38 American Journal of Health Promotion

T H E S C I E N C E O F H E A L T H P R O M O T I O N

Building the Methods

Reliability and Validity of Two InstrumentsDesigned to Assess the Walking and BicyclingSuitability of Sidewalks and RoadsJames Emery, MPH; Carolyn Crump, PhD; Philip Bors, MPH

Abstract

Purpose. Public health professionals hypothesize that when community environmentsprovide suitable walking and bicycling conditions, community members will be more active.Measurement indicators and assessment instruments are needed to evaluate suitability.This study determined the reliability and validity of two instruments to assess the suitabili-ty of sidewalks for walking and roads for bicycling.

Methods. Two data collectors used walking and bicycling suitability assessment instru-ments to collect data on 31 road segments. In addition, three transportation experts used a7-point Likert response system to subjectively evaluate walking and bicycling conditions forthe same segments. Intraclass correlations determined the reliability of each assessment in-strument and the reliability of the Likert response system. Pearson correlations (researchstaff assessments with expert assessments) were calculated to determine the criterion-relatedvalidity of the suitability measures.

Results. Intercoder reliability (intraclass) correlations for the walking and bicycling as-sessment instruments were r 5 .79 and .90, respectively. Intercoder reliability of the ex-perts’ Likert response system was r 5 .73 for the walking form and r 5 .77 for the bicy-cling form. Criterion-related validity (Pearson) correlations for the walking and bicyclingassessment instruments were r 5 .58 and .62, respectively.

Conclusion. Although some variables have lower reliability and validity than is ideal,the walking and bicycling suitability assessment instruments appear promising as instru-ments for community members and professionals to systematically assess key aspects of thephysical environment. (Am J Health Promot 2003;18[1]:38–46.)

Key Words: Physical Activity, Walking, Pedestrian, Bicycling, Community De-sign, Active Living, Active Transportation, Prevention Research

James Emery, MPH, is a research associate in the School of Public Health, the University ofNorth Carolina at Chapel Hill, Chapel Hill, North Carolina. Carolyn Crump, PhD, is aresearch assistant professor in the School of Public Health, the University of North Carolinaat Chapel Hill, Chapel Hill, North Carolina. Philip Bors, MPH, is a project officer for Ac-tive Living by Design, School of Public Health, the University of North Carolina at ChapelHill, Chapel Hill, North Carolina.

Send reprint requests to James Emery, MPH, Research Associate, Department of Health Behaviorand Health Education, School of Public Health, The University of North Carolina at Chapel Hill,CB# 7506, Chapel Hill, NC 27599-7506.

This manuscript was submitted October 16, 2002; revisions were requested December 13, 2002; the manuscript wasaccepted for publication April 23, 2003.

Copyright q 2003 by American Journal of Health Promotion, Inc.0890-1171/03/$5.00 1 0

INTRODUCTION

The 1996 U.S. Surgeon General’sreport recommends accumulating 30minutes of regular moderate physicalactivity on most days of the week,since moderate levels of physical ac-tivity can help reduce the risk forchronic disease.1 However, almost60% of adults are still not sufficientlyactive to achieve health benefits, andclose to 30% report no leisure-timephysical activity in the past month.2Moderate physical activity may bemore easily initiated and sustainedwhen activities can be incorporatedinto daily routines, such as walkingand bicycling for transportation andleisure. The physical environmentsurrounding and connecting residen-tial, school, and business locationsmay determine the extent to whichcommunity members can incorporatewalking and bicycling into their dailyroutines. Public health proponentshypothesize that when communityenvironments are designed to pro-vide suitable facilities for walking andbicycling, community members willbe more active.3,4

Within the last decade, transporta-tion and public health program goalshave focused attention on walkingand bicycling behavior and have be-gun studying the physical environ-ment that supports or challengesthese healthy behaviors. In 1994 theU.S. Department of Transportationestablished a goal to double adultwalking and bicycling trips. The re-cent Task Force on Community Pre-ventive Services recommended in-creasing environmental opportunities

September/October 2003, Vol. 18, No. 1 39

for physical activity.5 Healthy People2010 enhanced the Healthy People2000 community-environment objec-tive addressing community walkingand bicycling opportunities and add-ed objectives 22-14 and 22-15 to in-crease the number of walking and bi-cycling trips. Interdisciplinary profes-sional collaborations have increasedresearch into how the built environ-ment influences people’s travel deci-sions.6–13

Transportation planners and engi-neers have been studying pedestrianand bicycling environmental designissues since the 1970s.9–11,14–18 Howev-er, it was not until the 1980s that in-struments were developed to assessthe ‘‘suitability’’ of sidewalks androads for walking and bicycling. Mostinstruments focused on urban streetsand rural roads, although a few as-sessed the suitability of trails for walk-ing and bicycling.14,19–21 Suitability as-sessment instruments for quantitativedata collection generally group thedesign variables into five basic trans-portation environment categories:traffic volume, traffic speed, facilitywidth, surface factors, and locationfactors.14,22–35 To our knowledge onlytwo studies have developed auditmethods for use by community resi-dents.28,36

Public interest in more ‘‘livable’’communities designed to promotewalking and bicycling has encour-aged new alliances between profes-sionals from various disciplines andcommunity members interested inimproving the opportunities for phys-ical activity.37–43 Community memberscan advocate more effectively for in-creased opportunities for walkingand bicycling if they have tools forassessing the walkability and bike-abil-ity of their neighborhood, leisure,school, retail, and work environ-ments. Despite the increased promo-tion of walking and bicycling behav-ior and the need for supportive phys-ical environments, ‘‘gold standard’’tools with known psychometric prop-erties for assessing the suitability ofthose environments for physical activ-ity do not yet exist. Valid indicatorsto assess the progress of communitysupport for the recommended behav-iors or to evaluate built environmentsuitability for compliance with recom-

mendations for activity have yet to beidentified.2,44–48

In previous research we identified22 potential bicycling suitability as-sessment instruments in the profes-sional and internet literature butfound no appropriate pedestrian suit-ability assessment instruments.49

Therefore, the previous study con-cluded by developing an instrumentfor assessing the walking suitability ofsidewalks (Appendix 1) and analyzedthe 22 methods for assessing the bi-cycling suitability of roads to deter-mine the most desirable assessmenttool for future community use (Ap-pendix 2).

The purpose of the current studywas to test the reliability and validityof the two assessment instruments.

METHODS

Assessment InstrumentsEach assessment instrument is one

page in length and is completed bytraveling a road segment under 2miles in length. Overall walking orbicycling conditions from both sidesof the road are observed. The ratermarks appropriate response catego-ries for each variable and then uses asimple algorithm to calculate the fi-nal suitability score. For this project,the unit of analysis was the road seg-ment.

The walking suitability assessmentinstrument (Appendix 1) is com-prised of 11 variables that measurebasic transportation environmentcharacteristics including traffic vol-ume, traffic speed, sidewalk and buff-er width, surface quality, and sup-portive amenities such as curb rampsand street lighting.49 Recorded nu-meric values are summed to producethe overall walking suitability scorefor that road segment. If no sidewalkexists for the segment, the instru-ment produces a minimum score of99 to denote the lack of pedestrianfacilities. The one-page form also in-cludes dichotomous questions aboutintersection characteristics encoun-tered while assessing the sidewalksegment and provides space for re-cording isolated problem spots. Theintersection information is not partof the final suitability score, buthelps to identify design problems.

The Eddy bicycling suitability as-sessment instrument (Appendix 2)includes 27 variables grouped intothree main categories on a one-pageform (e.g., general road factors,pavement factors, and location fac-tors).30 The general road factors in-clude traffic volume, number ofthrough lanes, traffic speed, width ofthe outside lane, and width of anystriped bicycle lane. An algorithmgraphic at the bottom of the formprovides clearly designated boxes torecord several values from the scor-ing sheet, and then provides the for-mula for calculating the final bicy-cling suitability score. In this study,the Eddy instrument was adapted bydeveloping a ‘‘no’’ check box foreach dichotomous variable to ensurethat data collectors had to record aresponse for every variable and bymoving the algorithm graphic to thebottom of the page.

Data CollectionIn order to test the reliability and

validity of the two assessment instru-ments, 31 urban and rural road seg-ments ranging in distance from .1 to2.0 miles in length were identifiedwithin 10 miles of the University ofNorth Carolina at Chapel Hill. Roadsegments were purposefully selectedto equally represent major and mi-nor arterial highways, neighborhoodcollectors, and local streets. Two re-search staff members used the instru-ments to assess the sampled road seg-ments. One person had extensive ex-perience using both instruments inprior research; the other had neverused either form. Based on earlier as-sessment experience, the researchstaff decided to complete the assess-ments using a motor vehicle, ratherthan walking or bicycling the roadsegments. The assessment staff trav-eled the sampled road segments to-gether in the same motor vehiclefrom 8:00 A.M. until 7:00 P.M. overseveral nonholiday weekdays duringfair weather in the autumn of 2001.From within the car, they drove theroad segment once while observingthe environmental characteristics ofinterest. Parking the vehicle off road,the researchers independently com-pleted the assessment forms. Theythen traveled each road segment

40 American Journal of Health Promotion

again to observe any characteristicsnecessary to complete the assessmentforms. The assessors left the vehicleat least once per segment in order tomeasure the width of the sidewalk ortravel lane, observe or ‘‘feel’’ thegrade of the hill, or experience thephysical effects (e.g., the surface con-dition) of the road segment charac-teristics. They used a steel measuringtape to determine the width of theoutside travel lane, the sidewalk, andthe buffer area between the sidewalkand road.

Because no gold standard tools ex-ist for criterion-validity comparisons,we created a proxy standard usingthe opinions of three experts to in-dependently assess the same roadsegments. Two were transportationengineers with extensive research ex-perience in bicycle and pedestriandesign, and one was a county-levelcoordinator of bicycle and pedestriantransportation planning. The trans-portation experts subjectively assessedthe same walking and bicycling suit-ability characteristics using new formsdesigned specifically for the project.The two forms did not restrict theexperts’ assessment to specific designprompts. Instead, each variable wasmeasured on a 7-point Likert scaleranging from ‘‘conducive’’ to ‘‘inhib-itive’’ with a neutral midpoint.50

Due to an oversight, the transpor-tation experts’ walking suitability didnot include traffic volume, speed,and number of lanes. Therefore, inorder to calculate the criterion-validi-ty correlations for those variables, weused the values from the transporta-tion experts’ bicycling suitability as-sessments as proxy measures for thewalking suitability form. We assumedthat the validity correlations wouldbe similar between walking and bicy-cling, although possibly weaker forthe walking assessment because pe-destrians do not share the travel lanewith motor vehicles.

AnalysisThe two assessment instruments

each applied an algorithm that com-bines the assessed values to producea final suitability score for the roadsegment. Intercoder reliability corre-lations were calculated using SPSS10.1.0 for Windows (SPSS, Inc., Chi-

cago, Ill) to compute one-way analy-sis of variance intraclass correlationsfor each variable. The reliability andvalidity scores of the two researchstaff members and the three trans-portation experts were averaged foreach variable and for the overall suit-ability scores to provide an estimateof the ‘‘true’’ scores by each group.Criterion-related validity was definedas the strength of the association be-tween the research staff’s and thetransportation experts’ averaged suit-ability scores. Pearson product-mo-ment correlations were calculated todetermine the validity for both thewalking suitability assessment instru-ment and the bicycling suitability as-sessment instrument. Because only 17of the 31 road segments had side-walks, the walking instrument’s validi-ty could only be meaningfully esti-mated with the reduced sample ofroad segments. Validity testing of thebicycling instrument was performedon the entire sample of 31 segments.

RESULTS

Walking Suitability AssessmentFor the 17 road segments with

sidewalks, the intercoder reliabilitywas acceptable for both the overallwalking suitability assessment score (r5 .79) and the transportation ex-perts’ overall ratings (r 5 .73). Thecriterion-related validity correlationfor the overall walking suitability as-sessment score was r 5 .58.

Criterion-related validity for indi-vidual variables from the 17 segmentsranged from r 5 .15 to .84, with halfthe variables demonstrating validitycorrelations greater than r 5 .60(e.g., presence of sidewalk, presenceof streetlights, sidewalk material, traf-fic volume, and posted speed limit).Table 1 lists the variables in order ofvalidity correlations from most toleast robust.

Only two of the five variables withvalidity correlations greater than r 5.60 had low reliability. Sidewalk pres-ence and sidewalk material had lowintraclass correlations among thetransportation experts. The five vari-ables with weaker validity correlationsalso demonstrated weaker reliabilityby one or both assessment groups.They were buffer width, number of

through lanes, sidewalk condition,sidewalk width, and presence of curbramps.

Bicycling Suitability AssessmentIntercoder reliability was high for

the research staff’s overall bicyclingsuitability assessment score (r 5 .90)and acceptable for the transportationexperts (r 5 .77). The criterion-relat-ed validity correlation for the overallbicycling suitability score was r 5 .62.Reliability and validity correlationsfor individual variables ranged from r5 .004 to .82; over one third of thevariables demonstrated validity corre-lations greater than r 5 .60. Table 2lists the variables in order of validitycorrelations.

For both groups, variables thatwere more reliable tended to havehigher validity correlations. Objectiveobservations (e.g., measured width,posted speed limit, type of on-streetparking, street curb or shoulder, andtypes of turn lanes) were more reli-ably assessed than subjective assess-ments by the Eddy instrument. Reli-ability correlations for these variablesranged from r 5 .77 to r 5 1.0. Lessreliable variables were derived fromthe data collector’s subjective ratingof physical characteristics such assteepness of grade, quality of railroadcrossing, difficulty of intersectioncrossing, and density of drivewaysand intersections. Their correlationsranged from r 5 .03 to .58. Thetransportation experts disagreed withone another on the suitability of thethrough lanes, center turn lane,parking, curbing, drainage, andshoulders of the road segments.Their disagreement was reflected intheir low intraclass correlations andthe subsequent low validity for the as-sociated variables. This disagreementsuggests that the experts consider theimpact of complex design character-istics on walking and bicycling suit-ability in unique ways.

DISCUSSION

This study is a continuation of re-search meant to provide communitymembers and professionals with toolsto determine the suitability of localsidewalks and roads for walking andbicycling. Because no systematic data

September/October 2003, Vol. 18, No. 1 41

Table 1

Intercoder Reliability and Validity Results for Walking Suitability Assessment Instrument (n 5 17 Road Segments)

No. Walking Variable

Criterion-RelatedValidity (PearsonProduct-Moment)

Research Staffand Transpor-tation Experts

Reliability(Intraclass Correlation)

Research Staff(n 5 2)

TransportationExperts(n 5 3)

1234

Sidewalk presenceLightSidewalk materialAADT†

20.84**0.81**0.66**0.65**1

0.99***0.70***1.0***

Secondary data source

0.160.63***0.45**0.82*** (Bicycling proxy)

5 Speed† 0.63**1 1.0*** 0.80*** (Bicycling proxy)6789

10

Buffer widthNumber thru lanes†Sidewalk conditionSidewalk widthRamp

0.57*0.52*1

0.49*0.34

20.15

0.87***0.78***

20.100.53*0.89***

0.31*0.22* (Bicycling proxy)0.50***0.57***0.42**

Final walking suitability score 0.58* 0.79*** 0.73***

† Walking instrument did not originally include these three variables, which were included after the transportation experts had completed the Likert-response assessments. For this analysis, proxy scores were used from the transportation expert’s bicycle assessments.

* p , 0.05.** p , 0.01.*** p , 0.001.

collection tools previously existed forassessing the walking suitability ofsidewalks, this study advances thefield by determining that a new as-sessment instrument has acceptablereliability even if it is only moderate-ly valid. Confirmation in a largersample is necessary before criterionvalidity can be established. The Eddybicycling suitability assessment instru-ment does not require data collectorsto have engineering expertise or ap-ply logarithmic transformations dur-ing the final suitability algorithm. Itsease of use together with its strongreliability scores are pluses, althoughthe instrument’s validity is weakerthan desirable for final acceptance ofthe measure.

The modest reliability and validityof the overall walking suitabilityscores, the low validity of the bicy-cling instrument, and the low reli-ability or validity of some individualvariables on both instruments may bepartly explained by sampling and var-iable characteristics. The sample sizefor the bicycling suitability assess-ment analyses was 31 road segments,but only 17 road segments (with side-walks) for the walking suitability as-sessment analyses.

Although limited variability of thesampled suitability scores could beanother explanation for suppressedreliability correlations, our purposivesample results did not demonstratethose conditions. For the walking in-strument, the final group suitabilityscores ranged from 5 (good) to 13(poor) with a mean of 8.32 (fair).Approximately 35% of scores werepoor or worse, and 65% were fair orbetter. For the bicycling instrument,the final group suitability scoresranged from 1.75 (very good) to10.35 (very poor) with a mean of6.09 (poor). Approximately 60% ofbicycling scores were poor or worse,and 40% were fair or better. In gen-eral, our sample included slightlymore suitable walking conditions andless suitable bicycling conditions.These findings correspond with ourinformal observations of usual side-walk and road conditions.

Some variables in each instrumentrelied on more subjective assessmentsthat may have contributed to the low-er reliability scores. Attempts weremade to reduce subjectivity by pro-viding clear definitions of values(e.g., surface condition values weredefined using descriptions by the

Federal Highway Administration) andphotographs of examples.

Low reliability and validity couldalso be due to challenges experi-enced by the assessment staff. Duringdata collection, staff reported that itwas difficult to assess permanent suit-ability characteristics of the built en-vironment without being unduly in-fluenced by temporary conditions re-lated to maintenance. For example,debris accumulated on a sidewalkmay have impacted the rating forthat segment. Although codingguidelines attempted to anticipateand clarify these issues, a better ap-proach might be to modify the as-sessment forms to provide scoring fortemporary maintenance conditions.

Another challenge for both the re-search staff and the transportationexperts was to rate the suitability foran entire segment of sidewalk orroadway when variations existed with-in the segment, such as slight chang-es in sidewalk or lane width. Datacollectors were instructed to rate theoverall condition of the segment andsubdivide it into additional segmentsfor assessment when gross character-istics changed (e.g., speed limit,number of travel lanes, outside lane

42 American Journal of Health Promotion

Table 2

Intercoder Reliability and Validity Results for N. Eddy Bicycle Level of Service (n 5 31 road segments)

No. Bicycle Variable

Criterion-Related Validity(Pearson Product-Moment)

Research Staff andTransportation Experts

Reliability(Intraclass Correlation)

Research Staff(n 5 2)

Transportation Experts(n 5 3)

12345

SpeedBike laneFrequent curvesSevere gradesAnnual average daily traffic

0.82**0.82*0.82**0.76**0.74**

1.01.00.56***0.49**

Secondary data

0.80***0.47***0.73***0.68***†0.82***

6789

10

No sidewalksNumber of through lanesOutside lane widthRestricted sight distancePavement condition

0.70**0.60**

20.55**0.53**0.49**

0.74***0.87***1.00.53***0.58***

0.49***‡0.22*0.44***0.64***0.59***

1112131415

Numerous drivewaysDifficult crossingsCenter turn laneParallel parkingModerate grades

0.49**0.44*0.410.38

20.37*

0.58***0.171.01.00.08

0.52***0.58***§0.110.39***\0.68***†

1617181920

Angle parkingNumerous stopsRight turn lanePaved shoulderStorm drain grate

0.290.140.13

20.130.07

1.00.43**0.77***0.80***0.81***

0.39***\0.58***§0.65***0.39***¶0.39***¶

2122

Curb and gutterOne sidewalk only

0.0120.004

0.93***0.74***

0.39***¶0.49***‡

23 Physical median Unable to correlate because re-search staff mean value is con-stant

0.84*** 0.74***

24 Bike lane width N/A 1.0 Experts did not collect this variable25 Commercial land use N/A 0.62*** Experts did not collect this variable26 Rough railroad crossing N/A 20.03 Experts did not collect this variable27 Industrial land use N/A 0 Experts did not collect this variable

Final bicycle suitability score 0.62** 0.90*** 0.77***

† Planners assessed the suitability impact of the ‘‘grade,’’ which for validity testing was compared with both types of grade collected by researchstaff.

‡ Planners assessed the suitability impact of any ‘‘sidewalks,’’ which for validity testing was compared with both types of sidewalk presence col-lected by research staff.

§ Planners assessed the suitability impact of ‘‘Intersections,’’ which for validity testing was compared with both ‘‘difficult crossings’’ and ‘‘numerousstops’’ as collected by research staff.

\ Planners assessed the suitability impact of any ‘‘on-street parking,’’ which for validity testing was compared with both types of parking collectedby research staff.

¶ Planners assessed the suitability impact of ‘‘curb/gutter/drain or shoulder,’’ which for validity testing was compared with each of these character-istics collected by research staff.

* p , 0.05.** p , 0.01.*** p , 0.001.

width). The challenge arose whenchanges occurred in isolated spots.For example, a half-mile sidewalk 5feet in width was generally quite suit-able for pedestrian use. However,utility poles located in the center ofthe sidewalk could compromise thesegment’s suitability score. One ratermight conservatively rate the entiresidewalk width as 2 feet, since a

wheelchair could not easily use thesidewalk. Another rater might consid-er such poles to be isolated problemspots and note them separately. Ob-server training can provide guidancefor determining the ‘‘tipping point’’when the entire segment’s suitabilityrating for a particular variable is sig-nificantly changed by the cumulativeeffect of isolated problem spots.

Finally, low reliability and validityamong the transportation expertsmay have been partly due to theopen-ended structure of the form,which attempted to capture the ex-pertise of the transportation engi-neers and planners without prompt-ing them about what to consider im-portant for each variable. The ex-perts’ form provided a 7-point Likert

September/October 2003, Vol. 18, No. 1 43

scale for subjectively assessing thewalking and bicycling environment.Because the variables did not promptfor specific design details, respon-dents could consider multiple aspectsat once. For example, characteristicsof the travel lanes had very low reli-ability among the experts (e.g., num-ber of through lanes, center turnlane, and curb/gutter/drain/shoul-der). Comments written on theforms revealed that some experts as-sessed the impact not only of the in-tended design feature but of its phys-ical condition, placement, striping,and appropriateness, thus unequallyinfluencing their assessment.

Although the walking and bicy-cling suitability assessment instru-ments cannot substitute for carefulengineering of streets and intersec-tions to design improvements, theycan provide community membersand local planning staff with system-atic data to identify streets in need ofdesign improvements. These assess-ment tools can help people providesignificant input to public officials onways to improve the local built envi-ronment in order to increase activetransportation opportunities. Futureresearch should assess predictive va-lidity by considering the environmen-tal suitability for walking and bicy-cling compared with observabletransportation behavior.

LimitationsThis research was limited by its

use of a relatively small sample ofroad segments. This small sampleproduced several unstable reliabilityand validity correlations, especiallyfor the less prevalent variables (e.g.,industrial land-use and rough rail-road crossings), making it harder todraw firm conclusions about the in-struments’ utility.

The data on traffic volume wereannualized average daily counts.These do not provide a perspectiveon the specific travel stress inducedduring high-volume times of day,such as morning and evening com-mute hours. Hourly traffic volumevariability is hidden within the aver-age daily count, but it provides a sim-pler estimate of the impact of trafficvolume.

Although the data collectors trav-

eled each road segment two or moretimes, limitations prevented themfrom walking or bicycling the seg-ments, rather than observing themfrom an automobile. Some ruralroad segments were 2 miles inlength. The data collectors believedthat only two variables (grade andsurface condition) required assess-ment by disembarking from the vehi-cle at appropriate locations. If timeand funding permit, however, imple-menting assessments by foot and bi-cycle should improve reliability andvalidity. Another method could in-clude one community volunteer driv-ing while a second assessed the envi-ronmental conditions.

The research staff who assessedthe road segments were between 25and 40 years of age. A more diversegroup of assessors might capturewalking and bicycling suitability per-spectives of youth, older adults, andpeople with physical or visual disabili-ties. Broader representation of per-spectives might increase the general-izability of these tools across diversegroups of users.

A final limitation of these instru-ments is their failure to account forsupportive environmental characteris-tics (e.g., benches, water fountains,and even nearby destinations withshowers and changing facilities) andaesthetic characteristics (e.g., land-scaping, shady trees, scenery, or ab-sence of fumes). Future tools mightconsider including an assessment ofthese characteristics.

Future ResearchTransportation experts or engi-

neers may be highly trained observ-ers of pedestrian or bicycling physicalactivity conditions. However, their ob-servations may differ greatly fromwhat untrained ‘‘end-users’’ thinkabout the suitability of sidewalks androads. If one of the goals of publichealth is to increase walking and bi-cycling in the general population,the opinions of diverse communitymembers should also be incorporat-ed into the development of assess-ment instruments. Our study lackedsufficient funding to provide for re-cruiting and training communitymembers as observers. Future re-search could invite community mem-

bers to help identify additional vari-ables for study (e.g., supportive ame-nities and the aesthetics of the envi-ronment). Additional research couldalso expand the sample of ‘‘experts’’used for testing validity to includecommunity members who walk andbicycle. We are aware of three instru-ment-development studies that havebeen validated using communitymembers’ opinions about walkingand bicycling suitability characteris-tics.24,25,36 Future research could usesimilar methods to validate the in-struments in this study.

The Eddy bicycling assessment in-strument contains individual vari-ables that may be part of larger con-structs (e.g., types of parking, severityof grade, curb/shoulder zone ofpavement). Further research mightdetermine whether some of thesespecific constructs might be more re-liably or validly assessed as broadervariables.

Intersections present complicatedconditions for assessment. The walk-ing suitability instrument records ba-sic intersection data to highlight sig-nificant problem areas, but does notquantitatively compare intersections.Currently, the authors are collaborat-ing on formative research to developan intersection suitability assessmentinstrument to enhance collaborationbetween community members andplanning professionals, and provide ascientific basis for advocacy efforts bycommunity groups wishing to im-prove design.

CONCLUSION

The two assessment instrumentsdescribed in this article, despite theirmoderate psychometric properties,appear promising for promotingcommunity assessments of sidewalkand road suitability for walking andbicycling. With these instruments,community members and profession-als can compile data on the walkingand bicycling environment and usethose data to identify areas for im-provement that will ultimately createsupportive environments for morephysically active lifestyles. The assess-ment instruments and a projectguidebook are available for download

44 American Journal of Health Promotion

via the internet at http://www.unc.edu/;jemery/WABSA.

SO WHAT? Implications forHealth Promotion Practitionersand Researchers

This study suggests that two in-struments developed to assess thewalking and bicycling suitability oflocal sidewalks and roads are ac-ceptable for use by communitymembers and professionals. If con-firmed by others, these instru-ments will be useful in helpingcommunities perform simple, reli-able, and valid assessments of localenvironments for walking and bi-cycling. Professionals and trainedvolunteers can use the two instru-ments to assess road conditionsfor walking and bicycling andhighlight areas for improving thebuilt environment. Ideally, in col-laboration with local governmentofficials, health promotion practi-tioners can coordinate volunteer-led assessment projects that resultin advocacy for improvements.Health promotion researchers canuse the assessment instruments forcross-sectional and longitudinal re-search on the impact of the builtenvironment on physical activity.

Acknowledgments

The authors would like to thank the North CarolinaCardiovascular Health Program for its support of thisresearch, and the assistance of the following individu-als. For data collection, Rebecca Gordon, MPH; DavidHarkey, PE; William Hunter, MCE; Erica Steibelt; Ron-ald Svejkovsky, MUP; and Beth Timson, ASLA. Forstatistical advice, Robert DeVellis, PhD; and Chris Wie-sen, PhD. And for reviewing the article prior to publica-tion, Jo Anne Earp, ScD; Kelly Evenson, PhD; SaraHuston, PhD; Jane Stutts, PhD; two anonymous review-ers; and the editor.

References

1. US Department of Health and Human Servic-es. Physical Activity and Health—A Report of theSurgeon General Executive Summary. Atlanta, Ga:US Department of Health and Human Servic-es, Centers for Disease Control and Preven-tion, National Center for Chronic DiseasePrevention and Health Promotion; 1996.

2. US Department of Health and Human Servic-es, Centers for Disease Control and Preven-tion, Public Health Service, National Centerfor Chronic Disease Prevention and HealthPromotion, Activity DoNaP. Promoting PhysicalActivity—A Guide for Community Action. Cham-paign, Ill: Human Kinetics; 1999.

3. King AC. Community and public health ap-proaches to the promotion of physical activity.Med Sci Sports Exerc. 1994;26:1405–1412.

4. Schmid TL, Pratt M, Howze E. Policy as inter-

vention: environmental and policy approach-es to the prevention of cardiovascular disease.Am J Public Health. 1995;85:1207–1211.

5. Task Force on Community Preventive Servic-es. Recommendations to increase physical ac-tivity in communities. Am J Prev Med. 2002;22(4S):67–72.

6. US Department of Transportation. The Nation-al Bicycling And Walking Study—TransportationChoices for a Changing America: Final Report.Washington, DC: US Dept of Transportation;1994. FHWA-PD-94-023.

7. Humpel N, Owen N, Leslie E. Environmentalfactors associated with adults’ participation inphysical activity: A review. Am J Prev Med.2002;22(3):188–199.

8. King AC, Stokols D, Talen E, Brassington GS,Killingsworth R. Theoretical approaches tothe promotion of physical activity: forging atransdisciplinary paradigm. Am J Prev Med.2002;23(2S):15–25.

9. Handy SL, Boarnet MG, Ewing R, Killing-sworth R. How the built environment affectsphysical activity: views from urban planning.Am J Prev Med. 2002;23(2S):64–73.

10. Berrigan D, Troiano RP. The association be-tween urban form and physical activity in USadults. Am J Prev Med. 2002;23(2S):74–79.

11. Cervero R. Built environments and modechoice: toward a normative framework. Trans-port Res Part D. 2002;7:265–284.

12. Ewing R, Cervero R. Travel and the built en-vironment: a synthesis. Transport Res Rec.2001;1780:87–114.

13. Greenwald MJ, Boarnet MG. The Built Envi-ronment as a Determinant of Walking Behavior:Analyzing Non-work Pedestrian Travel in Port-land, Oregon. Irvine, Calif: University of Cali-fornia, Irvine; 2001.

14. Turner SM, Shafer CS, Stewart WP. BicycleSuitability Criteria: Literature Review and State-of-the-Practice Survey. Austin, Tex: Texas Depart-ment of Transportation; 1997. TX-97/3988-1

15. Frank LD, Engelke P, Hourigan D. How LandUse and Transportation Systems Impact PublicHealth: An Annotated Bibliography. Atlanta, Ga:Centers for Disease Control and Prevention;2001.

16. Craig CL, Brownson RC, Cragg SE, Dunn AL.Exploring the effect of the environment onphysical activity: a study examining walking towork. Am J Prev Med. 2002;23(2S):36–43.

17. Brownson RC, Baker EA, Housemann RA,Brennan LK, Bacak SJ. Environmental andpolicy determinants of physical activity in theUnited States. Am J Public Health. 2001;91:1995–2003.

18. Shriver K. Influence of environmental designon pedestrian travel behavior in four Austinneighborhoods. Transport Res Rec. 1996;1578:64–75.

19. New South Wales Health and Centre for Pop-ulation Health MAHS. Footpaths, Walkways,and Cyclepaths Checklist. Dubbo, New SouthWales: New South Wales Health and Centrefor Population Health, Macguire Area HealthService; September 1997.

20. Main Roads. Safety Audit Checklist for Dual-UsePaths. East Perth, Western Australia; MainRoads Western Australia; 1997.

21. Pein WE. Trail Intersection Design Guidelines.Chapel Hill: University of North CarolinaHighway Safety Research Center; 1996.

22. Guttenplan M, Landis BW, Crider L, McLeodDS. Multi-modal Level of Service (LOS) Analysisat a Planning Level. Washington, DC: Trans-portation Research Board; 2001. TRB Paper01-3084.

23. Landis BW. Bicycle level of service (BLOS)evaluation methodology. Sprinkle Consulting,Inc. Available at: www.fhiplan.com/mdpbikep

pedpplan/pdfdocs/bicyclelos.pdf. AccessedSeptember 20, 2002.

24. Landis BW, Vattikuti VR, Ottenberg RM, Mc-Leod DS, Guttenplan M. Modeling the RoadsideWalking Environment: A Pedestrian Level of Ser-vice. Washington, DC: Transportation Re-search Board; 2001. TRB Paper 01-0511.

25. Landis BW, Vattikuti VR, Brannick MT. Real-time human perceptions: Toward a bicyclelevel of service. Transport Res Rec. 1997:119–126.

26. Bonk D, Luxenberg S. Bicycle Suitability Index.Chapel Hill, NC: Chapel Hill Planning De-partment; 1999.

27. Harkey DL, Reinfurt DW, Sorton A. Develop-ment of the Bicycle Compatibility Index: A Level ofService Concept—Implementation Manual. Mc-Lean, Va: Federal Highway Administration;1998. FHWA-RD-98-095.

28. Crump C, Benson EB, Stutts JC. AssessingCommunity Policies and Environmental Supportfor Physical Activity: Pilot Test of Methods. ChapelHill, NC: The University of North CarolinaCenter for Health Promotion and DiseasePrevention; 1998.

29. Turner SM, Shafer CS, Stewart WP. BicycleSuitability Criteria For State Roadways in Texas.Austin, Tex: Texas Department of Transporta-tion; 1997. TX 97/3988-S.

30. Eddy N. Developing a level of service for bi-cycle use. Paper presented at: ProBike Pro-Walk ’96; September 3–6, 1996; Portland, Me.

31. Dixon LB. Bicycle and pedestrian level-of-ser-vice performance measures and standards forcongestion management systems. Transport ResRec 1538. 1996:1–9.

32. McCalip S, Tedder L. Orange County BicycleSuitability Study. Orange County, Calif: OrangeCounty Department of Planning; October1995.

33. Landis BW. Bicycle interaction hazard score:A theoretical model. Transport Res Rec 1438.1994:3–8.

34. Epperson B. Evaluating suitability of roadwaysfor bicycle use: Toward a cycling level-of-servicestandard. Transport Res Rec 1438. 1994:9–16.

35. Sorton A, Walsh T. Bicycle stress level as atool to evaluate urban and suburban bicyclecompatibility. Transport Res Rec 1438. 1994:17–24.

36. Pikora TJ, Bull FC, Jamrozik K, Knuiman M,Giles-Corti B, Donovan RJ. Developing a reli-able audit instrument to measure the physicalenvironment for physical activity. Am J PrevMed. 2002;23(3):187–194.

37. King AC, Bauman A, Abrams DB. Forgingtransdisciplinary bridges to meet the physicalinactivity challenge in the 21st century. Am JPrev Med. 2002;23(2S):104–106.

38. Burden D, Wallwork M, Sides K, Trias R, RueHB. Street Design Guidelines for Healthy Neighbor-hoods. Sacramento, Calif: Center for LivableCommunities; January 1999.

39. Local Government Commission Center forLivable Communities. Why People Don’t Walkand What City Planners Can Do About it. Sacra-mento, Calif.

40. Local Government Commission Center forLivable Communities. New Thinking for a NewTransportation Age. Sacramento, Calif: LocalGovernment Commission Center for LivableCommunities; 2001.

41. Burden D. Ten steps to create a walkablecommunity. Paper presented at: Triad Walka-ble Communities Regional Conference; Octo-ber 27–28, 1999; Winston-Salem, NC.

42. Burden D. Walkable Communities: A Search ForQuality. 1997.

43. Pedestrian Federation of America. Walk Trail: ACitizen’s Guide to Walkable Communities. Washing-ton, DC: Pedestrian Federation of America;1995.

September/October 2003, Vol. 18, No. 1 45

Appendix 1

Walking Suitability Assessment Form (developed by J. E.)

J. Emergy Walking Suitability Assessment Scores:Very good (less than 3.0): Color blue. These sidewalks are generally good for walking and wheelchair use.Good (3.0–5.9): Color Green. These sidewalks provide basic walking access, but could be upgraded to make them better walking and wheelchair

environments. Improvements might include enhancing the surface material or condition and installing more lighting.Fair (6.0–8.9): Color yellow. These sidewalks need improvements to the walking environment. Improvements might include lowering the posted

speed limit, improving the surface material or condition, or installing or widening buffers.Poor (9.0–26.0): Color orange. These sidewalks need major improvements to enable safe use. These types of improvements include replacing

unfirm surfaces (e.g., gravel, dirt, heaving bricks); repairing broken sidewalk sections; constructing curb ramps for wheelchair access; orconstructing a continuous sidewalk on at least one side of the street.

No sidewalk on quiet street (99.0): Color pink.No sidewalk on busy street (more than 99.0): Color Red. When there is no sidewalk, travel beside the road on foot or wheelchair is not safe or

comfortable (especially if the street has frequent, high-speed traffic). If these roads provide links between residential areas and frequentdestinations, the need for sidewalks is greater.

44. Sallis JF, Owen N. Physical Activity and Behav-ioral Medicine. Vol 3. Thousand Oaks, Calif:Sage; 1999.

45. Sallis JF, Bauman A, Pratt M. Environmentaland policy interventions to promote physicalactivity. Am J Prev Med. 1998;15(4):379–397.

46. Troped PJ, Saunders RP, Pate RR, ReiningerB, Ureda JR, Thompson SJ. Associations be-tween self-reported and objective physical en-vironmental factors and use of a communityrail-trail. Prev Med. 2001;32:191–200.

47. Stokols D. Translating social ecological theoryinto guidelines for community health promo-tion. Am J Health Prom. 1996;10:282–297.

48. Fawcett SB, Sterling TD, Paine-Andrews A, etal. Evaluating Community Efforts to Prevent Car-diovascular Diseases. Atlanta, Ga: Centers forDisease Control and Prevention, NationalCenter for Chronic Disease Prevention andHealth Promotion; 1995.

49. Emery JE, Crump CE, Bors P. Improvingopportunities for walking and bicycling incommunities. Paper presented at: ProBikeProWalk 2000; September 7, 2000; Philadel-phia, Pa.

50. Singleton RA, Straits BC, Straits MM. Ap-proaches to Social Research. New York, NY: Ox-ford University Press; 1993.

46 American Journal of Health Promotion

Appendix 2

Bicycling Suitability Assessment Form (adapted from Nils Eddy30)

Note: The form was adapted from the original. We created a ‘‘yes/no’’ assessment-marking process to ensure that every variable would be ratedby data collectors. We moved the algorithm to the bottom of the sheet and the comments area to the top right corner, and deleted the brief sidewalkassessment.N. Eddy Bicycle Suitability Assessment Scores:

Very good (less than 3.00): Color blue. A road that is bicycle friendly and usable by all levels of bicyclists. There are few improvements needed.Good (3.00–3.99): Color Green. A road that can be used safely by most bicyclists. Minimal improvements may be needed.Fair (4.00–4.99): Color yellow. A road that has some hazards, but still can be used by adults for bicycling. Not recommended for children. Specific

improvements are needed.Poor (5.0–6.99): Color orange. This road has many hazards and would require adult bicyclists to be very careful. Not safe for children. Many

improvements are needed.Very poor (higher than 6.99): Color Red. This road has many hazards, heavy traffic, and bad road conditions. Not safe for any bicyclists.

Improvements are greatly needed.

A fusion of the best of science and the best of practice —

together, to produce the greatest impact.

Stay on top of the science and art of health promotion withyour own subscription to the American Journal ofHealth Promotion.

“ The American Journal of Health Promotion provides a forum for that rare commodity— practical and intellectual exchange between researchers and practitioners. ”

Kenneth E. Warner, PhDAvedis Donabedian Distinguished University Professor of Public HealthSchool of Public Health, University of Michigan

“ The contents of the American Journal of Health Promotion are timely, relevant, andmost important, written and reviewed by the most respected researchers in our field. ”

David R. Anderson, PhDVice Programs and Technology, StayWell Health Management

Definition of Health Promotion

“Health Promotion is the science and art of helping peoplechange their lifestyle to move toward a state of optimalhealth. Optimal health is defined as a balance of physical,emotional, social, spiritual and intellectual health. Lifestylechange can be facilitated through a combination of effortsto enhance awareness, change behavior and createenvironments that support good health practices. Of thethree, supportive environments will probably have thegreatest impact in producing lasting change.”

(O’Donnell, American Journal of Health Promotion, 1989, 3(3):5.)

DIMENSIONS OFOPTIMAL HEALTH

Subscribe today...

Editor in Chief Michael P. O’Donnell, PhD, MBA, MPHAssociate Editors in Chief Bradley J. Cardinal, PhDDiane H. Morris, PhD, RDJudy D. Sheeska, PhD, RDMark G. Wilson, HSD

SECTION EDITORSInterventionsFitness

Barry A. Franklin, PhDMedical Self-Care

Donald M. Vickery, MDNutrition

Karen Glanz, PhD, MPHSmoking Control

Michael P. Eriksen, ScDWeight Control

Kelly D. Brownell, PhDStress Management

Cary Cooper, CBEMind-Body Health

Kenneth R. Pelletier, PhD, MD (hc)Social Health

Kenneth R. McLeroy, PhDSpiritual Health

Larry S. Chapman, MPHStrategiesBehavior Change

James F. Prochaska, PhDCulture Change

Daniel Stokols, PhDHealth Policy

Kenneth E. Warner, PhDApplicationsUnderserved Populations

Ronald L. Braithwaite, PhDHealth Promoting Community Design

Jo Anne L. Earp, ScDResearchData Base

David R. Anderson, PhDFinancial Analysis

Ron Z. Goetzel, PhDMethod, Issues, and Results in Evaluationand Research

Lawrence W. Green, DrPHQualitative Research

Marjorie MacDonald, BN, PhDMeasurement Issues

Shawna L. Mercer, MSc, PhD

The Art of Health PromotionLarry S. Chapman, MPH

ANNUAL SUBSCRIPTION RATES: (Good through 12/31/05)

Individual Institution

U.S. $99.95 $144.85

Canada and Mexico $108.95 $153.85

Other Countries $117.95 $162.95

CALL 800-783-9913 (U.S. ONLY) or 818-760-8520OR FIND US ON THE WEB AThttp://www.HealthPromotionJournal.com