two previous surveys (a Washington State survey in 1994, and a League of American Wheelmen survey in 1974). A "study" would include being double-blind, and include much more than one group to examine the problems. A "study" should be reproducable, but this survey is not. It is not accurate either. To quote from the report:
Please note that there are several other thing about this survey which handicap it. This is a survey of the adult League of American Bicyclists. It is not inclusive of all bicyclists, and therefore any claims for relative danger apply really only to this group. Other factors may influence other groups. Also, this is self-reported data. There is no way to check this data at all (and we know the range of errors is 80%; you know, 2% to 5% is probably acceptable, but 80%?). Finally, the author of the study decided on the criterion for a "serious" crash as one that caused at least $50 of property damage or medical expenses (few medical visits go now for $50 for anything).
Concerning the statements on sidewalks above, they were not made in the report of the survey. They were apparently made by John S. Allen to support his thesis. The abstract of the survey stated:
There are several things to be said about the data on sidewalks in the survey. First, this statement is in the abstract, but not in the survey report itself. Second, it lumps all the crashes into one category, whereas if it were broken out, you could easily see that sidewalks are less hazardous than roads without bike facilities. The statement above that riding on sidewalks is 24.8 times more hazardous than riding on a major road without bicycle facilities is based upon a table in the 1998 report of the same title presented to the Transportation research Board (I found both on a Google search). This table shows, for all crashes, the crash rate per million kilometers is 1026 for "Other (most often 'sidewalk')" and 41 for "Major w/o bike facilities." By the way, 1026 divided by 41 is 25.0, not 24.8, and neither of these numbers are used in the study itself. They were derived from the report's Table 4. Crashes by facility type.
But this is biased data. First, I doubt that League of American Bicyclists ride much on these "Other" areas, and probably ride them very fast when they do. But if you look at the data for the percentage of the total crashes, it is 5% and 4% respectively for "Serious" and "Minor" crashes on "Other (most often 'sidewalk')." It is 29% and 17% respectively for "Serious" and "Minor" crashes on "Major w/o bike facilities." And it is 41% and 43% respectively for "Serious" and "Minor" crashes on "Minor w/o bike facilities." ("Major" and "minor" in these cases meaning major and minor roads.) It should also be pointed out that the LAB 96 survey is the only one cited in this paper that has data on "Other (most often 'sidewalk')." The other two surveys (WA 94 and LAW 74) list "N/A" for this category, as they had no data on it at all.
LittleBigMan, you asked that we provide scientifically validated data to back our claims that riding sidewalks are as safe or safer than riding on the roadway. The data in the study you provide suggests that, if we use LAB data, the LAB cyclists experienced inordinent numbers of minor crashes on sidewalks compared to the number of miles (km) ridden on them. But this data is really suspect. I know that I had a hard time remembering what I had done for a year previous in my bicycling. I don't think many of us could produce data anywhere near accurate a year later about our bicyling habits (# km/type of roadway each day).
Now, from a practical aspect, I will talk about my today's ride (which I do still remember pretty well). I rode my "auto avoidance" route today, as on Thursday evening I treat autos as if it were a Friday (and I still do not bicycle on Friday--two trips to emergency rooms have quelled my enthusiam for Friday riding). I rode ten miles this evening, about 1.5 of which was on sidewalks, 0.5 on bicycle paths, and the rest on open roadway (both with and without bike lanes). I was passed by approximately 35 cars during that time on the road. My estimate is that about 10% of auto drivers are impared for some reason (cell phone usage, baby, drink, radio, alcohol, etc.). That means that probably three or four of those cars passing me were impared. One guy, in a sedan with a child next to him, talked to me about my recumbant as we were stopped side-by-side at a stop light; as the light turned, he answered a cell phone and proceded through the intersection. I headed to the sidewalk, as the next stretch of my commute was uphill, with no shoulder or bike lane whatsoever. During my commute home, I had (if I give ~3000 pounds/car) about 52.5 tons of vehicles pass me within ten feet at a regular speed (45-55 mph), and 4.5 tons were most probably driven by impared drivers. That is a lot of mass and force.
When I was on the sidewalk, I was separated by these vehicles by a curb, sometimes a light pole, and most often by about 10-20 feet of ground (sometimes with trees between me and the road). Sure, I could hit one of those poles, or the post in the middle of the bike path I took. But if I did, my mass going at 12.5 mph would be about 250 pounds. This is about 18.3 feet per second. My 250 pounds (with the bike) hitting at 12.5 mph would be 4583 foot pounds per second would create an impact, but probably not life-threatening.
If I am on a road, and hit by a car, then (assuming the car is going my direction at 45 mph, and I'm fast at 18 mph) the closing speed is (45 mph - 18 mph) about 40 feet per second. The car weighs 3000 pounds though (if I'm lucky), and so the impact will be at 118,800 foot pounds per second. I don't think I would survive a direct hit.
So I use the auto avoidance when I think they are driving in a hazardous manner (and that is on Thursday afternoon and all day Friday). If I were to hit a person (and I saw none today on the sidewalks, and two on the bike path I took), it would be at that slow speed and low force. It would hurt, but probably not kill anyone. In other words, those above who said that you were not looking correctly at the risk management are correct in their thinking. The risks simply are not as great.
I am nearing 60 years old now, and hope to be cycling to 100 years old. I will continue to use these techniques to ensure I don't tangle again with a car. Some of you who have been riding "aggressively" don't perhaps know what it feels like to awaken in an ER and have the Doc tell you he had just completed the CAT scan, but you didn't remember it. But I have had that experience, and it is not pleasant. It has made me re-think all my cycling--for instance, I ride a recumbant now, and the sidewalks are not as hazardous to me as you who don't as I'm closer to the ground, and traveling feet-first. Ever tried swimming through a river rapids head-first--that isn't pleasant either, so why do we go head first on bikes when we have cars to duck instead of rocks?
Well, the dishes becken, so I need to wrap this up. I will continue to use sidewalks when it is to my advantage (it is legal here), and bike paths too. They may hurt you, but they are very unlikely to kill you. I can heal...but not if I'm dead!
PS--I am a Certified Safety Professional, and applied a lot of my 25 years in the safety field to analyzing my bicycling.