ERIC EJ767618: Impact of a Comprehensive Safety Program on Bicycle Helmet Use among Middle-School Children

2007·0.4 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2007, 40, 239–247 NUMBER2 (SUMMER2007)
IMPACT OF A COMPREHENSIVE SAFETY PROGRAM ON BICYCLE
HELMET USE AMONG MIDDLE-SCHOOL CHILDREN
RON VAN HOUTEN
WESTERNMICHIGANUNIVERSITY
AND
JOY VAN HOUTEN AND J. E. LOUIS MALENFANT
CENTERFOREDUCATIONANDRESEARCHINSAFETY
Abicyclehelmetprogramwasevaluatedinthreemiddleschoolsusingamultiplebaselineacross
schools design. Two of the three schools had histories of enforcement of helmet use. During
baselinemanystudentsridingtheirbikestoandfromschooldidnotweartheirhelmetsorwore
them incorrectly. A program that consisted of peer data collection of correct helmet use,
education on how to wear a bicycle helmet correctly, peer goal setting, public posting of the
percentageofcorrecthelmetuse,andsharedreinforcers,allofwhichwereimplementedbythe
school resource officer, increased afternoon helmet use and afternoon correct helmet use in all
threeschools.Probedatacollectedadistancefromallthreeschoolsindicatedthatstudentsdid
notremovetheirhelmetsoncetheywerenolongerincloseproximitytotheschool,andprobe
datacollectedinthemorningattwooftheschoolsshowedthatthebehaviorchangetransferred
tothe morning.
DESCRIPTORS: bicycle helmet use, goal setting, enforcement, public posting, peer data
monitoring,bicycle safety
_______________________________________________________________________________
Itisestimated that70% ofchildrenaged5 to injuredwereundertheageof16.Thesedataalso
14 yearsridebicycles(Sacks,Kresnow,Houston, show that the 10- to 15-year-old age group had
& Russell, 1996). Although riding bicycles has the highest fatality and injury rates, with fatality
significanthealthbenefits,bicyclesareassociated rates54%andinjuryrates140%higherthanthe
with more injuries than any other consumer average rate for all bicyclists.
product except motorized vehicles (Hoover- Head injuries account for 60% of bicycle-
Wilson, Baker, Teret, Shock, & Garbarino, related deaths and more than two thirds of
1991).DatafromtheNationalHighwayTraffic bicycle-related hospital admissions (Brewer et
Safety Administration (2005) indicate that 725 al., 1995). Bicycle helmets have been docu-
bicyclistswerekilledand41,000wereinjuredin mented to reduce the risk of head injury by
2004,and21%ofthosekilledand32%ofthose 85% and brain injury by 88% (Thompson,
Rivara, & Thompson, 1989). Other studies
have shown that the introduction of bicycle
This research was supported by a NHTSA contract
from the Florida Department of Transportation Safety helmet legislation is associated with both
Office. We thank school resource officers Les Harris and increased helmet use and reductions in bi-
Ken Jamison of the Saint Petersburg Police Department
cycle-related deaths and injuries (Graitcer,
and Robert Tillotson of the Lee County Sheriff’s Office
fortheir assistance incarrying outthis study. Kellerman, & Christoffel, 1995; Mackinan &
Address correspondence to Ron Van Houten, Psychol- Medenorp, 1994; MacPherson, To, & Mac-
ogy Department, 3700 Wood Hall, Western Michigan
arthur, 2002). These data suggest that helmet
University, Kalamazoo, Michigan 49008 (e-mail: ron.
[email protected]). A copy of the DVD explaining use should be a major intervention target for
howtocarryouttheprogrammaybeobtainedfromJanet middle-school children.
Degner at the Florida Technology Transfer Center, P.O.
The most commonly employed counter-
Box116587,Gainesville, Florida32611.
doi:10.1901/jaba.2007.62-06 measures to increase helmet use include legis-
239
240 RON VAN HOUTEN et al.
lation, helmet giveaway programs, and educa- bicycles. Two schools were located in Saint
tionprograms.Areviewevaluatingtheeffectsof Petersburg, and the third school was located in
helmet legislation has documented variable Bonita Springs. One of the schools (Riviera
increases in helmet use following the introduc- Middle School) had been exposed to the
tion of helmet laws (Karkhaneh, Kalengal, program the previous year as the pilot school.
Hagel, & Rowe, 2006). Another strategy for The two remaining schools (Meadowlawn
increasing helmet use is the helmet giveaway Middle School and Bonita Springs Middle
program.Twostudies haveevaluatedtheeffects School) had no prior experience with the
of this program on helmet use. One found that program.
an extensive giveaway program failed to pro-
Departure Measures
duce an increase in helmet use by low-income
students (Parkin et al., 1995). The other found Adult observers. The first and second authors,
that a giveaway program increased helmet use with the assistance of a research assistant,
byelementary-schoolstudentsfrom3%to38% recorded several behaviors as the children left
but had no effect on helmet use by middle- school at the end of each day. Observers seated
school students (Logan et al., 1998). Jointly in vehicles parked at the exit next to the bicycle
these studies suggest that giving middle-school compound recorded helmet use at the end of
students helmets alone will not increase their the school day at the two Saint Petersburg
helmet use. Parkin et al. (1993) evaluated the schools. The school resource officer collected
effectsofaneducationalprogramonhelmetuse data at a location with a view of the students
at18Canadianschools.Theyfoundanincrease leaving the bicycle compound at Bonita
in helmet use from 3.4% to 16% following the Springs.
introduction of the educational intervention. Observersrecordedwhetherthehelmetwason
Behavior strategies to change transportation the student’s head and whether the helmet was
safety behaviors related to the use of safety oncorrectly.Tobescoredascorrectlyworn,the
equipment have typically focused on safety-belt helmet had to be buckled snugly (the loop
use. Some interventions that have increased formed by the buckle must not form a loop the
safety-belt use are posted feedback (Malenfant, observer estimated would accommodate more
Wells, Van Houten, & Williams, 1996), thanafewfingers),andthehelmetneededtobe
enforcement (Van Houten, Malenfant, & Roli- level (if the forehead was exposed because the
der,1985),peermonitoring(Cooper&Phillips, helmetwastippedupinthefront,orthebackof
2004), and incentives and rewards (Geller, the head was exposed because the helmet was
raked forward, it was not scored as level).
Kalsher, Rudd, & Lehman, 1989). We have
not found any systematic behavioral study to The percentage of students wearing a bicycle
increase middle-school bicycle helmet use. The helmet each day was computed by dividing the
purpose of the present study was to evaluate number of children wearing a helmet by the
a treatment package that included some of the total number of children bicycling. The per-
above-mentionedbehavioralelementstoincrease centageofhelmetsworncorrectlywascalculated
by dividing the number of children wearing
bicycle helmet use by middle-school children
a helmet correctly by the total number of
who ride their bicycles to school.
children bicycling. The first author trained all
adult observers by illustrating each of the
METHOD
possible response outcomes for correct helmet
Participants and Setting use.
Participants were children from three Florida Peer observers. The officer at each school
schools who commuted to school on their selected students to observe and record bicycle
HELMET SAFETY PROGRAM 241
helmet use. Students recorded helmet use only Interobserver Agreement
during the treatment condition at Riviera and A measure of interobserver agreement was
Meadowlawn, whereas students at Bonita obtained for three to five sessions during each
Springs collected data throughout the experi- condition of the experiment at each site.
ment. Five students were selected at Riviera, 8 Interobserver agreement data for distance
students were selected at Meadowlawn, and 8 probes were collected for half the distance
students were selected at Bonita Springs. The measures at Riviera and Meadowlawn. No
officer selected students who he believed would interobserver agreement data were collected
be reliable and responsible to observe and score for probe measures at Bonita Springs or for
helmet use. morning probe measures.
During the treatment condition, 1 or 2 peer
observers were assigned to observe helmet use Experimental Design
each day. Helmet use was observed and A multiple baseline across schools design was
recorded the same way by student observers as employed. After collecting baseline data at all
itwasbytheadultobservers.Theofficertrained three schools, the treatment condition was
observers to record helmet use by demonstrat- introduced at Bonita Springs, while the other
ingexamplesofcorrectandincorrecthelmetuse two schools remained in the baseline condition.
and showing the children a video on correct OncecorrecthelmetusehadstabilizedatBonita
helmet use. Student observers were then taken Springs, the treatment condition was intro-
outside as a group on the 1st day of the duced at Riviera, while Meadowlawn remained
intervention and observed and recorded helmet in the baseline condition. Once helmet use had
useofstudentsdepartingschoolonbicycles;the stabilized at Riviera, the treatment condition
officer then reviewed whether each helmet was was introduced at Meadowlawn.
scored correctly or incorrectly. Students were Baseline. During baseline, the officers at
trained to use the same definitions for target Riviera and Meadowlawn periodically wrote
behaviors employed by adult observers. tickets for failure to wear a helmet as they had
done during previous years. Tickets were not
Probe Measures written at Bonita Springs (Florida law requires
Adult observers collected two probe mea- all children and youth under the age of 16 to
sures. The first, collected at two schools, wear a bicycle helmet).
involved children riding to school at a specified Helmet-useprogram.Atthestartofthehelmet
time (morning). The second, collected at all program,theofficercalledallstudentswhorode
three schools, involved children riding home their bicycles to school to an assembly. During
from school at specified distances (approxi- this assembly, the officer explained the impor-
mately 0.5 mile from the school). The distance tanceofhelmetuseandreviewedthereasonsfor
measure was included to determine if the wearingabicyclehelmet.Healsoshowedavideo
children removed their helmets after leaving onthecorrectfittingofbicyclehelmets.Students
theschoolarea.Bothmeasureswereincludedto were told that peers wouldbe collecting data on
assess whether the treatment generalized over helmetuse,andthepercentageofcorrecthelmet
time and was maintained over distance. Dis- useeachweekwouldbepostedonachartinthe
tance probe data were collected by research cafeteria along with the record, and another
assistants who sat in a car parked along the displayed at the administrative office at the
route (to decrease the likelihood that students school entrance. They were then asked to select
would notice that they were being observed). a helmet-use goal for the school.
Morning data were scored by adults parked Goal setting was done by consensus, and the
across from the school. officer asked if they could do better if someone
242 RON VAN HOUTEN et al.
initially set a low goal. The goal set at Riviera Data on the mean number of riders were
and Meadowlawn was 80%, and the goal set at computed each day for baseline and the
Bonita Springs was 70%. Students were further treatment condition. At Bonita Springs the
instructed that if they met the goal before the number of riders per day averaged 30 during
endoftheschoolyeartheywouldcelebratetheir baseline and 30 during treatment. At Riviera
success with a party with pizza, ice cream, soda, thenumberofridersperdayaveraged25during
and small prizes. They were also told that baseline and 23 during treatment. At Meadow-
abicyclehadbeendonatedandthatitwouldbe lawn the number of riders per day averaged 34
raffled off at the party. during baseline and 37 during treatment. The
At the end of the meeting, free helmets were baseline mean level of helmet use at Riviera the
given to students who did not have helmets (six previous year was 76%, and the mean level of
to eight helmets were distributed at each of the helmet use during the treatment condition the
schools), and the officer fitted helmets forthose previous year was 98%.
students. The peer data-collection procedure
Probes
was initiated during the afternoon of the
assembly, and the charts were put up showing Distance probe data were collected approx-
the baseline mean level of correct helmet use. imately 0.5 mile from the school at all three
Each week the officer met briefly with the peer schools. The officer suggested locations that
observers to collect their data sheets. After this a majority of the students needed to pass on
meeting,thepercentagesdisplayedonthecharts theirwayhomethatmetcriteriaforthedistance
were changed based on data collected by the probe. School departure helmet use during
student observers. these probes averaged 26% at Bonita Springs,
72%atRiviera,and68%atMeadowlawn.After
the introduction of the treatment condition,
RESULTS
helmet use increased to 36% at Bonita Springs,
School Departure Helmet Use 100% at Riviera, and 100% at Meadowlawn.
The percentage of students wearing bike During baseline, correct helmet use at probe
helmets when departing from the three middle locations averaged 15% at Bonita Springs, 30%
schools at the end of the school day ispresented atRiviera,and30%atMeadowlawn.Following
inFigure 1.Helmetuseaveraged14%atBonita treatment, correct helmet use at the distance
Springs, 82% at Riviera, and 52% at Meadow- probe sites increased to 32% at Bonita Springs,
lawn during baseline. During treatment, helmet 70% at Riviera, and 80% at Meadowlawn.
use increased to 45% at Bonita Springs, 98% at Note that probe data could differ from data
Riviera, and 95% at Meadowlawn. collectedattheschoolthesamedaybecausenot
The percentage of correct helmet use by all children needed to pass the probe site on
students departing from the three middle their way home.
schoolsattheendoftheschooldayispresented The percentage of morning helmet use
in Figure 2. During baseline the percentage of during the single morning probe at Bonita
correct helmet use at departure averaged 9% at Springs was 23%, and correct helmet use was
Bonita Springs, 64% at Riviera, and 30% at 17%.Duringthetreatmentcondition,morning
Meadowlawn. It should be noted that correct helmetwas48%,andcorrectusewas35%.The
helmet use showed an upward trend during percentages of helmet use during two morning
baseline at Meadowlawn. During treatment, probes at Meadowlawn were 81% and 83%
correct helmet use increased to 40% at Bonita during baseline and 97% during a single
Springs, 80% at Riviera, and 78% at Meadow- morning probe during treatment. The percent-
lawn. ages of correct helmet use during two morning
HELMET SAFETY PROGRAM 243
Figure 1. The percentage of students wearing bicycle helmets at all three middle schools. The last 4 days of the
helmet program at Riviera were after the party, as was the last day of the helmet program at Meadowlawn. Gray
diamonds show the percentage of helmet use during the distance probes taken after school. Gray triangles show the
percentageofhelmet use during the morningprobes.
244 RON VAN HOUTEN et al.
Figure2. Thepercentageofstudentswearingbicyclehelmetscorrectlyatallthreemiddleschools.Thelast4 daysof
the helmet program at Riviera were after the party, as was the last day of the helmet program at Meadowlawn. Gray
diamonds show the percentage of helmet use during the distance probes taken after school. Gray triangles show the
percentageof helmet use during the morningprobes.
HELMET SAFETY PROGRAM 245
probes at Meadowlawn were 31% and 50% an increasing trend at Meadowlawn, although
during baseline and 87% during treatment. they were somewhat stable over the last 5 days.
Unfortunately,itwasnecessarytointroducethe
Peer Data Collection treatment at this site before this trend stabilized
It was not possible to make an observation- because the end of the school year was rapidly
by-observation comparison of student-recorded approaching.
data with adult-recorded data because the The results also indicated that the schools
students and adults scored behavior from two with a baseline history of enforcement for
different vantage points. However, it is possible nonuse of helmets had higher baseline levels of
to compare daily mean helmet use recorded by helmet use and that the intervention produced
student observers and adult observers. At near-perfect levels of helmet use and relatively
Riviera student-recorded data averaged 3% highlevelsofcorrecthelmetuseatbothofthese
higher for helmet use and 16% higher for
schools.Theprogramalonewasabletoproduce
correct helmet use. At Meadowlawn student-
a moderate level of helmet use and correct
recorded data were 4% higher for helmet use,
helmet use similar to baseline levels at the two
and for correct helmet use were 6% higher.
schools with a history of bicycle helmet
enforcement. The number of citations written
Interobserver Agreement
at Meadowlawn for violating the helmet law
Interobserver agreement was calculated for
averaged 16 per month prior to treatment, but
helmet use and correct helmet use (data
none were written after the treatment was
collected by adults) for cyclists departing school
introduced.
and for probe data by dividing the number of
At Riviera, eight citations were written per
agreements on the occurrence of the behavior
month during baseline, and only two were
by the number of agreements on occurrence
issued during treatment, with both issued
plusthenumberofdisagreements.Interobserver
during the 1st week. Because the program was
agreement on the occurrence of helmet use at
more effective at the two schools with a history
Bonita Springs averaged 98% (range, 91% to
of enforcement, fewer citations were written at
100%), and interobserver agreement on the
these two schools following the introduction of
occurrence of correct helmet use averaged 91%
the program.
(range, 67% to 100%).
Although the threat of citations likely had
Interobserver agreement on helmet use was
some effect on behavior, many bicyclists re-
100% at Riviera, and agreement on correct
ceived multiple citations. The officers said the
helmet use averaged 93% (range, 84% to
students rarely paid the fines, even though they
100%). At Meadowlawn, agreement on helmet
were informed that the Department of Motor
use averaged 98% (range, 87% to 100%), and
Vehicles would require payment of all unpaid
agreement on correct helmet use averaged 93%
fines before they could get a driver’s license.
(range, 67% to 100%).
Thedifferenceintreatmentefficacybetweenthe
schools with and without a history of enforce-
DISCUSSION
ment may have been the resultof rule-governed
The results of this experiment showed that behaviorinfluencingasubsetofstudents,orthe
theinterventionwasassociatedwithaconsistent aversiveness of being stopped by the officer.
increase in bicycle helmet use and correct Although it is possible that the program
bicycle helmet use at all three middle schools. improved helmet use by discouraging students
Data for helmet use were stable for all three from riding their bicycles to school, this seems
schools, but data for correct helmet use showed highly unlikely because the number of children
246 RON VAN HOUTEN et al.
riding to school remained constant over the and that the effects of the treatment had
course of the experiment. The observers also transferred to times when the peer observers
notedthe same students riding their bicycles on were not present. Because the adults who
a regular basis throughout the study period. recorded probe data did so from parked cars,
At Riviera, baseline measures used in this it is unlikely that the students were aware of
study were collected 1 year later, and they being watched. Another feature of this study
showedapartialreversal oftheeffectsproduced that has not been implemented in other studies
by treatment. One reason for the decline may with middle-school studentswastheuseof peer
have been that some components of the observers. This made the treatment more cost
treatment were reversible. Another explanation effective and may have contributed to the
is the turnover of a third of the students from efficacy of the program.
year to year. We evaluated an intervention that contained
It is also interesting that children were multiplecomponentsbecausewecouldnotfind
frequently observed riding their bicycles with prior behavioral research addressing this prob-
a helmet in their possession, usually attached to lem and wanted to increase the probability of
the handlebars or backpack. On several occa- obtaining a successful outcome. We were
sions observers recorded whether children had unable to perform a component analysis
a helmet in their possession that was not worn. because of the limited time available before
The incidence of this phenomenon averaged the end of the school year. One element of the
20% during baseline. However, no reliability package was the helmet giveaway to students
data were recorded for this measure because who said they did not have a helmet. Although
only one of observers who collected interob- studies have demonstrated that helmet give-
server agreement data recorded it. It is likely aways alone or in conjunction with traditional
that children took their helmets with them educational programs are ineffective, it is
because of parental pressure to wear a helmet. possible that this element contributed to the
This explanation is consistent with the some- efficacy of the behavioral intervention even
what higher percentage of helmet use observed thoughithasbeendocumentedtobeineffective
during morning probe measures than in the in isolation. We recommend that future re-
afternoon. search address the question of how much each
Moststudentsscoredasnotwearingahelmet component contributes to the success of this
correctly did not have it buckled or had the intervention package.
strap too loose. Of the smaller proportion that It appears from this study that helmet use is
didnothavethehelmetlevel,inalmostallcases a much easier target behavior to change than
the helmet was higher in the front (leaving the correct helmet use. Perhaps training methods
forehead exposed). Based on informal observa- should be used that require more active
tion and analysis it appeared that after the responding on the part of the trainees. The
intervention was introduced, helmets scored as FloridaDepartmentofTransportation(FDOT)
not correctly worn were most often loose rather prepared a program package consisting of an
than unbuckled. Further research should be interactive DVD and other materials needed to
conducted to determine why students did not carry out this program. The dissemination of
fasten their helmets more securely. this program is now being funded by the
Several features added to the strength of this FDOT Safety Office and is being implemented
study. First, the use of distance probes and through the Florida Technology Transfer
morning probes confirmed that students were Center at the University of Florida with the
not taking off their helmets on the way home, assistance of the Florida Bicycle Association.
HELMET SAFETY PROGRAM 247
REFERENCES MacPherson,A.K.,To,T.M.,&Macarthur,C.(2002).
Impact of mandatory helmet legislation on bicycle-
Brewer, R. D., Fenley, M. A., Protzel, P. I., Sacks, J. J., related head injuries in children: A population based
Thornton, T. N., & Nowak, N. D. (1995). Injury study. Pediatrics,110, 60–65.
controlrecommendations:Bicyclehelmets.Morbidity Malenfant,L.,Wells,J.K.,VanHouten,R.,&Williams,
andMortality Weekly Report,44, 1–17. A.F.(1996).Theuseoffeedbacktoincreaseobserved
Cooper, M. D., & Phillips, R. A. (2004). Explanatory daytimeseatbeltuseintwocitiesinNorthCarolina.
analysis of the safety climate and safety behavior Accident AnalysisandPrevention,28,771–777.
relationship. Journal of Safety Research, 35, 497– National Traffic Safety Administration. (2005). Traffic
512. safety facts 2004 data. Washington, DC: Center for
Geller,E.S.,Kalsher, M.J.,Rudd,J. R.,& Lehman,G. Statistics andAnalysis.
R. (1989). Promoting safety belt use on a university Parkin,P.C.,Hu,X.,Spence,L.J.,Kranz,K.E.,Shortt,L.
campus:Anintegrationofcommitmentandincentive G.,&Wesson,D.E.(1995).Evaluationofasubsidy
strategies. Journal of Applied Social Psychology, 19(1), programtoincreasebicyclehelmetusebychildrenof
3–19. low-incomefamilies.Pediatrics,96,283–287.
Graitcer, P. L., Kellerman, A. L., & Christoffel, T. A. Parkin,P.C.,Spence,L.J.,Hu,X.,Kranz,K.E.,Shortt,
(1995). A review of educational and legislative L. G., & Wesson, D. E. (1993). Evaluation of
strategies to promote bicycle helmets. Injury Pre- a promotional strategy toincrease bicycle helmet use
vention,1,122–129. bychildren. Pediatrics,91, 772–777.
Hoover-Wilson,M.H.,Baker,S.P.,Teret,S.P.,Shock, Sacks, J. J., Kresnow, M., Houston, B., & Russell, J.
S.,&Garbarino,J.(1991).Savingchildren:Aguideto (1996).BicyclehelmetuseamongAmericanchildren.
injuryprevention.New York:Oxford Press. Injury Prevention,2, 258–262.
Karkhaneh,M.,Kalengal,J.C.,Hagel,B.E.,&Rowe,B. Thompson, R. S., Rivara, F. P., & Thompson, D. C.
H. (2006). Effectiveness of bicycle helmet legislation (1989). A case-control study of the effectiveness of
to increase helmet use: A systematic review. Injury bicycle safety helmets. The New England Journal of
Prevention,12, 76–82. Medicine,320, 1361–1367.
Logan, P., Leadbetter, S., Gibson, R. E., Schieber, R., Van Houten, R., Malenfant, L., & Rolider, A. (1985).
Branche, C., Bender, P., et al. (1998). Evaluation of Increasing driver yielding and pedestrian signalling
a bicycle helmet giveaway program in Texas. throughtheuseoffeedback,promptingandenforce-
Pediatrics,101,578–582. mentprocedures.JournalofAppliedBehaviorAnalysis,
Mackinan,M.L.,&Medenorp,S.V.(1994).Association 18,103–115.
between bicycle helmet legislation, bicycle safety
education and use of bicycle helmets in children. Received April 18,2006
Archives of Pediatric Adolescent Medicine, 148, Final acceptanceNovember 1, 2006
255–259. Action Editor,KenSilverman