Motorcycle Helmet Effectiveness Revisited-Methodology

2.0 METHODOLOGY

In 1989, using Fatality Analysis Reporting System (FARS) data for 1982 through 1987, Wilson estimated that motorcycle helmets were 29 percent effective in preventing occupant fatalities in motorcycle crashes. She based her calculation on a paired comparison study of crashes resulting in at least one fatality and involving one or more motorcycles each of which carried one rider and one passenger. The methodology was based on previous work by Evans on the effect of vehicle mass on driver fatalities (1984); Kahane (1986) and Partyka (1987, 1988) on the effectiveness of seat belts and other restraint systems; and Evans and Frick (1988), on the effectiveness of motorcycle helmets. Since 1989, NHTSA has used the 29 percent effectiveness to calculate the number of lives saved through the use of motorcycle helmets.

Motorcycle crashes with fatal outcomes involving one rider and one passenger were compared using a matched pairs procedure. Limiting the crashes to those involving a rider and passenger combination gives the following four possibilities:

Table 1
Rider/Passenger Helmet Use Combinations

	Passenger Not Helmeted	Passenger Helmeted
Rider Not Helmeted	Neither rider nor passenger helmeted	Rider not helmeted, passenger helmeted
Rider Helmeted	Rider helmeted, passenger not helmeted	Both rider and passenger helmeted

For each of the four possible combinations, ratios were calculated for rider to passenger fatalities and passenger to rider fatalities in a given year. For example, in those cases in which neither the rider nor the passenger used a helmet, the Fatality Ratio (FR) of rider to passenger deaths, is:

FR _(RNH) =Rider _(NH) /Psgr _(NH)

And in those cases where only the rider was helmeted, the Fatality Ratio is:

FR_(RH) =Rider^_(H) /Psgr _(NH)

The fatality ratios were used to calculate separate helmet effectiveness, expressed as a percentage, for riders and passengers. It is assumed that any difference between the fatality ratios for unhelmeted and helmeted motorcycle occupants in each of the possible scenarios is due to the effectiveness of the helmet. For example, the effectiveness of a helmet in preventing fatal injuries to a rider accompanied by an unhelmeted passenger would be:

Effectiveness = (FR_(RNH) - FR_(RH)) / FR_(RNH)

Wilson calculated the fatality ratios and effectiveness for each year of the six-year period from 1982 through 1987 (Table 2) to show the variation that occurs from year to year as the result of the small number of cases in some of the cells. To control for this variation, multiple years of data were used to calculate the overall effectiveness for the different combinations of riders and passengers. Aggregating data over the 5-year period from 1982 through 1987 results in an overall 29 percent effectiveness of helmets in preventing fatal injuries in otherwise fatal motorcycle crashes (lower right hand cell in Table 3).

Table 2
Rider/Passenger Fatalities in Motorcycle Crashes*
By Helmet Use Status 1982-1987

Year	Helmet Used		Number of Deaths		Fatality Ratio
Year	Rider	Passenger	Rider	Passenger	Rider/Psgr	Psgr/Rider
1982	No	No	297	237	1.253	0.798
	No	Yes	19	15	1.267	0.789
	Yes	No	37	44	0.841	1.189
	Yes	Yes	183	155	1.181	0.847
1983	No	No	263	224	1.174	0.852
	No	Yes	25	12	2.083	0.480
	Yes	No	24	31	0.774	1.292
	Yes	Yes	160	131	1.221	0.819
1984	No	No	243	232	1.047	0.955
	No	Yes	21	13	1.615	0.619
	Yes	No	30	36	0.833	1.200
	Yes	Yes	147	123	1.195	0.837
1985	No	No	265	236	1.123	0.891
	No	Yes	17	9	1.889	0.529
	Yes	No	31	36	0.861	1.161
	Yes	Yes	144	143	1.007	0.993
1986	No	No	263	247	1.065	0.939
	No	Yes	20	12	1.667	0.600
	Yes	No	26	31	0.839	1.192
	Yes	Yes	158	125	1.264	0.791
1987	No	No	279	251	1.112	0.900
	No	Yes	20	11	1.818	0.550
	Yes	No	20	19	1.053	0.950
	Yes	Yes	125	102	1.225	0.816
Total	No	No	1610	1427	1.128	0.886
	No	Yes	122	72	1.694	0.590
	Yes	No	168	197	0.853	1.173
	Yes	Yes	917	779	1.177	0.850

*Crash must involve both a rider and a passenger on the same motorcycle.
Source: National Center for Statistics and Analysis, NHTSA - FARS 1982-1987

Table 3
Helmet Use Effectiveness (Percent) For Motorcycle Riders
and Passengers 1982 through 1987

Effectiveness For:	Control	1982	1983	1984	1985	1986	1987	All Years
Rider	Unhelmeted Passenger	33	34	20	23	21	05	24
	Helmeted Passenger	07	41	26	47	24	33	31
	Average	20	38	23	35	26	19	27
Passenger	Unhelmeted Passenger	01	44	35	41	36	39	33
	Helmeted Passenger	29	37	30	14	34	14	28
	Average	15	40	33	28	35	26	30
Rider and Passenger	Unhelmeted Passenger	17	39	28	32	29	22	29
	Helmeted Passenger	18	39	28	31	29	23	29
	Average	17	39	28	31	29	23	29

Source: National Center for Statistics and Analysis, NHTSA - FARS 1982-1987