THE TRUTH ABOUT SEAT BELTS 
The present report shows, with more than 150 pages of documentation, much of it taken from NHTSA's own files, that not only do seatbelts not save lives but that wearing a seatbelt increases the chance of being killed or severely injured in an automobile accident. 
An anaysis of the FARS data for 2001 shows that 54% of automobile occupants killed in head-on collisions were wearing seatbelts at the time. (ftp://ftp.nhtsa.dot.gov/FARS/) No evidence is presented to indicate that if the the remainder had been wearing seatbelts they would not also have been killed. Nor has the government presented any evidence that seatbelts save lives in any other type of collision. To force 200,000,000 people to wear seatbelts at a cost of a billion dollars a year without credible evidence would be a serious enough matter. But to show that there is no evidence that seatbelts save lives does not, in itself, prove that they are dangerous. In order to do this, we have to look at both the physical and statistical evidence. 
	a)The physical evidence. The original justification put forward by seatbelt proponents was that seatbelts would save lives in head-on collisions by preventing occupants from being "thrown through the windshield". According to the FARS data base for 2001, out of a total of 36,281 vehicle occupants killed that year in traffic accidents, 144 were killed by being "ejected through the windshield". The total number "ejected through the windshield" is given as 182. If we include "partial ejections", the total killed is given as 205. Even in this small percentage of cases, there is no evidence that wearing a seatbelt would have saved these people, due to the injuries caused by the seatbelt itself (see below). If we look at all head-on collisions, we find that, according to NHTSA's own figures (table 32, p9), head-on collisions account for only 1.8% of all collisions and only 13.7% of fatal collisions. An inspection of available accident reports indicates that most of those killed in head-on collisions are killed as a result of being crushed, not as a result of being thrown through the windshield (p26, case 10.122; p27, case 10.174; p35, case 10.41; p40, case 10.20;) An analysis of FARS data, combined with photographs and accident reports, indicates that approximately 83% of those killed in head-on collisions were killed as a result of being crushed. Moreover, there are  many recorded cases of head-on collisions where those wearing a seatbelt were killed while those not wearing a seatbelt survived (p26, case 10.122; p.35, case10.41; p.38, cases IA-6 and WI2; p39, cases 10.286 and 10.15.1; p40, case 10.20; p.47, 5th case; p.43, 8th case; p44, 13th case). In addition, there are numerous recorded cases of occupants being killed by internal injuries caused by the seatbelt or killed by fire, or drowning, after a crash as a result of not being able to open their seatbelts.(pp 25-49). Nor does this figure take into account the decreased mobility caused by the seatbelt which makes it harder for a driver to avoid a collision in the first place. These examples do not represent exceptional or unusual cases. 
One of the principal fallacies of the seatbelt advocates is their failure to understand Newton's Second Law of Motion. Newton's Law states that the force is equal to the mass times the rate of change of the velocity in the direction of the rate of change. A person weighing 150 pounds and wearing a seatbelt, for example, coming to a stop in one second from 44 miles per hour, would experience a force of 300 pounds from the seatbelt. If the person came to a stop from 30 miles per hour in one foot, as shown in the NHTSA dummy tests, the force of the seatbelt on the person would be 4,509 pounds for .045 seconds, the time it takes for the person to stop. 
	A 4,000 pound vehicle coming to a stop in one second from a speed of 44 mph would exert, and experience, a force of 8,000 pounds. If the vehicle came to a stop in one foot, the force on the vehicle would be 258,133 pounds for .031 seconds. The effect on most cars is like stomping on an empty tin can. We know from a study of actual accidents that when the impact speed is 45 miles per hour or higher, the dashboard in most cars is driven all the way back to the back of the front seat, crushing the front seat occupants to death. A study of the FARS data indicates that the average fatal head-on collision takes place at 55 miles per hour. In the case of persons being killed by being "ejected through the windshield", the average speed for 2001 was 66 miles per hour. 

When their first rationale for seatbelts was questioned, seatbelt proponents then asserted that seatbelts would save lives in a roll-over accident. But it is a fact that this assertion is false because in roll-over accidents occupants are killed by being crushed when the roof caves in, and this is true in almost all standard production automobiles and light pickup trucks. (p25, case 10.155; p28, cases 10.179 and 10.219; p36, case 10.97). The only hope of survival an occupant has in such a case is to either duck down, jump out, or be "ejected" (see below), all of which are made more difficult by wearing a seatbelt. It is obvious, therefore, that in the case of rollover accidents, wearing a seatbelt increases the chance of being killed. An analysis of the FARS data, combined with photographs and accident reports, shows that, in fatal roll-over accidents, almost 100% of all vehicles involved are severely crushed. 
Seatbelt proponents then asserted that seatbelts would save lives by preventing occupants from being "ejected" in a crash. We have already seen how, in the case of head-on collisions, such claims are highly questionable. For a person to be "ejected" from a motor vehicle  requires an opening for the person to be "ejected" through. An analysis of the force vectors involved in various types of collisions indicates that it is highly unlikely that a person would be "ejected" though a side or rear window, making such claims highly questionable. That leaves only cases of "ejection" through open doors or roofs, which, according to the government's own data, are quite rare. Moreover, the preponderence of the evidence is that a person prevented from being "ejected" by a seatbelt in such cases would be as likely, or even more likely, to be killed. For a person to be "ejected" from a motor vehicle by a collision requires a crash of tremendous force.  In crashes of such intensity the vehicle is likely to be badly crushed or even torn apart, as we know from actual accident reports (pp 25-49). If a person is wearing a seatbelt in such a crash, the force exerted by the seatbelt on the body is likely to result in severe injury or death, even if the person is not crushed by the vehicle. Verified cases of "ejection" are actually quite rare (see 
p. 7). NHTSA now coined the term "partial ejection". But the term "partial ejection" is an oxymoron. If a person is still partially in the vehicle he has not been "ejected". Nevertheless, as we see in table 72, p24, NHTSA now began to lump "partial ejections" and "total ejections" together under "ejections". By this definition, a person who's hand is sticking out the car window has been "ejected". (Notice that it does not say that the people were killed as a result of the "ejection") According to NHTSA document Inf GR/ DL/3/2 (p66) the number of people killed each year by "ejections" through door openings in non roll-over accidents constitute only about 2.5% of fatalities. (Even this is doubtful. See note (2), p. 7).As we have seen in the previous paragraph, in a rollover accident a person's chances of survival would be improved by being "ejected". Moreover, some people are saved as a result of "ejections" in other types of accidents (p26, case10.112; p27, case10.174; p35, case 10.34; p38, case WI-2) while many others are killed as a result of being trapped in the car by the seatbelt or by internal injuries caused by the seatbelt (pp25-49). So there is no evidence that seatbelts constitute a net benefit by preventing "ejections". 
Finally, and most recently, seatbelt proponents have put forward the argument that seatbelts would save lives because they would "better enable the driver to control the vehicle". In fact, the opposite is true. Seatbelts restrict the driver's mobility in an emergency and thus make it more difficult for him to avoid an injury accident. 
In any analysis of the effects of wearing a seatbelt in a crash, the following factors must be considered. We may summarize the effects under 5 headings: 
1. Effect of being trapped in the car after a crash as a result of not being able to open the seatbelt. Numerous persons have been burned to death or drowned by this effect. (pp25-49). This is not primarily the result of defective seatbelts. It frequently happens as a result of a crash that and occupant cannot open his seatbelt because a)he is pinned in the wreck and cannot reach the buckle, b)his arms, hands or wrists are broken c)the car and/or the occupant is on fire and the buckle is too hot to touch or d)the buckle itself has been damaged or twisted in such a way that it cannot be opened. 
2. Fatal injuries caused by the belt in a crash, including broken spleens, burst intestines, broken necks, slit throats, decapitation, crushed chest, crushed pelvis and broken spine. (cases 10.140, 10.154, p25; cases 10.169(1) and 10.169(2), p27; cases 8(2) and 9, p32; case 10.99, p36; cases 10.286 and 10.15.1, p39; p46, 3 cases; p49, 3 cases; p43, 2 cases; p44, 4 cases; p45, 1 case; p46, 4 cases). 
3. In roll-over accidents,  a seatbelt assembly makes it more difficult tp avoid being crushed as the roof caves in. In side impact collisions, which consitute 33% of all passenger car fatalities (Table 71, p.17 ) (p26, case 10.108, p35, case 10.68, p.38, NJ2) the occupant is prevented from sliding, or being pushed away from the impact thus greatly increasing his injuries and his chance of being killed. It also frequently happens that one side of the vehicle is crushed more than the other. The seatbelt prevents the occupant from taking evasive action. The same is true for objects coming through the windows or windshield (p39, case 5-12-77). The severe crush rate for fatal side impact collisions is almost 100%. 
4. Decreased driver mobility. A seatbelt restricts a driver's mobility in an emergency, making an injury accident more likely. This is consistent with the statistical evidence which shows that a higher percentage of belted drivers are involved in injury accident than unbelted ones. (See also section VIII, page 6). 
5. In rear-end collisions the whiplash effect caused as a result of wearing a seatbelt has been well documented (p104) and can cause severe, although not necessarily fatal,  neck injuries. But there are documented cases of people being killed by seatbelts even in rear-end collisions (p36, case 10.79, p46, last case). 

The physical evidence shows that wearing a seatbelt does much harm and little good. There is, of course, no proof that seatbelts save any lives at all. The preponderence of the physical evidence is that wearing a seatbelt increases the danger of being killed or severely injured in a motor vehicle accident. 
b) The statistical evidence. A study made by Adams, (p69), comparing fatality rates for european states with and without seatbelt laws, both before and after such laws were introduced, shows that these laws had an adverse effect on fatality rates in europe. A similar study which we have made of the corresponding data in the United States of America, (pp70-103), indicates that the introduction of seatbelt laws and, hence, the increased use of seatbelts, had no discernable effect on the fatality rates in the United States, based on NHTSA's own data. The raw data actually shows a slight increase, on average, but this increase is within the margin of error of the data. In the United States, fatality rates are based on estimates of total vehicle miles driven which are not known precisely. Likewise, if we look at the effect of the passage of seatbelt laws in individual states (pp70-103) we find that some states experienced a slight increase in fatality rates following the passage of the law, while other experienced a slight decrease. Again, these results are within the margin of error of the data. Thus, while the American data does not prove that there was an increase in fatality rates following the passage of seatbelt laws, neither does it show decrease. A study by Levine and Basilevsky, (pp104-105), based on the Canadian data, showed a substantial increase in fatalities after passage of seatbelt laws in Canada. This study was based on a smaller sample than the two studies mentioned above but is significant because the authors had full access to police and hospital records in the province of Manitoba over a seven year period during which the seatbelt laws were passed in Canada. The study showed that almost the entire increase in fatalities was among drivers wearing seatbelts. 
On May 21, 2003, Mr. Jeffrey Runge, the NHTSA Administrator, testifying under oath before the Subcommittee on Highways, Transit and Pipelines of the United States House of Representatives, made the statements given in quotation marks below. He repeated most of these statements in testimony given before the Subcommittee on Surface Transportation and Merchant Marine of the United States Senate on May 22, 2003. He again repeated most of these statements, with minor variations, in testimony before the same sub-committee on April 5, 2005. Because these statements have often been cited as justification for forcing people to wear seatbelts, we must examine them carefully. Let us take them one at a time: 
I. "Wearing safety belts is the number one offensive and defensive step all individuals can take to save lives." The above evidence shows that seatbelts save few, if any, lives and that wearing a seatbelt has numerous adverse effects. Furthermore, NHTSA data states that 63% of those involved in fatal crashes were driving in an unsafe manner. (p16, table 65).The "number one step that all drivers can take", therefore, is not to wear seatbelts but to drive safely. 
II. "Buckling belts is not a complex vaccine, doesn't have unwanted side effects and doesn't cost any money". In fact, equipping all automobiles and light trucks with seatbelts adds more than a billion dollars a year to the cost of the cars and light trucks sold in this country. As the published and eye witness accounts which we have submitted show, seatbelts have killed hundreds and, beyond a reasonable doubt, thousands of people. One cannot, therefore, maintain that they have "no unwanted side effects". 
III. "It is simple, it works, and it's life saving". While we have presented over a hundred cases of people being killed by seatbelts, NHTSA has not presented a single verified case of a person's life being saved by a seatbelt in an automobile accident. Moreover, the statistical data which we have presented shows that the wearing of seatbelts has no discernable effect on fatality rates in the U.S. and an adverse effect in europe. The statement that (buckling seatbelts) is "life saving" is, therefore, not supported by the evidence. 
IV."Safety belt use cuts the risk of death in a severe crash in half." This statement is demonstrably false. In head-on collisions approximately 83% of those killed were killed by being crushed, according to NHTSA's own data. Many of those wearing seatbelts are killed while many of those who are not wearing seatbelts survive. There is no evidence that seatbelts reduce the risk of death even in a head on collision. In all other types of collisions, seatbelts have either no effect or increase the risk of death or severe injury. According to NHTSA, at least 43% of those killed in automobile accidents were wearing seatbelts at the time. In some states, the total is as high as 65%. There is no evidence that if the rest had been wearing seatbelts they would not also have been killed. Furthermore, the statistical data shows that increased seatbelt use had no effect on the fatality rates in the United States and an adverse effect in europe. This is not surprising because the available evidence indicates that approximately 90% of all automobile accident victims are killed by being crushed. 
V. "Most passenger vehicle occupants killed in motor vehicle crashes continue to be totally unrestrained." This statement is presumably based on the data in table 88 (p12) which states that of those killed in passenger car crashes in 2001, 42.8% were wearing seatbelts. It also states, however, that in 8.6% of cases, seatbelt use was "unknown". When we limit ourselves to those cases in which seatbelt use was actually claimed to be known, the percentage use immediately goes up to 47%. Attention is called to the footnote which states that police reports may be biased in favor of seatbelts. Table 83 (p21) states that 64% (not 57.7% as stated in column 2) of "drivers in fatal crashes" were wearing seatbelts, among those where seatbelt use was known. (68% of fatal accident victims are drivers (Table 86, p.11)). For 9.7% it is stated to be unknown. It also repeats the caveat that police reports are not to be trusted where seatbelts are concerned. The obvious implication of the statement is that if more were wearing seatbelts more would be saved. But the preponderence of the evidence shows that this is not true. In short, the statement is both questionable and misleading. (See also paragraph VIII below). 
VI. "If safety belt use were to increase from the national average of 75% to 90% - an achievable goal - nearly 4,000 lives would be saved each year". A saving of 4,000 lives a year for a 15% increase in seatbelt use implies a saving of 26,680 lives a year in going from 0% to 100%, or a saving of 20,000 lives a year in going from 0% to 75%, the present NHTSA estimate of seatbelt use. But, as NHTSA's own data shows, not only did the introduction of seatbelt laws in states in this country not lead to a dramatic decline in their fatality rate relative to those states which had no such laws, it made no statistically discernable difference at all. Nor was there a statistically discernable difference between those states which had primary laws and those states which had secondary laws. (pp. 67, 68). In view of the fact that going from 0% to 75% made no discernable difference, The claim that going from 75% to 90% would save 4,000 lives a year not only has no basis in fact but is contradicted by both the the physical and the statistical evidence. 
VII. "For every one percentage point increase in safety belt use - that is, 2.8 million people "buckling up" - we would save hundreds of lives, suffer significantly fewer injuries and reduce economic costs by hundreds of millions of dollars a year". Mr. Runge did not state how he arrived at these conclusions but let us take them one at a time: 
(a)"hundreds of lives would be saved". As we have already seen, both the physical and the statistical evidence indicate that increased seatbelt use has either no effect or a negative effect on fatality rates. So the statement is contradicted by the evidence. 
(b) "suffer significantly fewer injuries". The references on pp 106-146, as well as the previous discussion and the cases we have cited, (pp25-49), show, however, that seatbelts cause numerous, severe and even fatal injuries, while there is no evidence of any offsetting benefit. As is pointed out in the discussion below, the second part of table 86 indicates that 88% of those injured were wearing seatbelts at the time, as opposed to an overall seatbelt use rate of 75% for the same year (2001). This is consistent with the physical evidence. Assuming there are 200,000,000 passenger car riders in the country, a seatbelt use rate of 75% implies 150,000,000 are using seatbelts and 50,000,000 are not. In table 86, the number of unbelted drivers injured in 2001 is given as 201,000, for a ratio of .00402. Had the number of belted drivers been injured at the same rate, the number of belted drivers injured would have been 150,000,000 x .00402 = 603,000. In fact, the total number of belted drivers injured is given by NHTSA as 1,570,000, an increase of 967,000. This number is the same regardless of the total number of passenger car occupants assumed. (603,000,000 is to 201,000 as 75 is to 25). (c)"reduce economic costs by hundreds of millions of dollars". In order for this to be true, seatbelts would have to reduce the number of accidents and/or make the existing accidents less severe. But, as we have already pointed out, a seatbelt reduces a driver's mobility in an emergency making an accident more likely. Moreover, the statistical evidence shows that seatbelts do not decrease the fatality rate, and, as shown in the previous paragraph, they increase the injury rate. Thus, seatbelts increase, not decrease, the total cost of accidents. And this does not include the extra billion dollars a year the American people must pay for cars as a result of the seatbelt laws. It must be concluded, therefore, that seatbelts do not "reduce the economic costs by hundreds of millions of dollars". 
Seatbelt advocates have used the data on table 86, p.11, to argue that because the table shows that, among automobile occupants killed in traffic accidents, 42.8%  were wearing seatbelts at the time, this proves that seatbelts save lives because 42.8% is lower than 75% which NHTSA claims was the percentage of all automobile occupants using seatbelts in 2001. This is a specious argument at best. One could just as easily argue that if 42.8% were wearing seatbelts but were killed anyway, the virtues of seatbelts are questionable. The crux of the matter is that no cause and effect relationship is shown. Our analysis of the physical and statistical evidence shows that wearing a seatbelt cannot be the cause of the difference. If these numbers are correct, we would have to look elsewhere for the cause. It could be, for example, that that tiny percentage subgroup of truly reckless drivers, the kind who get themselves and other people killed, are less prone to wear seatbelts than most. The total number of drivers involved in fatal collisions in 2001 was only 0.00018 of the total number of drivers and of those, 63% were driving recklessly and 68% of those killed in fatal accidents were drivers (op. cit. above). When we look carefully at table 86, we find that 8.6% of cases are listed as "unknown". Thus, if we accept their data and consider only those cases in which restraint use is known, the percent of those killed wearing seatbelts immediately goes up to 47%. When we look at the second part of table 86, "Passenger car occupants injured", we find that fully 88% of those injured were wearing seatbelts at the time, among those where seatbelt use was known. This is considerably higher than the 75% seatbelt use that NHTSA claimed for 2001. Note that the number who are injured is almost 100 times as great as those who are killed. Here, however, the data is consistent with the physical evidence. Now NHTSA added a note at the bottom of table 86 which says, "Restraint use is determined by police and may be overreported (sic) for survivors". No similar caveat is given for the first part of the table regarding reported restraint use for occupants who were killed. Since NHTSA has questioned the credibility of their own reports, it seems that police, having been trained to repeat the false mantra "seatbelts save lives", would be more likely to understate the number of killed wearing seatbelts than to overstate the number injured wearing seatbelts. 
(1) FARS stands for Fatal Accident Reporting system and is  a record, compiled and maintained by the federal government, which contains (according to the compilers) detailed information on every fatal accident which has taken place in this country since 1975. FARS is available on line from: ftp://ftp.nhtsa.dot.gov/FARS/ To analyse this data, however, requires first of all an SAS reader to display the data. Secondly, all the data is in code so that one must learn the code (which is also available on line). Finally, one must have some familiarity with statistical analysis. A head-on collision is defined as a collision between two vehicles moving in opposite directions along the same axis. A front-end collision, on the other hand, may take place at any angle and may be a collision with fixed objects, pedestrians, or bicyclists. The crush rate for fatal front-end collisions for 2002 was 91%; that is, the percent of vehicles involved which were severely crushed, and 91% of vehicle occupants who were killed in front end collisions were killed in vehicles which were severely crushed, according to FARS data. After 2001, FARS stopped listing head-on collisions as a separate category, lumping them instead with front-end collisions. (Instructions for filling out the FARS forms, 2002 FARS C&V Manual, p. A18 of the manual). 
 (2) When we read the instructions to police officers and emergency personnel for filling out the FARS data forms, we learn that all persons who fell off the bed of a pickup truck or fell off a snow-mobile or a three-wheel or four-wheel ATV or from a go-cart (!) are to be listed as having been "ejected". (2002 FARS C&V Manual, p. 447 of the manual). Moreover, there is no evidence to prove that all the persons who are listed as having been "ejected" actually were. When we look at the data to find out how all these people came to be "ejected", we see a category entitled "ejection path" with such sub-categories as "through the door", "through the window" and so on. When we look at the actual data we find that most of these data points are coded as "9" which is the FARS code in this category for "unknown". In other words, all they really know in most cases is that the victims was outside the vehicle when they arrived on the scene. How he got there, they have no idea. He might have jumped clear, he might have crawled out of the vehicle after the accident and expired on the sidewalk, he might have been pulled from the wreck by passers-by. Yet all these people are listed as having been "ejected".