ARTICLE
10.1177/0193841X02250524
EVALUATION REVIEW / APRIL 2003
Stolzenberg, D'Alessio / "BORN TO BE WILD"
"BORN TO BE WILD"
The Effect of the Repeal of Florida's
Mandatory Motorcycle Helmet-Use Law
on Serious Injury and Fatality Rates
LISA STOLZENBERG
STEWART J. D'ALESSIO
Florida International University
In response to political pressure, the state of Florida repealed its mandatory motorcycle helmet-
use law for all operators and passengers older than the age of 21, effective July 1, 2000. Using
monthly data and a multiple time-series design, the authors assessed the effect of this law change
on serious injury and fatality rates for motorcycle riders aged 21 and older. Controls for serious
injury and fatality rates for motorcycle riders younger than 21 years of age were included in the
analyses. Maximum-likelihood results showed that the repeal of the mandatory helmet-use law
in Florida had little observable effect on serious injuries or on fatalities that resulted from
motorcycle crashes. Policy implications of these findings are discussed, and explanations are
given as to why the repeal of the mandatory motorcycle helmet-use law in Florida was
inconsequential.

Keywords: mandatory motorcycle helmet-use law; motorcycle crashes; motorcycle injuries;
motorcycle fatalities
Mandatory motorcycle helmet-use laws have become a popular strategy
to improve public safety. The primary purpose of these laws is to reduce seri-
ous injuries and fatalities that result from motorcycle crashes by requiring
operators of motorcycles and their passengers to wear protective helmets.
The underlying rationale for mandatory motorcycle helmet-use laws ema-
nates from research that finds that motorcycle riders wearing protective hel-
mets are much less likely to suffer injuries, particularly head injuries, and are
less apt to be killed in a crash (Evans and Frick 1988; Fleming and Becker
1992; Gabella et al. 1995; Sarkar, Peek, and Kraus 1995; Weiss 1992). Helmet
use is also reported to decrease the overall cost of medical care (Rowland
et al. 1996).
131
EVALUATION REVIEW, Vol. 27 No. 2, April 2003 131-150
DOI: 10.1177/0193841X02250524
2003 Sage Publications
Twenty-one states currently have full helmet-use laws, while another 25
states have limited helmet-use laws that exempt adult riders from wearing
helmets (see the appendix). Adult riders are usually defined as persons older
than 18 or 21 years of age. Only 4 states do not have any helmet-use laws.
These states include Colorado, Illinois, Iowa, and New Hampshire. Five
states also require motorcyclists to be covered by an insurance policy provid-
ing for at least $10,000 in medical benefits for injuries incurred as a result of a
crash while operating or riding as a passenger on a motorcycle.
Despite a plethora of studies that report that mandatory motorcycle helmet-
use laws are effective in reducing motorcycle crash injuries and fatalities
(Chenier and Evans 1987; Chiu et al. 2000; Fleming and Becker 1992;
Graham and Lee 1986; Hartunian et al. 1981; Kraus et al. 1994; Robertson
1976; Sosin and Sacks 1992; Sosin, Sacks, and Holmgreen 1990; Watson,
Zador, and Wilks 1980, 1981), a public backlash has developed against these
laws. Opponents of mandatory motorcycle helmet-use laws maintain that
protective helmets hinder a motorcycle operator's vision (Gordon and Prince
1975) and hearing (Purswell and Dorris 1977), thereby making accidents
more likely to occur. They also argue that motorcycle helmet use increases
the probability of certain types of injuries that transpire during a crash, pri-
marily neck and spinal cord injuries (Cooter et al. 1988; Goldstein 1986;
Huston and Sears 1981; Konrad et al. 1996; Krantz 1985).
It is also proffered that the repeal of mandatory helmet-use laws, by offer-
ing individuals a choice in whether to wear protective helmets, increases the
attractiveness and convenience of motorcycle use. This in turn affects miles
traveled and sales of motorcycles. For example, Kraus, Peek, and Williams
(1995) reported that after an unrestricted helmet-use law went into effect in
California in 1992, there was a substantial decrease in the number of motor-
cycles observed traveling on roads in various California cities. Finally, many
empirical evaluations of helmet-use laws are thought to be methodologically
flawed (Adams 1983; American Motorcyclist 1991; Perkins 1981). Prob-
lems include small unrepresentative samples, specification error, a myopic
focus on head injuries resulting from motorcycle crashes, and a continued
failure to employ multivariate statistical procedures, to name a few.
Determined and sustained political pressure from motorcycle enthusiasts
and advocacy groups has led a number of states to repeal their mandatory hel-
met-use laws. Florida is one such state. Florida repealed its mandatory
motorcycle helmet-use law on July 1, 2000, by authorizing that an individual
older than 21 years of age could operate or ride a motorcycle without wearing
protective headgear securely fastened on his or her head if such person is cov-
ered by an insurance policy providing for at least $10,000 dollars in medical
benefits for injuries incurred as a result of a crash while operating or riding on
132
EVALUATION REVIEW / APRIL 2003
a motorcycle. Individuals younger than 21 are still required to wear protec-
tive helmets when riding a motorcycle.
A continuing concern among the medical community is whether the
repeal of Florida's mandatory helmet-use law has increased serious injuries
and fatalities resulting from motorcycle crashes. To our knowledge, the only
study conducted to date that examines the impact of the repeal of Florida's
mandatory helmet-use law is a simple pretest-posttest analysis of head inju-
ries and fatalities that occurred before and after the law's repeal undertaken
by Hotz et al. (2002) at the University of Miami.
In this study, Hotz et al. (2002) gathered data on all motorcycle-related
cases that were admitted at either the Ryder Trauma Center or at the Univer-
sity of Miami/Jackson Memorial Medical Center's Surgical Emergency
Room for a 6-month period prior to the law's repeal (July 1, 1999, to Decem-
ber 31, 1999) and for a 6-month period following the law's repeal (July 1,
2000, to December 31, 2000). Information was obtained on 54 subjects prior
to the law's repeal and on 94 subjects after the law's repeal. Motorcycle-
related cases taken to other trauma centers or directly taken to the morgue by
paramedics were not included in the study. Using bivariate statistical proce-
dures, Hotz et al. found that the repeal of Florida's mandatory helmet-use law
engendered a substantial decrease in helmet use among motorcycle riders
and a sizable increase in the number of brain injuries and motorcycle fatali-
ties. They also noted an 81% increase in the number of motorcycle crash vic-
tims admitted to the two trauma centers following the repeal of the helmet-
use law.
Although the results generated from this study appear to support the posi-
tion that the repeal of Florida's mandatory motorcycle helmet-use law
increased motorcycle crash injuries and fatalities, it is important to recognize
that simple comparisons of prelaw and postlaw repeal means are only sugges-
tive. Further evidence is needed before accepting these findings as definitive
because it is not possible to say whether the observed changes in motorcycle
crash injuries and fatalities were due to the repeal of the law or whether they
resulted from a preexisting trend or some other salient factor. As Charles
Branas and C. William Schwab (2002) of the University of Pennsylvania
Medical Center pointed out,
Numerous shortcomings in this work should probably limit its interpretation to, at best, a
very rough, unadjusted guide to the impact of repealing Florida's helmet law. Changes in
numerous other variables (i.e., the number of registered motorcycles, roadway types,
alcohol consumption, weather, available police resources) . . . might have also affected
the occurrence of motorcycle helmet crash victims presenting to the University of Miami
Medical Center. (Pp. 473-74)
Stolzenberg, D'Alessio / "BORN TO BE WILD"
133
In this article, we investigate further the impact of Florida's repeal of its
mandatory motorcycle helmet-use law, correcting for the methodological
problems encountered in earlier research. We do not, however, aim to repli-
cate the Hotz et al. (2002) study that analyzed clinical data. Rather, for analyt-
ical and practical reasons, we contribute to the literature by using monthly
data over a 192-month period and a multiple time-series design to better esti-
mate the effect of Florida's repeal of its mandatory helmet-use law on motor-
cycle crash injuries and fatalities.
The type of data and the analytic strategy used in this study have several
methodological advantages. First, although no research design guarantees
correct inferences, the multiple time-series design is considered an effective
quasi-experimental design for drawing causal inferences (Johnson and
Christensen 2000). This design, which is depicted below, involves compari-
sons of series of observations (O) over time expected to be affected by an
intervention (X) with a control series not expected to be influenced by the
same intervention:
O
A1
O
A2
O
A3
O
A4
O
A5
O
A6
O
A7
O
A8
O
A9
O
A10
O
B1
O
B2
O
B3
O
B4
O
B5
O
B6
O
B7
O
B8
O
B9
O
B10
.
This design helps to rule out a large number of plausible alternative explana-
tions for a hypothesized causal relationship because the comparison series
acts to reduce the possibility of history effects.
In addition, although the simple diagram shown above depicts only one
experimental and one control series, multiple experimental and control series
are examined in this study. Specifically, we compare the trends in the serious
injury and fatality rates for motorcycle riders aged 21 and older before the
helmet-use law was repealed with trends after the law was repealed. If the dif-
ferences between the preintervention and postintervention series are positive
and greater than one should expect from chance, then a significant effect of
the law's repeal on serious injuries and fatalities can be inferred. Of course,
because the repeal of the motorcycle helmet-use law applied only to motorcy-
cle operators and passengers older than 21 years of age, the repeal of the law
should have little if any effect on the serious injury and fatality rate series for
motorcycle riders younger than 21 years of age. The use of the younger-than-
21 serious injury and fatality rates as statistical controls in the analyses helps
us to avoid attributing significance to the repeal of the law that should, more
accurately, be attributed to some other independent but coincidental event
such as a change in weather conditions.
We also use month rather than year as our unit of analysis because monthly
data are considered superior for interpreting change and for reducing the
134
EVALUATION REVIEW / APRIL 2003
confounding of history effects (Tiao and Wei 1976). Analyzing monthly data
also permits greater flexibility in applying more sophisticated and efficient
statistical procedures because of the increased number of observations.
Finally, because this analysis compares changes in serious injury and fatality
rates in one state over time and does not predict changes across different
states, potential biases resulting from dissimilarities in state accident-report-
ing practices are minimized.
The policy implications of this study are noteworthy. If the repeal of
Florida's mandatory motorcycle helmet-use law is related positively to seri-
ous injuries and fatalities that result from motorcycle crashes, once trend and
seasonal fluctuations are taken into account, then we can conclude that man-
datory helmet-use laws are effective in improving motorcycle safety. Yet if
the repeal of Florida's mandatory helmet-use law has no observable impact
on motorcycle crash injuries and fatalities, then motorcycle safety might be
improved more effectively through alternative mechanisms.
DATA
The data used in this study were obtained from the Traffic Crash Database
maintained by the Florida Department of Highway Safety and Motor Vehi-
cles, Office of Management and Planning Services (OMPS). The OMPS
Traffic Crash Database encompasses data from 1986 through 2001. This
database contains information on motor vehicle crashes that occur on public
roads in the state of Florida.
Two dependent variables are used in the analysis. The first endogenous
variable of theoretical import is the serious injury rate for motorcycle riders
older than 21 years of age. This variable is operationalized as the monthly
number of motorcycle crashes with serious injuries involving operators and
passengers aged 21 and older occurring on interstates, highways, county
roads, and municipal roads divided by motorcycle registrations and multi-
plied by 100,000. Serious motorcycle crash injuries are defined as incapaci-
tating injuries in which there are visible sign(s) of injury from the crash (e.g.,
bruises, abrasions, limping, etc.) and the person(s) was assisted from the
crash scene.
The second dependent variable of interest, the fatal motorcycle crash rate,
is measured as the monthly number of fatal motorcycle crashes involving
operators and passengers aged 21 and older occurring on interstates, high-
ways, county roads, and municipal roads divided by motorcycle registrations
Stolzenberg, D'Alessio / "BORN TO BE WILD"
135
and multiplied by 100,000. A fatal injury occurs when an injury sustained in a
motorcycle crash results in the death of the individual within 90 days.
We analyze the effect of Florida's repeal of its mandatory motorcycle
helmet-use law with a dummy variable coded 0 before July 2000 and 1 other-
wise. The two control variables are measured as the monthly number of
motorcycle crashes with serious injuries or fatalities involving operators and
passengers younger than 21 occurring on interstates, highways, county
roads, and municipal roads divided by motorcycle registrations and multi-
plied by 100,000.
DESCRIPTIVE ANALYSIS
Figures 1 and 2 compare the mean changes between preintervention and
postintervention periods for the serious injury and fatality rate series. For
these comparisons, we use the 174 months preceding the repeal of the helmet-
use law (January 1, 1986, to June 30, 2000) and the 18 months following the
repeal of the law (July 1, 2000, to December 31, 2001).
A cursory glance at Figure 1 indicates little support for the assertion that
the repeal of Florida's mandatory helmet-use law caused an increase in the
serious injury rate. In fact, serious injuries actually decreased following the
repeal of the law. The mean level of the preintervention serious injury rate for
motorists older than 21 was 56.61; the mean level of this series after the helmet-
use law's repeal decreased to 45.52. Further examination of Figure 1 shows
that the mean level of the serious injury rate for motorists younger than 21
also decreased substantially following the repeal of the helmet-use law. The
preintervention serious injury rate for motorists younger than 21 was 16.95.
After the helmet-use law's repeal, the mean level of the series was reduced to
5.37. The exact reasons for these decreases are not readily apparent.
Figure 2 depicts mean changes between the preintervention and
postintervention periods for the two fatality rate series. The overall means for
the 21 and older series and for the younger than 21 series during the
preintervention period were 6.25 and 1.25, respectively. After the repeal of
the law, the fatality rate for riders aged 21 and older increased to 6.94, and the
fatality rate for motorists younger than 21 years of age decreased to 0.87.
However, both of these rate changes are small and appear to be the result of
random fluctuations.
Although these initial results appear to support the position that the repeal
of the mandatory motorcycle helmet-use law did not increase the serious
injuries and fatalities substantially in Florida, it is important to recognize that
136
EVALUATION REVIEW / APRIL 2003
simple comparisons of preintervention and postintervention means are only
suggestive at best. Further empirical evidence is necessary before accepting
these findings as definitive because it is not possible to say whether these
Stolzenberg, D'Alessio / "BORN TO BE WILD"
137
Figure 1: Means for Preintervention and Postintervention Motorcycle Serious
Injury Rates in Florida
Figure 2: Means for Preintervention and Postintervention Motorcycle Fatality
Rates in Florida
observed changes in motorcycle crash injuries and fatalities are due to the
repeal of the law or whether they are simply the result of preexisting trends.
To illustrate this point, we constructed graphical charts that depict the seri-
ous injury and fatality rates over time (see Figures 3 and 4). The vertical line
depicted in each of the figures represents the repeal of the helmet-use law in
Florida. A visual examination of Figure 3 suggests that the serious injury
rates were trending downward slightly prior to the repeal of law. Conse-
quently, a question remains as to whether the large observed reductions in the
serious injury rates during the postintervention period, as depicted in Fig-
ure 1, resulted from the repeal of the helmet-use law or from a preexisting
downward trend. In contrast, both fatality rate series appear to have remained
relatively stable over time.
MULTIVARIATE
TRANSFER FUNCTION ANALYSES
We began the multivariate transfer function analyses by constructing
univariate autoregressive integrative moving-average (ARIMA) models for
the each of the four series for the 174-month period preceding the repeal of
the mandatory helmet-use law. The univariate ARIMA, which typically is
developed through an iterative model-building strategy, accounts for the sto-
chastic processes associated with a series (Box, Jenkins, and Reinsel 1994).
Several factors must be considered in selecting an appropriate univariate
ARIMA model. One important consideration is whether the series has a sin-
gle constant variance throughout its course. A nonstationary variance is
engendered by dramatic fluctuations in variation between observations in a
series. To determine whether each of the four series was stationary in vari-
ance, we consulted a rule-based expert system in the statistical software pro-
gram Forecast Pro (Stellwagen and Goodrich 2000). This system, which uses
a goodness-of-fit measure to compare competing models, recommended that
a natural logarithm transformation was necessary to stabilize the variance of
the older-than-21 fatality rate series.
Another salient consideration is whether a series has a single constant
level throughout its course. That is, a series should not "trend" or "drift"
upward or downward over time. A commonly used test for the presence of an
unstable level is the "augmented" Dickey-Fuller test (Dickey, Bell, and
Miller 1986). This test, which assesses whether a series has a unit root, indi-
cated that the younger-than-21 serious injury and fatality rate series were
trended and required first-order differencing.
138
EVALUATION REVIEW / APRIL 2003
Figure 3: Motorcycle Serious Injury Rates in Florida
NOTE: Florida's repeal of its mandatory helmet-use law became effective on July 1, 2000.

139
140
Figure 4: Motorcycle Fatality Rates in Florida
NOTE: Florida's repeal of its mandatory helmet-use law became effective on July 1, 2000.
A third consideration is whether a series has any cyclical or periodic fluc-
tuation that repeats itself each time at the same phase of the cycle or period.
This repetitive variation, commonly known as seasonality, is most likely to
occur at yearly intervals with monthly data. Our examination of each of the
series autocorrelation functions (ACFs) at lags of 12 months, 24 months, 36
months, and 48 months indicated that none of the series needed to be season-
ally differenced.
Once the four series were determined to be stationary in variance and
level, we examined each series ACF and partial autocorrelation function for
autoregressive and for moving-average processes. In an autoregressive pro-
cess, the current value in a series is influenced by an exponentially weighted
sum of one or more previous values. That is, the effect of one or more prior
observations (i.e., the order of the autoregressive parameter) on the current
observation diminishes over time: Y
t
=

1
Y
t 1
+ ... +

p
Y
t p
+ a
t
. In contrast,
each value in a moving-average process is determined by the average of the
current disturbance and one or more previous disturbances. The effect of a
moving-average process lasts for a finite number of periods (i.e., the order of
the moving-average parameter) and then vanishes abruptly: Y
t
= a
t


1
a
t 1

...

q
a
t q
.
After constructing the univariate ARIMA models, we used the automatic
multivariate transfer function procedure in Autobox (Reilly 1999) to assess
the impact of the mandatory motorcycle helmet-use law's repeal on the seri-
ous injury and fatal crash rates for operators older than 21 years of age.
1
Both
of the younger-than-21 series were employed as control variables in these
equations. Three different intervention models were assessed. First, we con-
sidered the possibility that the older-than-21 motorcycle crash injury and
fatality rates declined sharply after the repeal of the mandatory motorcycle
helmet-use law went into effect and then remained at these lower levels over
time. We also investigated the possibility that the repeal of the law had a small
initial impact on the serious injury and fatality rates that grew larger over
time. This transfer function has two effect parameters. The intervention or
omega parameter (
) measures the degree of change in the level of the serious
injury and fatality series; the delta parameter (
) estimates the amount of time
required for these changes to be actualized. The larger the value of the delta
parameter, the more gradual the impact of the law's repeal. Conversely, a
small delta coefficient would suggest that the effect of the law's repeal
occurred more rapidly. Finally, we tested whether serious injuries and fatali-
ties were reduced initially by the law's repeal but then returned to preexisting
levels as time passed.
2
Our analyses indicated that for the older-than-21 serious injury and fatal-
ity rate series, the abrupt permanent change model provided the best fit to the
Stolzenberg, D'Alessio / "BORN TO BE WILD"
141
data. The zero-order transfer function models the law's repeal as having an
abrupt and permanent effect on motorcycle crash injuries and fatalities. If the
intervention coefficients are positive and statistically significant, the proposi-
tion that the repeal of Florida's mandatory motorcycle helmet-use law
increased the serious injury and fatality rates above that expected on the basis
of preexisting trends would be supported.
Tables 1 and 2 present the maximum-likelihood coefficients along with t
values to evaluate statistical significance. A Box-Ljung Q statistic (Ljung and
Box 1978), which tests the null hypothesis that a set of sample
autocorrelations is associated with a random process, indicated that the resid-
uals for the models were uncorrelated (i.e., constituted "white noise").
3
The
results displayed in Table 1 indicate that the repeal of the helmet-use law did
not increase the serious injury rate above that expected on the basis of preex-
isting trends. The intervention coefficient is negative and is not statistically
significant. Visual examination of Table 1 also shows that the serious injury
rate for motorcycle riders younger-than-21 years of age has a substantive
effect. Both the omega and the delta coefficients are statistically significant,
thereby indicating that the younger-than-21 control series has a permanent
but gradual effect on the older than 21 serious injury motorcycle crash rate.
However, as correctly noted by an anonymous reviewer, it is possible that the
relationship between the younger-than-21 series and the older-than-21 series
is spurious; both series may be affected by one or more common causes such
as weather conditions.
Table 2 also reveals little evidence that the repeal of the mandatory helmet-
use law affected the fatal motorcycle crash rate. Although the intervention
variable is related positively to the fatal crash rate, its effect is not statistically
significant. This finding runs counter to the prediction that the repeal of the
helmet-use law in Florida influenced motorcycle crash fatalities above that
predicted on the basis of preexisting trends. Our analysis also demonstrates
that the effect of the younger-than-21 fatality rate control series is inconse-
quential. In sum, our results indicate that preexisting trends rather than the
repeal of the motorcycle helmet-use law account for the changes in the seri-
ous injury and fatality rates observed initially in Figures 1 and 2.
DISCUSSION AND CONCLUSION
We began this article by noting that although the state of Florida recently
repealed its mandatory motorcycle helmet-use law, only one study has been
published to our knowledge that evaluates its impact. This study found that
142
EVALUATION REVIEW / APRIL 2003
the repeal of Florida's mandatory helmet-use law engendered a substantial
decrease in helmet use among motorcycle riders and a large increase in the
number of brain injuries and motorcycle crash fatalities. However, because
this study failed to account for differences between preintervention and
postintervention periods, its findings are questionable. The factors engender-
ing injury and death in motorcycle crashes are much more complex than
Stolzenberg, D'Alessio / "BORN TO BE WILD"
143
TABLE 1: Maximum-Likelihood Coefficients for the Motorcycle Serious Injury
Rate (Age > 21) Equation
Model Component
Coefficient
Standard Error
p
Value
t
Value
Constant
1.290
0.775
.098
1.66
AR(1)
0.194
0.065
.003
2.98
serious injury rate (age < 21)
Delta
0.970
0.014
.000
69.92
Omega
0.678
0.142
.000
4.78
Law change
Omega
1.150
2.420
.636
0.47
NOTE: The equation for the full model is
Y
( )
T
= 1.597 + [
X
1( )
T
][(1
B
**1)][(1 .970B**
1)]**1 [(+ .678)] + [
X
2( )
T
][(1.150)] + [(1 .194B** 1)]**1 [
A
( )
T
], where
Y
= serious
injury rate (age > 21),
X
1 = serious injury rate (age < 21), and
X
2 = law change.
= first
difference. To produce more robust models, Autobox's outlier detection procedure was
used to identify potential interventions (i.e., level shifts, seasonal pulses, and single-
point outliers). Although not shown in the table, dummy variables were included in the
analyses to model these interventions.
TABLE 2: Maximum-Likelihood Coefficients for the LN Motorcycle Fatality Rate
(Age > 21) Equation
Model Component
Coefficient
Standard Error
p
Value
t
Value
Constant
1.734
0.025
.000
67.83
fatality rate (age < 21)
Omega
0.009
0.022
.675
0.42
Law change
Omega
0.133
0.085
.119
1.57
NOTE: The equation for the full model is
Y
( )
T
= 1.734 + [
X
1( )
T
][(1
B
**1)][(+.009)] +
[
X
2( )
T
][(+.133)] + [
A
( )
T
], where
Y
= natural logarithmetic fatality rate (age > 21),
X
1 =
fatality rate (age < 21), and
X
2 = law change.
= first difference.To produce more robust
models, Autobox's outlier detection procedure was used to identify potential interven-
tions (i.e., level shifts, seasonal pulses, and single-point outliers). Although not shown in
the table, dummy variables were included in the analyses to model these interventions.
simply whether the motorcycle rider was wearing a protective helmet during
the crash.
Against this backdrop, we analyzed time-series data drawn from the entire
state of Florida to determine the effect of Florida's repeal of its mandatory
motorcycle helmet-use law on serious injuries and fatalities resulting from
motorcycle crashes. Our results indicate that the law's repeal had little
observable influence on the serious injury or fatality rate for motorcycle rid-
ers older than 21 years of age. The absence of an effect is rather surprising,
considering the findings published in previous research studies.
The question that begs answering is why the repeal of Florida's mandatory
helmet-use law had no substantive effect on motorcycle crash injuries and
fatalities. We offer a few explanations that warrant discussion. One possibil-
ity is that because motorcycle crashes are often very severe, a diminishing
marginal return can be expected by further increases in levels of motorcycle
safety. An example drawn from the airline industry may be instructive. An
airliner suffers engine trouble at 30,000 feet and is spiraling out of control
toward the ground. A stewardess informs the passengers to fasten their safety
belts. However, whether the passengers obey the stewardess's instructions is
relatively meaningless because the odds of surviving the plane crash are
extremely low notwithstanding whether a passenger is wearing his or her
safety belt. Because many fatal motorcycle crashes are extremely severe, it
seems unlikely that the repeal of a motorcycle helmet-use law would have
any statistically discernible effect on serious injury or fatality rates simply
because it raised the level of driver safety slightly further. Supporting this
logic, it has been shown that motorcycle helmets are "most effective in less
severe collisions that result in mild to moderate injuries" (Offner, Rivara, and
Maire 1992, 640).
By the same token, the effectiveness of protective helmets in reducing
serious injury and fatality rates depends on the proportion of motorcycle
crash fatalities that result from head injuries. That is, for helmet use to be
effective in saving lives, head injuries in motorcycle crashes must account for
a large percentage of all motorcycle crash fatalities. However, this appears
not to be the case. Although Sosin and Sacks (1992) found that helmet-use
laws were effective in preventing nonfatal head injuries in nonfatal motorcy-
cle crashes, they did not observe any difference between states with and with-
out mandatory helmet-use laws in reference to overall fatality rates from
motorcycle crashes. Their findings suggest that although motorcycle helmets
may save lives, the number of lives saved is so small that it has no statistically
discernable impact on the overall number of motorcycle crash deaths.
A third possibility relates to the relationship between helmet use and
motorcycle crashes. Let us accept for the moment that motorcycle helmets
144
EVALUATION REVIEW / APRIL 2003
are effective in saving lives in motorcycle crashes as reported in many clinical
studies. However, if motorcycle helmets also impair the vision and/or hearing
of motorcycle operators, helmeted motorcycle riders may also have a greater
chance of getting into crashes than do unhelmeted drivers. Consequently, any
beneficial effect of motorcycle helmets in saving riders' lives would be coun-
terbalanced by making motorcycle accidents more likely for helmeted riders.
In a similar vein, it is also plausible that the adverse consequences of helmet
nonuse in crashes is being negated by increased care on the part of the indi-
vidual when riding a motorcycle. Sosin and Sacks (1992) reported a substan-
tially higher motorcycle crash rate in states that have mandatory helmet-use
laws, whereas Lin, Hwang, and Kuo (2001) noted that unhelmeted riders have
fewer collisions with moving cars than do helmeted motorcycle riders. We
also observed that the motorcycle crash rate for riders older than 21 decreased
by approximately 21% following the repeal of the helmet-use law in Florida.
A fourth possibility is that the repeal of the motorcycle helmet-use law had
little effect on whether motorcyclists wore their protective helmets. It is esti-
mated that even without a motorcycle helmet-use law, between 40% and 60%
of motorcycle riders still wear helmets (Kraus, Peek, and Williams 1995).
Following the repeal of Florida's mandatory helmet-use law, approximately
46% of the individuals older than age 21 involved in motorcycle crashes were
still wearing protective helmets.
Although our results have important policy implications, it would be pre-
mature to accept them as definitive and final. Certain caveats must be enter-
tained. First, our finding of a null effect of the law's repeal asks the question
"Why?" How can it be that the repeal of the mandatory motorcycle helmet-
use law in Florida had no observable effect on serious injury and fatality
rates, when many other empirical studies report precisely the opposite? The
most likely explanation relates to methodological considerations. Previous
research has been handicapped by a continued reliance on small and unrepre-
sentative samples, specification problems, and a failure to use multivariate
statistical procedures. The current study addressed most of these problems.
The weak influence of the law's repeal on serious injury and fatality rates
might also be attributable to our multivariate transfer function (ARIMA)
analysis, which is considered to be a relatively conservative statistical
procedure.
Second, there will always remain a question as to whether the evidence
presented here suffices to sufficiently discredit the utility of mandatory
motorcycle helmet-use laws. For example, one could make a reasonable
argument that a longer postintervention would be needed to evince an inter-
vention effect. Although it would have been desirable to extend the period of
analysis, difficulties in obtaining additional data precluded extending the
Stolzenberg, D'Alessio / "BORN TO BE WILD"
145
series. Nevertheless, other studies that use even shorter postintervention peri-
ods report that either the implementation or the repeal of mandatory helmet-
use laws influence motorcycle crash injuries and/or fatalities (Chiu et al.
2000; Fleming and Becker 1992; Hotz et al. 2002; Kraus et al. 1994). In addi-
tion, because the monthly number of motorcycle fatalities in Florida is rather
small, the fatality rate series is somewhat unstable. Thus, it is important that
readers view our findings in reference to the impact of the law's repeal on the
fatality rate for motorcycle riders older than 21 years of age with some
healthy skepticism.
Third, the data analyzed here were drawn from one large state. Would our
failure to find a strong positive effect of the repeal of the mandatory motorcy-
cle helmet-use law on serious injuries and fatalities in Florida hold for other
states? Does state size or state demographics play a role in determining the
impact of motorcycle helmet-use legislation? Others should consider repli-
cating this study in other states that have recently repealed their mandatory
helmet-use laws. The more frequently such research is conducted, the greater
confidence we can place in the generalizability of our findings.
Notwithstanding these caveats, the implications of our findings for policy
are clear. The notion that Florida's repeal of its mandatory helmet-use law
increased significantly serious injuries and fatalities resulting from motorcy-
cle crashes has been assumed implicitly by policy makers and social scien-
tists alike, but such a belief is not supported by this study. When motorcycle
registrations, preexisting trends, and seasonal factors are taken into account,
the repeal of Florida's helmet-use law has little observable effect on serious
injury and fatality rates.
It would seem particularly fruitful for other researchers to identify more
precisely the specific mechanisms responsible for our findings. It also seems
clear that those who share our interest in improving motorcycle safety might
be better served by shifting their attention away from mandatory helmet-use
laws to other potentially more effective measures. Driver training and educa-
tional programs might be considered a viable alternative. As Perkins (1981,
295) asserted, "Prevention through rider and driver education may be consid-
erably more cost-effective and save many more lives than mandatory helmet
laws." Other measures that help increase the visibility of motorcyclists in
traffic may also prove effective.
The effect of mandatory helmet-use laws on motorcycle crash injuries and
fatalities is an important question that is raised frequently, with scant and
often questionable empirical evidence on which to base definitive answers.
The purpose of this study was to shed additional light on this issue. Because
our findings show that Florida's repeal of its mandatory motorcycle helmet-
use law did not increase the serious injury or fatality rates, we conclude that
146
EVALUATION REVIEW / APRIL 2003
policy makers should probably consider revising or repealing these types of
laws. We fully expect that our findings and conclusions will be elaborated
and challenged in future empirical work.
APPENDIX
Helmet-Use Law Statutes by State
State
Type of Helmet-Use Law
Alabama
Helmet law, all riders
Alaska
Helmet law, age exemptions
Arizona
Helmet law, age exemptions
Arkansas
Helmet law, age exemptions and insurance requirement
California
Helmet law, all riders
Colorado
100% helmet law free
Connecticut
Helmet law, age exemptions
Delaware
Helmet law, age exemptions
Florida
Helmet law, age exemptions and insurance requirement
Georgia
Helmet law, all riders
Hawaii
Helmet law, age exemptions
Idaho
Helmet law, age exemptions
Illinois
100% helmet law free
Indiana
Helmet law, age exemptions
Iowa
100% helmet law free
Kansas
Helmet law, age exemptions
Kentucky
Helmet law, age exemptions and insurance requirement
Louisiana
Helmet law, age exemptions and insurance requirement
Maine
Helmet law, age exemptions
Maryland
Helmet law, all riders
Massachusetts
Helmet law, all riders
Michigan
Helmet law, all riders
Minnesota
Helmet law, age exemptions
Mississippi
Helmet law, all riders
Missouri
Helmet law, all riders
Montana
Helmet law, age exemptions
Nebraska
Helmet law, all riders
Nevada
Helmet law, all riders
New Hampshire
100% helmet law free
New Jersey
Helmet law, all riders
New Mexico
Helmet law, age exemptions
New York
Helmet law, all riders
North Carolina
Helmet law, all riders
North Dakota
Helmet law, age exemptions
Ohio
Helmet law, age exemptions
Oklahoma
Helmet law, age exemptions
Oregon
Helmet law, all riders
Stolzenberg, D'Alessio / "BORN TO BE WILD"
147
Pennsylvania
Helmet law, all riders
Rhode Island
Helmet law, age exemptions
South Carolina
Helmet law, age exemptions
South Dakota
Helmet law, age exemptions
Tennessee
Helmet law, all riders
Texas
Helmet law, age exemptions and insurance requirement
Utah
Helmet law, age exemptions
Vermont
Helmet law, all riders
Virginia
Helmet law, all riders
Washington
Helmet law, all riders
West Virginia
Helmet law, all riders
Wisconsin
Helmet law, age exemptions
Wyoming
Helmet law, age exemptions
Notes
1. Box, Jenkins, and Reinsel (1994) provided a comprehensive discussion of the iterative
procedures used in estimating multivariate transfer function models. Suffice it to say that once
the univariate autoregressive integrative moving-average model for each stochastic series is
identified, each of the stochastic series is then prewhitened. Prewhitening is a process that
applies a given set of autoregressive and moving-average factors to a stationary series. Each sto-
chastic input series is prewhitened by its own autoregressive and moving-average factors. The
output series is then prewhitened by the input series autoregressive and moving-average factors.
If there are multiple input series, than the stationary output series is prewhitened once for each
different input series. Prewhitening is advantageous because it helps to remove the
intrarelationship in a series, thereby allowing a more accurate appraisal of the interrelationship
between the input and output series.
2. The intervention variables were coded as pulse functions in these analyses.
3. Although reported in Table 1, the nonsignificant parameters were eliminated from the
models prior to the calculation of the Box-Ljung Q statistics.
REFERENCES
Adams, J. G. U. 1983. Public safety legislation and the risk compensation hypothesis: The exam-
ple of motorcycle helmet legislation. Environment and Planning C: Government Policy
2:193-203.
148
EVALUATION REVIEW / APRIL 2003
Appendix (continued)
State
Type of Helmet-Use Law
American Motorcyclist. 1991. New helmet study; same old flaws. American Motorcyclist
45:34-35.
Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 1994. Time series analysis: Forecasting and con-
trol. 3d ed. Englewood Cliffs, NJ: Prentice Hall.
Branas, C. C., and C. W. Schwab. 2002. Motorcycle helmet laws require careful study. Journal of
Trauma, Injury, Infection, and Critical Care
52:437-74.
Chenier, T. C., and L. Evans. 1987. Motorcyclist fatalities and the repeal of mandatory helmet
wearing laws. Accident Analysis and Prevention 19:133-39.
Chiu, W., C. Kuo, C. Hung, and M. Chen. 2000. The effect of the Taiwan motorcycle helmet use
law on head injuries. American Journal of Public Health 90:793-96.
Cooter, R. D., A. J. McLean, D. J. David, and D. A. Simpson. 1988. Helmet-induced skull base
fracture in a motorcyclist. Lancet 1:84-85.
Dickey, D. A., W. R. Bell, and R. B. Miller. 1986. Unit roots in time series models: Tests and
implications. American Statistician 40:12-26.
Evans, L., and M. C. Frick. 1988. Helmet effectiveness in preventing motorcycle driver and pas-
senger fatalities. Accident Analysis and Prevention 20:447-58.
Fleming, N. S., and E. R. Becker. 1992. The impact of the Texas 1989 motorcycle helmet law on
total and head-related fatalities, severe injuries, and overall injuries. Medical Care 30:832-42.
Gabella, B., K. L. Reiner, R. E. Hoffman, M. Cook, and L. Stallones. 1995. Relationship of hel-
met use and head injuries among motorcycle crash victims in El Paso County, Colorado,
1989-1990. Accident Analysis and Prevention 27:363-69.
Goldstein, J. P. 1986. The effect of motorcycle helmet use on the probability of fatality and the
severity of head and neck injuries. Evaluation Review 10:355-75.
Gordon, S., and J. Prince. 1975. Field of view with and without motorcycle helmets. Technical
report DOT HS-801-75. Washington, DC: National Highway Traffic Safety Administration.
Graham, J. D., and Y. Lee. 1986. Behavioral response to safety regulation: The case of motorcy-
cle helmet-wearing legislation. Policy Sciences 19:253-73.
Hartunian, N. S., C. N. Smart, T. R. Willemain, and P. L. Zador. 1981. The economics of safety
deregulation: An analysis of excess death and costs due to motorcycle helmet law repeal in
the United States. Journal of Health Politics, Policy and Law 8:86-78.
Hotz, G. A., S. M. Cohn, C. Popkin, P. Ekeh, R. Duncan, E. W. Johnson, F. Pernas, and J. Selem.
2002. The impact of a repealed motorcycle helmet law in Miami-Dade County. Journal of
Trauma, Injury, Infection, and Critical Care 52:469-74.
Huston, R. L., and J. Sears. 1981. Effect of protective helmet mass on head/neck dynamics. Jour-
nal of Biomedical Engineering 103:18-23.
Johnson, B., and L. Christensen. 2000. Educational research: Quantitative and qualitative
approaches. Boston: Allyn & Bacon.
Konrad, C. J., T. S. Fieber, G. K. Schuepfer, and H. R. Gerber. 1996. Are fractures of the base of
the skull influenced by mass of the protective helmet? A retrospective study in fatally injured
motorcyclists. Journal of Trauma 41:854-58.
Krantz, K. P. G. 1985. Head and neck injuries to motorcycle and moped riders with special regard
to the effect of protective helmets. Injury 16:253-58.
Kraus, J. F., C. Peek, and A. Williams. 1995. Compliance with the 1992 California motorcycle
helmet use law. American Journal of Public Health 85:96-99.
Kraus, J. F., C. Peek, D. L. McArthur, and A. Williams. 1994. The effect of the 1992 California
motorcycle helmet-use law on motorcycle crash fatalities and injuries. Journal of the Ameri-
can Medical Association 272:1506-11.
Lin, M., H. Hwang, and N. Kuo. 2001. Crash severity, injury patterns, and helmet use in adoles-
cent motorcycle riders. Journal of Trauma 50:24-30.
Stolzenberg, D'Alessio / "BORN TO BE WILD"
149
Ljung, G. M., and G. E. P. Box. 1978. On a measure of lack of fit in time series models.
Biometrika 65:297-303.
Offner, P. J., F. P. Rivara, and R. V. Maire. 1992. The impact of motorcycle helmet use. Journal of
Trauma 32:636-41.
Perkins, R. J. 1981. Perspectives on the public good. American Journal of Public Health
71:294-95.
Purswell, J. L., and A. L. Dorris. 1977. Auditory and biomechanical performance requirements
for motorcycle helmets. In Proceedings of human factors society--21st annual meeting,
226-30.
Reilly, D. P. 1999. Autobox for Windows, 5.0. Hatboro, PA: Automatic Forecasting Systems.
Robertson, L. S. 1976. An instance of effective legal regulation: Motorcycle helmet and daytime
headlamp laws. Law and Society 10:476-77.
Rowland, J., F. Rivara, P. Salzberg, R. Soderberg, R. Maier, and T. Koepsell. 1996. Motorcycle
helmet use and injury outcome and hospitalization costs from crashes in Washington State.
American Journal of Public Health 86:41-45.
Sarkar, S., C. Peek, and J. F. Kraus. 1995. Fatal injuries in motorcycle riders according to helmet
use. Journal of Trauma 38:242-45.
Sosin, D. M., and J. J. Sacks. 1992. Motorcycle helmet-use laws and head injury prevention.
Journal of the American Medical Association 267:1649-51.
Sosin, D. M., J. J. Sacks, and P. Holmgreen. 1990. Head injury-associated deaths from motorcy-
cle crashes: Relationship to helmet-use laws. Journal of the American Medical Association
264:2395-99.
Stellwagen, E. A., and R. L. Goodrich. 2000. Forecast Pro. Belmont, MA: Business Forecast
Systems.
Tiao, G. C., and W. W. S. Wei. 1976. Effect of temporal aggregation on the dynamic relationship
of two time series variables. Biometrika 63:513-23.
Watson, G. S., P. L. Zador, and A. Wilks. 1980. The repeal of helmet use laws and increased
motorcycle mortality in the United States, 1975-1978. American Journal of Public Health
70:579-85.
. 1981. Helmet use, helmet use laws, and motorcyclist fatalities. American Journal of
Public Health 71:297-300.
Weiss, A. A. 1992. The effects of helmet use on the severity of head injuries in motorcycle acci-
dents. Journal of the American Statistical Association 87:48-56.
Lisa Stolzenberg is an associate professor in the School of Policy and Management at Florida
International University. Her work has appeared in a variety of journals, including the Ameri-
can Sociological Review, Social Forces, Social Problems, and Criminology. Her research has
been supported by the National Institute of Justice and several other funding agencies.
Stewart J. D'Alessio is an associate professor in the School of Policy and Management at
Florida International University. His current research focuses on economic inequality and
crime.
150
EVALUATION REVIEW / APRIL 2003