Measuring the accessibility and equity of public parks: a ...staff.washington.edu/kwolf/Archive/Classes/ESRM304_SocSci/304 Soc Sci... · Measuring the accessibility and equity of

Measuring the accessibility and equity ofpublic parks: a case study using GIS

Sarah NichollsDepartment of Recreation, Park and Tourism Sciences

Texas A&M University, USA

Geographic information systems (GIS) can provide leisure service age ncies with numerousopportunities to enhance the planning and manage ment of their facilities. This paper demonstratesone such application, to the measurement of levels of accessibility and distributional equity offeredby a system of public parks . The methods proposed are relatively simple; nevertheless, they do offersubstantial improvements upon those previously utilized by leisure service providers. They facilitateidenti� cation of poorly served areas and populations, and sugge st where new facilities might best besited so as to maximize access and equity. These methods are illustrated by a case study of the parksystem in Bryan , Texas .

Many other applications of GIS to leisure service provision are available , several of which arebrie � y discussed. Toge ther, they could enable age ncies to function more effectively and , ultimately,to provide better levels of service to the public .

INTRODUCTION

The use of geographic information sys tems(GIS) in public planning and management hasproliferated over the past decade , yet itsadoption within the � eld of leisure servicesappears to have been relatively limited .Wicks et al. (1993) offe red an appraisal of thepotential applications of GIS for park market-ing, management and planning, but the in-frequency of subsequent ar ticles devoted toGIS in the leisure services literature suggeststhat the potential they identi� ed has notbeen realized .

Since 1993, the ease of implementing andusing GIS has improved signi� cantly. Greatadvances have been made in both the num-ber and power of capabilities provided asstandard functions in GIS packages , and theamount of easily available data, much of itdownloadable over the Internet, has in-creased . These improvements have enabledthe development of more sophisticated ana-lytical applications , including many that are

per tinent to researchers and professionals inthe � eld of leisure services .

The primary aim of this paper is to illus-trate the utility of GIS as a tool for measuringlevels of accessibility and distributionalequity within a sys tem of public parks . Levelof access to public parks is an impor tantindicator of the effectiveness of their provi-sion . Similarly, the degree of equity, or fair-ness, affo rded by the distribution of facilitiesis a central concern of public leisure serviceproviders .

Despite the inherently spatial nature of theconcepts of access and equity, little researchappears to have been conducted into theapplication of GIS technology to their meas-urement . While Nedovic-Budic et al. (1999)do provide a sample of applications devel-oped within the context of the Illinois Recrea-tion Facilities Inventory, their contributionillustrates the breadth of these potentialapplications more than the speci� cs of theirimplementation . The focus of this paper is onexplaining the methods available within a GIS

Managing Leisure 6 , 201–219 (2001)

Managing Leisure ISSN 1360-6719 print /ISSN 1466-450X online © 2001 Taylor & Francis Ltdhttp :/ /www.tandf.co .uk/ journalsDOI: 10.1080/13606710110084651

environment to assess the levels of accessi-bility and equity of urban leisure facilities ,and on the interpretation of their results . Inaddition , the use of a straigh t line and anetwork measure of dis tance are compared .

There are , potentially, many ways of carr y-ing out such objectives . In this paper, rela-tively simple methods , in terms of the GISfunctionality and level of computational com-plexity required , are utilized , the aim being toillustrate their everyday utility to practi-tioners with limited time and resources .

The paper is divided into � ve major sec-tions . The � rs t explains the concepts ofaccessibility and distributional equity, whilethe second provides a de� nition of GIS and abrief description of how this technologyworks . Next , methods of measuring accessi-bility and equity are discussed . The four thsection consists of a case study illustratingthe application of GIS to the measurement ofthe two concepts of interest . The ar ticleconcludes with a discussion of the case studyresults and of the potential of GIS for improv-ing the planning and management of leisureservices and facilities in other manners .

ACCESSIBILITY AND EQUITY

The case study presented here illustrates theuse of GIS technology when assessing levelsof accessibility and distributional equity in asys tem of public parks . These two principleswere selected because they are both widelyrecognized as impor tant indicators of a well-functioning urban sys tem .

Accessibility refers to the ease with whicha site or service may be reached or obtained;it can thus be said to measure the relativeoppor tunity for interaction or contact with agiven phenomenon such as a park (Gregory,1986) . Pred (1977) speci� cally relates thequality of life within a city to the accessibilityof its inhabitants to nature and extensiverecreational open space oppor tunities .

Traditional studies of accessibility ap-proach this concept from a purely geometric

perspective, founded in location theory, inwhich the aim is to maximize the ef� ciency ofdistribution networks so as to minimizesys tem costs (McAllister, 1976; Morrill andSymons , 1977; Gregory, 1986). Such ef� -ciency-based analyses do not take into ac-count the distribution of outcomes orbene� ts among users . Rather, the seeminglyvalue-free models that they utilize can resultin signi� cant discrimination against cer taingroups and areas (Morrill, 1974; Morrill andSymons , 1977; Harvey, 1988) . It is , therefore ,impor tant to identify the social and eco-nomic dimensions of accessibility as theyrelate to users instead of concentratingsolely on geometry and sys tem pro� ts . Atthis point , consideration of the concept ofdistributional equity becomes relevant .

The term equity refers to the fairness orjustice of a situation or distribution (Smith ,1986) . With respect to the distribution ofpublic resources , it is concerned with thequestion , ‘Who gets what?’ or, normatively,‘Who ough t to get what?’ (Laswell, 1958, inWicks and Crompton , 1986, p . 342). Equity isclearly a subjective concept , open to mul-tiple , sometimes competing, interpretations .Indeed , Symons (1971, p . 59) has describedquestions regarding distributional equity asrepresenting ‘the scienti� c geographer’snemesis’ precisely due to the dif� culties ofde� ning what is ‘just’ or ‘fair’.

It is , never theless, necessary to adopt ade� nition of equity in order to analyse it .Typologies of equity such as those suggestedby Lucy (1981) and Crompton and Wicks(1988) are useful guides when attempting todo this . Between them , these authors identifyfour major classes of equity with regard tothe allocation of resources , each of whichcan be operationalized in one or more ways .As Fig. 1 illustrates , the four classes are: i)equality; ii) compensatory (Crompton andWicks) or need (Lucy); iii) demand (includingLucy’s category ‘preferences’); and , iv) mar-ket (including Lucy’s category ‘willingness topay’).

202 Nicholls

Equity based on equality can be oper-ationalized in one of two ways . The allocationof inputs to services in an equal manner,regardless of geographic area or the socio-economic characteris tics of residents , isanalogous to the notion of equal oppor tunity.All areas and residents receive equal facili-ties , whether measured in terms of expendi-ture , man-hours , or numbers of features .Output equality, in contrast , requires ‘equal-ity of condition after receipt of service’ (Line-berr y and Welch , 1974; p . 709) . It measuresthe bene� ts received by residents as a resultof public service provision . Paradoxically,input equality rarely results in output equal-ity, and vice versa, hence the notion ofequality is compromized either way. Never-theless , as demonstrated by Wicks andCrompton (1986, 1987) suppor t fo r this ap-proach to service provision is relatively highamongs t US park and recreation depar t-ments , city council members , and the generalpublic, a � nding the authors attribute toprevailing traditions of equality and freedomof choice, and political and legal expe-diency.

A compensatory, or need-based , approachto equity implies , as Lucy (1981; p . 448)notes, ‘that unequals should be treated un-equally’. Thus , disadvantaged residents orareas are awarded extra increments of re-sources so as to provide these groups withoppor tunities that they might not otherwisehave had . The role of the public sector is ,therefore , to redistribute resources in a com-pensatory manner. While the identi� cation of‘disadvantaged’ or ‘needy’ groups may besubject to debate, they are usually des-ignated according to socio-economic criteriasuch as income.

Demand as a conceptual basis for achiev-ing equity involves rewarding those whodemonstrate an active interest in a serviceor facility, whether th rough use (the eco-nomic model) or vociferous advocacy (thepolitical model). The notion of producinggoods and services to meet consumers’demands is grounded in microeconomic the-ory; this approach is also often favoured byelected of� cials who see it as an appropriatemeans by which to improve their publicaccountability.

Fig 1. Taxonomy of equity models (after Lucy (1981), and Crompton and Wicks (1988))

Measuring the accessibility and equity of public parks 203

The � nal group of equity models illustratesthe potential in� uence of marke t forces onservice distribution . Allocation of resourcesaccording to amount of taxes paid , will-ingness to pay (e .g., an entrance or user fee) ,or the least cost alternative are included inthis category. Market determination of ser-vice provision is most appropriate in thecommercial sector, and was not consideredhere.

Rather, a compensatory or need-basedapproach to the equitable provision of publicleisure services was adopted . The least ad-vantaged were de� ned according to thesocio-economic characteristics of age , in-come , race /ethnicity, and population densityof area of residence . The groups consideredmost in need with regards to the provision ofpublic leisure services and facilities were ,thus , the young, the elderly, minorities , andthose living in areas of higher populationdensity.

The choice of a need-based approachrecognizes the National Recreation and ParkAssociation’s (NRPA’s) call for increased at-ten tion to the lower-income and culturallydiverse por tions of society when planningand managing public leisure oppor tunities(Mertes and Hall, 1995) . It is also rooted inthe concept of social justice and the notionof socially just communities in which ‘allindividuals and groups are treated fairly’(Beatley and Brower, 1993; p . 18).

GEOGRAPHIC INFORMATION SYSTEMS

GIS have been de� ned as ‘automated sys temsfor the capture, storage , re trieval, analysis ,and display of spatial data’ (Clarke , 1995; p .13) . A GIS essentially consists of a set ofspatial or map information and a databasecontaining the attributes , both quantitativeand qualitative , of this geographic informa-tion . For example , the database attached to amap of a park might contain informationregarding its acreage , number of tenniscour ts and parking spaces , length of foot-

paths , and level of development These twosets of information are dynamically linked ,using geographic location as the commonidenti� er, such that the attribute data may beaccessed th rough the map or vice versa. It isthis embedded linkage between the spatialdata and its non-spatial attributes that im-bues GIS-based spatial analysis with a newlevel of value and meaning beyond thatpreviously obtainable by using either spatialinformation accessible from paper maps orstatistical analyses typically applied to data-bases , but with no means of ef� ciently com-bining the two . Moreover, the data utilizedneed not relate to only one speci� c theme ,such as parks . An underlying principle of GISis the ability to overlay diffe rent kinds ofinformation for a speci� ed geographic area(e.g. distributions of facilities and naturalresources , transpor tation networks , and de-mographic data) so that spatial relationshipsbetween them may be assessed .

These computer-based sys tems offer anumber of other advantages over the use ofanalogue , paper maps . For example , theexis tence of information in digital form al-lows fo r a dynamic environment in which theuser can interact with the data, manipulatingthem at will and visualizing the results eitheron paper or on the screen . Another keybene� t of GIS is the ability not only to mapand assess current situations , but also toenvisage future scenarios under diffe rentsets of hypothetical parameters . The speedand ease of this ability has prompted theadoption of GIS as a decision-making tool bymany public policy makers and planners .

MEASURING ACCESSIBILITY ANDEQUITY

To date , few authors have integrated theevaluation of accessibility and equity with-in a single study. Measurement issues relatedto each of these concepts are discussedindividually before attention is turned tomethods of combining their analysis .

204 Nicholls

Accessibility

Perhaps the most basic standard with re-gards to the provision of urban parks is theNRPA’s recommendation that 10 acres (4.1hectares) of open space be available per 1000residents . This compares to the 6 acre (2.4ha) standard advocated by the NationalPlaying Fields Association (NPFA) in the UK.Many cities calculate this ratio to obtain abroad picture of the adequacy of their level ofsupply. Some also divide an urban area intosmaller zones and calculate the amount ofparkland available to residents within each ofthese units . The problem with this ‘con-tainer’ approach (Talen and Anselin, 1998) isthat it assumes the bene� ts of servicesprovided are allocated only to residentswithin the prede� ned zone in which they aresituated , and that no spatial externalities tosurrounding areas occur. Similarly, it as-sumes that residents of an area have suf� -cient access that they all bene� t from theservices provided within it , an unrealis ticexpectation . A major disadvantage of thisquotient , therefore , is that it does not con-sider the spatial distribution of oppor tun-ities . The location of parks relative to theirpotential users is an impor tant factor in theassessment of accessibility, as is increasinglybeing recognized by both NRPA and NPFA.

Some park and recreation depar tments doproduce maps illustrating the accessibility oftheir facilities , thereby taking this spatialdimension into account . Access is typicallyde� ned according to each park’s servicearea, represented by a circle drawn aroundthe facility with a radius equivalent to themaximum desired distance of users from it .This distance might be determined accordingto NRPA’s public park and open space classifi-cation scheme , which recommends the ideallocation and size of various types of openspace relative to the surrounding population(Table 1) . This ‘radius’ method , as it isreferred to here , is an example of the cover-ing model of accessibility (Hodgar t, 1978) .

Residents are said to be ‘covered’ by, i.e.,have access to , a park if they are locatedwithin the speci� ed maximum distance of it ,but are deemed to have inadequate access ifthey are not .

This method does have advantages overthe computation of ratios of parkland area topopulation , but several problems arise fromits usage . First , the radius method can pro-vide only an approximate representation of apark’s service area since it assumes ‘as-the-crow-� ies’ movement . In reality, potentialusers cannot travel in straigh t lines . Theymove instead along prede� ned public righ tsof way, and must avoid barriers to travel suchas railway lines and rivers . Thus , the actualtravel distance is almost always greater thanthe direct distance. Clift (1994) , for example ,in a study of the accessibility of grocerystores par ticipating in a Women , Infants andChildren’s Program food checks scheme ,found that while some recipients lived only4500 feet from a store , their actual traveldistance to it amounted to more than th reeand a half miles, or 18,480 feet . Such differ-ences may cause considerable increases intime , cost and effor t for those with limitedmeans and mobility or for those with youngchildren . This point is especially per tinentfor mini and neighbourhood parks thatshould ideally be accessible by walking alongresidential streets and trails .

The second disadvantage of the radiusmethod is that it assumes parks to be open toaccess at all points along their boundaries .This is not always true; in many cases , usersmust travel out of their way to reach a pointof entr y. These � rs t two factors are bothlikely to lead to overestimation of the size ofa park’s service area. The third , however,may lead to its underestimation , and isrelated to measurement of the speci� eddistance from the centre of the park ratherthan its boundary. As the size of a parkincreases , and the distance between its cen-tre and its perimeter grows , underestimationof the service area becomes greater due to


Table 1 National Recreation and Park Association’s Public Park and Open Space Classi�cationScheme

Type of Park Description Location Criteria Site Criteria Population Served

Mini Park Used to addresslimited, isolated orunique recreationalneeds.

Service area usuallyless than a 1�4 mile(0.4 km) along trailsor low-volumeresidential streets.

Usually between2,500 square feetand 1 acre (0.4hectares); maximum5 acres (2 ha).

500 to 2,500

NeighbourhoodPark

The basic unit of apark system. Servesthe recreational andsocial focus of theneighbourhood.Emphasis is oninformal active andpassive recreation.

Walking distance ofa 1�4 to a 1�2 mile(0.4–0.8 km),uninterrupted bynon-residentialroads or otherphysical barriers.

Minimum of 5 acres(2 ha), 7 to 10 acres(2.8 to 4.1 ha)optimal.

2,000 to 10,000

Community Park Serves a broaderpurpose than aneighbourhoodpark. Focus is onmeeting community-based recreationneeds.

Usually serves 2 ormoreneighbourhoodswithin a 1�2 to 3 mile(0.8–4.83 km)distance.

Between 20 and 50acres (8.1 and 20.3ha).

10,000 to 50,000

Natural ResourceArea

Land set aside forthe preservation ofsigni�cant naturalresources, remnantlandscapes, openspace, and visualaesthetics/buffering.

Location determinedprimarily byresource availabilityand opportunity.

Variable Variable

Greenway Ties park systemcomponentstogether to form acontinuous parkenvironment.

Location determinedprimarily byresource availabilityand opportunity.

Variable Variable

Sports Complex Consolidates heavilyprogrammed athletic�elds andassociated facilitiesto larger and fewersites strategicallylocated throughoutthe community.

Strategically locatedthroughoutcommunity.

Usually a minimumof 25 acres (10.1ha), with 40 to 80(16.2 to 32.4 ha)being optimal.

Variable

Special UseFacility

Covers a broadrange of parks andrecreation facilitiesoriented towardsingle-purpose use.

Variable Variable Variable

Private Park/RecreationFacility

Parks and recreationfacilities that areprivately owned yetcontribute to thepublic park andrecreation system.

Variable Variable Variable

Source: Mertes and Hall, 1995, p. 94

206 Nicholls

the inclusion of the park itself within thiszone . The � nal disadvantage of drawing acircle around the centroid of a park is that itdoes not take into account the park’s shape .The less regular this becomes the higher thedegree of inaccuracy and misrepresentationof the service area; a linear park wouldprovide a good example .

The con� guration of a park, and the posi-tion(s) of its point(s) of access , as well as therealistic measurement of distance to it , areimpor tant factors not considered under thetraditional, radius method . An alternativethat minimizes these inaccuracies is there-fore desirable. One such approach is basedon the measurement of distance along theroads and other public righ ts of way sur-rounding parks so as to emulate as closely aspossible the actual routes that users arelikely to follow between their residences anddesignated points of access to facilities . This‘network analysis’ approach appears to re-spond to all the disadvantages of the radiusmethod . Distance can be measured to orfrom each of the access points to each park.In this way, dis tance is measured realis tically,and from the boundary of the park ratherthan its centre , and the park’s shape is alsotaken into consideration . This approach canbe implemented quite simply in a GIS envi-ronment using a shor test path algorithm pre-encoded into the software.

In this study, both measures of distance ,using the straight-line radius and the networkprocedures , were utilized . This allowed com-parison of the levels of service por trayed byeach , and discussion of the implications ofany differences. These might relate to theaccuracy of service assumed by agenciesusing the simpler, straight-line method , aswell as to the degree of equity indicated sincein this study equity is inherently dependentupon level and distribution of access .

Equity

Previous empirical equity analyses havebeen of two main kinds , which Scott and

Cutter (1996) label ‘outcome’ and ‘process’s tudies . ‘Outcome’ studies have focused onthe distribution of various resources relativeto the socio-economic characteris tics of resi-dents . Speci� cally, they have been concernedwith the adherence or otherwise of distribu-tions of various publicly provided resourcesto the ‘underclass hypothesis’ (Lineberr y,1977) . This hypothesis purpor ts that sys tem-atic and deliberate discrimination exis tsagainst cer tain socio-economically disadvan-taged groups and areas in the distribution ofgoods and services , resulting in their receiv-ing fewer and /or poorer quality resourcesrelative to more advantaged citizens . Underthe alternate view, ‘unpatterned inequality’,inequitable distributions are deemed to illus-trate no systematic bias against cer taingroups , i.e ., any inequities in terms of differ-ential receipt of goods and services seem tobe spread evenly th roughout the population .Early examples of ‘outcome’ studies includeAntunes and Plumlee (1977) , who examinedthe quality of neighbourhood streets in Hous-ton , and Mladenka and Hill (1977), whopresented an analysis of the distribution ofpublic parks and libraries , also in Houston .More recently, Koehler and Wrightson (1987)re-examined the distribution of Chicago’spark and recreation services , an analysisoriginally completed by Mladenka in 1980,and were responded to by Mladenka (1989).

‘Process’ studies , in contrast , have beenless concerned with who is or is not im-pacted by inequity, and more with the rea-sons underlying distributions of resources .Analyses such as those by Jones and Kauf-man (1974), Mladenka (1980), and Cingranelli(1981) have attempted to account for varia-tions in the distributions of public goods andservices th rough factors including racism ,political favouritism , institutionalized deci-sion-making rules , cer tain historical events ,and population shifts . They have discovereda complex variety of empirical patterns ,differing both between the various citiesstudied and the services examined . In this


study, however, the outcome of the distribu-tion of public leisure services , in terms ofvariations between different socio-economicgroups , rather than the processes underlyingthis distribution , is considered .

Despite the impor tance of both these typesof study, there remains a lack of analysis ofthe spatial aspect of equity, i.e., of studiesthat have explicitly investigated the relation-ships between the spatial distributions offacilities or services and the spatially refer-enced socio-economic characteris tics of thepopulations they serve . Although severalauthors (e.g., Mladenka , 1989; Wicks et al.,1993) have analysed the equity of distribu-tions of leisure services using various demo-graphic characteristics such as age andincome as independent variables , these stud-ies have tended to utilize the ‘container’approach to access and equity, as discussedabove , which compar tmentalizes leisurefacilities and their users in to discrete units ,allowing no spatial interaction betweenthem .

Combining accessibility and equity

Talen and Anselin (1998) have argued for theneed to replace the container approach withone that measures levels of accessibilitycontinuously over space. The level of acces-sibility to a set of facilities could then be usedas the criterion against which the degree ofequity is assessed . Investigating equity froma distinctly socio-spatial perspective, usinglevels of accessibility as the means of differ-entiation , directly responds to the type ofanalysis called for by critics of standardgeometric studies of accessibility. To date ,few authors have adopted this kind of socio-spatial approach to the analysis of equity,and only Talen and Anselin have investigatedthis issue in the context of parks (Talen , 1997;Talen , 1998; Talen and Anselin, 1998). Al-though the procedures these authors haveemployed are theoretically sound , they aresomewhat complicated in terms of computa-tional and software requirements and their

results are not always easy to interpret . Theaim of this paper is to demonstrate the utilityof a much simpler method with regards tocomputation , interpretation , and map dis-play.

The methods proposed consist of twostages . First , is the identi� cation of levels ofaccessibility using a relatively straightfor-ward buffering technique , with distancemeasured both as-the-crow-� ies and alongthe street network. Second , is the assess-ment of the degree of equity based uponthese levels of access . This involves compar-ing the characteris tics of those residentswithin a service area and who are consideredto have good access, with those of peopleoutside the service area for whom access isdeemed inadequate. This can be achievedusing a standard , two-sample statistical tes t .The approach draws on the methodologyused by Werner (1998) in his analysis of theequity implications of discontinuing a publicbus service in Ramsey County, Minnesota .

CASE STUDY: LEVELS OF ACCESSIBILITYAND EQUITY OF PUBLIC PARKS IN

BRYAN, TEXAS

The case study illustrates the application ofGIS technology to identi� cation of the levelsof accessibility and equity provided by asys tem of municipal parks in Bryan , Texas .The Bryan Parks and Recreation Depar tment(BPRD) manages 29 facilities in the city,including seven mini, nine neighbourhoodand seven community parks as de� ned ac-cording to the NRPA’s classi� cation . These 23facilities total 222 acres (89.9 hectares) ,which equals 4 acres (1.6 ha) per thousand ofthe population , a � gure well below NRPA’s 10acre (4.1 ha) recommendation (Table 2). Thisamount does rise when all facilities withinthe city are considered; these additionalamenities were not included in this paper,however, for reasons described below.

Single use (e .g., tennis cour ts) and specialuse (e.g., the athletic complex) facilities were

208 Nicholls

excluded from the analysis , as were privatesector sites , since the emphasis was on theaccessibility of everyday, outdoor recreationand open space oppor tunities . The regionalpark was also excluded since it is a long wayfrom the centre of town and is only acces-sible by car; in this study, the main concernwas for leisure sites within walking distanceof one’s residence.

The analysis was under taken at two levels .First , neighbourhood parks , as the basicunits of the sys tem (as described in Table 1)

were considered alone . Second , mini, neigh-bourhood and community parks were com-bined in order to include those areas lackinga neighbourhood facility but, never theless ,able to offe r some form of open space. Inboth cases , one half of one mile (0.8 kilo-metres) was used as the maximum recom-mended walking distance . While NRPA doesspecify up to 3 miles (4.8 km ) as an accept-able travelling distance fo r a communitypark, this assumes the user is driving. In thiss tudy, the ability to walk to a park wasemphasized . Thus , a distance of one half ofone mile (0.8 km ) was adopted th roughout.

This distance compares favourably to

those currently utilized by cities in the UKsuch as Dundee and London (both of whichlist 1.2 km as the desired maximum distancefrom local, neighbourhood-style parks) , butis more generous than Aberdeenshire Coun-cil’s ‘ideal hierarchy’ of open space , whichrecommends 0.5 km (Kit Campbell Asso-ciates , 2001).

MethodsThe discussion of methods is divided into� ve subsections: GIS software; data acquisi-tion; data preparation; implementation of GIStechniques; and , equity analysis .

GIS SoftwareThe case study was carried out using Arc-View GIS (Version 3.1), a package producedby the Environmental Sys tems Research In-stitute (ESRI). The network analysis tech-nique was implemented in Network Analyst .This is an ArcView extension that can beused to solve many common network prob-lems including the identi� cation of serviceareas .

Data AcquisitionA variety of data sources , summarized inTable 3, were utilized .

Table 2 Bryan Parks and Recreation Department Facilities

Type of Facility Total Number ofFacilities

Total Acres (Ha) ofFacilities

Acres (Ha) PerThousand of thePopulation

Mini Park 7 10 (4.1) 0.18 (0.07)Neighbourhood Park 9 75 (30.4) 0.82 (0.55)Community Park 7 137 (55.5) 2.49 (1.01)

Total For Mini,Neighbourhood andCommunity Parks

23 222 (89.9) 4.04 (1.63)

Regional Park 1 1060 (429.3) 19.27 (7.81)Athletic Complex 2 112 (45.4) 2.04 (0.83)Aquatic Centre 1 3 (1.2) 0.05 (0.02)Tennis Courts 1 1 (0.4) 0.02 (0.007)Golf Course 1 141 (57.1) 2.56 (1.04)Total For All Facilities 29 1539 (623.3) 27.98 (11.33)

Source: City of Bryan Park and Recreation Department Staff, 1996


A road map of Brazos County (in whichBryan is located ) created by the TexasDepar tment of Transpor tation (TxDOT), wasdownloaded from the Texas Natural Re-sources Information Sys tem (TNRIS) site(www.tnris .state .tx.us /pub /GIS/highway) .The census data utilized in the analysis ofequity were obtained from ESRI’s ArcDataOnline (ADOL) web facility (www.esri.com /data /online / tiger/ index) . In this study, analy-sis was carried out on census blocks so as toobtain as � ne a degree of spatial detail aspossible. Since this is the lowest level ofaggregation available , only a limited amountof socio-economic information is provided soas to maintain resident con� dentiality. How-ever, this provides a broad enough range ofvariables to analyse equity, as it has beende� ned in this study.

Data PreparationOnce all the relevant data had been col-lected , they were entered into the GIS asoverlaying ArcView shape� les . A ‘shape� le’ isthe name given within ArcView to the geo-graphical representation of a theme or layerof spatial information . In order for each

separate layer to be superimposed correctlyupon the others , each must be stored in thesame geographic projection and co-ordinatesys tem .

The next task was to locate all public parksin the study area and create a new ArcViewshape� le to contain them . Sites were identi-� ed using paper maps and site diagramsprovided by the BPRD (City of Bryan , 1996) ,and were then drawn into ArcView usingproper ty line data provided by the BrazosCounty Appraisal District as a base.

In addition , it was necessary to locate allthe points of access to each of the parks inorder to operationalize the accessibilitymeasure based on the positions of points ofentry. This involved extensive � eldworkrather than a simple reliance on paper maps .All potential entrances were identi� ed andrecorded so as to capture pedestrian pointsof access in addition to the primary,vehicular ones .

Implementation of GIS Techniques

The two methods of measuring accessibility,‘radius’ and ‘network analysis’, were appliedto each set of parks . The radius techniquewas implemented using the Create Bufferscommand under the Them e menu of ArcView.A dis tance of one half of one mile (0.80 kms)was speci� ed , measured from the geographiccentre of each park, and circular buffe rs ,representing each park’s service area, werecreated .

The network analysis technique also in-volved speci� cation of the travel distanceallowed for each type of park. This was thenmeasured outwards from each park’s speci-� ed points of access along the surroundingstreets . Since neighbourhood parks areideally accessible by walking along locals treets , only this type of street was includedin the analysis; travel along major routes wasnot permitted . A service area was created forevery point of access , and these were then

Table 3 Sources of Data

Type of Data Source

Street network Texas Department ofTransportation

Boundary lines(lots, etc.)

Brazos County AppraisalDistrict (BCAD)

Park boundaries BCAD and Bryan Park andRecreation Department

Park accesspoints

Personal observations ateach park

Socioeconomicinformation

U.S. Bureau of the CensusSummary Tape Files (1990),ArcData Online

Geographiccensusinformation

U.S. Bureau of the CensusTIGER Files, ArcData Online

210 Nicholls

joined into a single service area for each parkusing the Union command in the Edit menu .

The Select by Them e function in the Them emenu was then used to identify those censusblocks lying inside and outside each servicearea. To avoid the need for any fo rm of arealinterpolation , only those blocks with theirgeographic centre inside each service areawere identi� ed as being within it , i.e., theHave Their Centre In criterion was utilized .Although it should be recognized that thistechnique provides only an estimate of thetrue number and characteris tics of residentslocated within service areas , it was deemedaccurate enough for the level of detail likelyto be needed by a public leisure servicesdepar tment .

Equity Analysis

The equity analysis was carried out using theMann-Whitney U tes t procedure in SPSS. Anon-parametric procedure such as Mann-Whitney was required due to the lack ofnormality, equal variances , and independ-ence typical of spatial data . For each varia-ble , the median value for census blocksoutside the service area was compared tothe median value for blocks inside it , and theextent to which the two medians differed wascomputed . The nine variables utilized in theequity analysis were: i) population density;ii) per cent non-White (i.e., Blacks , Asians ,American Indians , and all other races); iii) per cent Black; iv) per cent Hispanic; v) percent under age 18; vi) per cent over age 64;vii) per cent of housing units renter occu-pied; viii) mean housing value (for owneroccupied units); and , ix) mean contract rent(for rental units). Housing tenure and valuewere used as a proxy for income sinceincome data are not available for censusblocks . Groups considered most likely to bein ‘need’ of better than average access toparks were non-Whites, those earning lowincomes (approximated by those who rent as

opposed to own their home, and those whoseproper ty or rental value is lower than aver-age ), the young and the elderly, and thoseresiding in more densely populated areas andless likely to have access to a private garden .Percent Black and Hispanic were used toallow a more detailed analysis of any par tic-ular racial or ethnic biases found in levels ofaccessibility. For reference purposes , Table 4contains a summary of socio-economic anddemographic information for the city ofBryan as a whole.

RESULTS

This section is divided into two par ts . First ,levels of access to Bryan park facilities are

Table 4 Selected Socio-Economic and Demo-graphic Data for Bryan, Texas

Characteristic Value(1990)

Total Population 55,002Age Distribution

per cent Under 18per cent Over 64

27.09.8

Racial Compositionper cent Whiteper cent Blackper cent Other

69.917.212.9

per cent Hispanicper cent Non-Hispanic

19.880.2

Housing Tenureper cent Occupied Housing UnitsOwner Occupiedper cent Occupied Housing UnitsRenter Occupied

48.351.7

Median Housing Value (owneroccupied units, $)

58,400

Median Contract Rent ($) 312Income

Median Household Income ($)Median Family Income ($)Per Capita Income ($)

22,57729,27711,691

Source: U.S. Bureau of the Census, 1990


discussed . Second , the equity of their dis-tribution is examined .

AccessFigure 2 illustrates service areas for the nineneighbourhood parks in Bryan according tothe radius and network analysis approaches .Access appears limited to those neighbour-hoods immediately surrounding each facility,leaving large areas th roughout the city with-out access . In most cases , as expected , the

service area according to the network analy-sis technique is much smaller and moreirregularly shaped than under the radiusmethod . The two service areas for the park inthe far nor th-west section of the city clearlyillustrate the impact of a barrier, in this casethe railway line , on reducing the actualdegree of access to a facility.

The two sets of service areas for all mini,neighbourhood and community parks areillustrated in Fig. 3. Both indicate a lack of

Fig. 2 Service areas for Bryan neighbourhood parks according to the radius and network analysismethods

212 Nicholls

coverage in the far nor th-west tip of the cityand in the east . In contrast , the nor thern andsouthern areas of Bryan appear to have goodaccess to a park. Once again , however, thetotal service area as de� ned by networkanalysis is visibly smaller than under theradius technique .

The location of Bryan’s park facilities , and ,hence , of the patterns of access produced(whether under the radius or network analy-sis method ) , might be considered incon-sequential if the unserved areas were

sparsely populated . The ability to combinethe accessibility maps with a map of thepopulation distribution using a simple func-tion in a GIS is , therefore , an effective meansof more meaningfully representing the levelof accessibility in numeric as opposed topurely visual terms . Table 5 shows the resultsof this procedure .

The propor tion of the population coveredby neighbourhood parks is small under boththe radius (19%) and network analysis (12%)methods . The variation in service in absolute

Fig. 3 Service areas for all Bryan parks according to the radius and network analysis methods


terms amounts to approximately 3,900 peo-ple . In the case of all th ree types of parkcombined , the propor tions of the populationserved according to the radius and networkanalysis techniques are 55% and 38%, re-spectively. In absolute terms , the number ofresidents considered to have adequate ac-cess to a park is 8547 fewer under thenetwork analysis procedure . The accept-ability of such � gu res depends upon in-dividual depar tments’ service deliveryobjectives , but they do appear low.

Both the maps and the � gures presentedhere demonstrate the differences in resultsthat can be obtained by using two differentmethods of measuring distance . They sug-gest that if leisure service agencies use thesekinds of techniques to assess levels of acces-sibility to their facilities , they should use themore accurate network analysis technique . Ifthey continue to utilize the traditional circlemethod , they should recognize the degree ofinaccuracy that may be por trayed .

EquityTable 6 contains results of the Mann--Whit-ney equity analyses . Since SPSS only repor tsresults of this tes t in a two-tailed manner, itwas necessary to compare the median valuesof the variables inside and outside serviceareas to determine whether equity or in-equity was indicated in those instanceswhere a signi� cant difference between thetwo groups was found (p , 0.05) . Giventhe de� nition of equity adopted here, an

equitable distribution was suggested whenpopulation density, the propor tion of youngor elderly residents , the propor tion of minor-ities , or the propor tion of home renters , wassigni� cantly higher within the service areathan outside of it , or when the mean housingvalue or contract rent was signi� cantly lower.In the opposite case , inequity was con-sidered to be demonstrated . When no sig-ni� cant difference was indicated (p , 0.05) ,the results were interpreted as suggestingthat those disadvantaged groups consideredmost critical in terms of their access to parkswere receiving equal oppor tunities com-pared to other por tions of the community.Thus , equality, rather than need-basedequity, was evident .

As Table 7 shows , the distribution of parksin Bryan appears to be quite equitable . Forneighbourhood parks alone, non-Whites (andBlacks in par ticular, though not Hispanics) ,as well as those with lower housing values orrents , have signi� cantly higher levels ofaccess to park facilities than do their White ,higher income neighbours . There appear tobe no signi� cant variations in access withpopulation density or age .

When all th ree types of parks are con-sidered together, those living in more denselypopulated areas also appear to be par tic-ularly well served by park facilities , in addi-tion to minorities and those on lowerincomes . While the elderly appear to be moreadvantaged than other ages under the radiusmethod , the young bene� t more according to

Table 5 Number of Blocks and Residents With Access to Bryan Parks Based Upon NRPA LocationStandard of One Half of One Mile

Type of Park Access Method

Number of Blocks

WithAccess

WithoutAccess

Number of Residents

WithAccess

WithoutAccess

Percentageof PopulationWith Access

Neighbourhood Centroid Radii 338 889 10,426 44,576 18.96Network Analysis 274 953 6,540 48,462 11.89

All Centroid Radii 825 402 30,173 24,829 54.86Network Analysis 671 556 21,026 33,376 38.23

214 Nicholls

network analysis . In no instance is an inequi-table distribution indicated; rather, if need-based equity is not evident, the situationappears to be one of at least equality.

Unexpectedly, the results of the equityanalyses are fairly constant across both

methods for both sets of parks , despite thedifferences in levels of accessibility foundbetween them . The only noticeable differ-ence is that for both categories of parks , theequity results under network analysis con-tain one more signi� cant variable (housing

Table 6 Mann–Whitney Analysis of Park Equity

Type of ParkAccessMethod Variable

Median Value of Variable

Within S.A. Outside S.A.Mann–Whitney U

2-tailedp

Neighbourhood CentroidRadii

Popn densityper cent Non-Whiteper cent Blackper cent Hispanicper cent Under 18per cent Over 64per cent ROMHVMCR

3571.4366.6716.0010.5727.2711.8640.00

363.00198.00

3809.5213.16

0.008.33

25.819.76

32.6348300.00

292.00

100416.060291.071489.598571.599045.097065.596000.076068.572306.0

0.5300.0000.0000.2720.3380.1510.1110.0000.000

NetworkAnalysis


3846.1572.2219.35

9.0928.3811.5442.86

35800.00198.00

3768.1214.29

0.008.47

25.8110.0032.10

47500.00288.00

90374.053098.563513.589920.587796.087848.581450.567429.064440.0

0.6030.0000.0000.5150.2400.2440.0070.0000.000

All CentroidRadii


3846.1530.22

5.779.09

27.2411.1133.33

40800.00246.00

3333.339.5150.008.37

25.007.75

28.8352500.00

311.50

96781.075543.085321.5

106352.5101454.0

99275.0100132.5

85829.082199.0

0.0100.0000.0000.7270.1360.0450.1070.0000.000

NetworkAnalysi

sPopn densityper cent Non-Whiteper cent Blackper cent Hispanicper cent Under 18per cent Over 64per cent ROMHVMCR

4032.2634.48

7.2510.3027.7811.1136.36

39550.00238.00

3200.0010.00

0.007.92

25.009.09

28.5752250.00

298.50

103866.585318.597989.5

116528.5111620.0119560.0111414.0

97240.599156.5

0.0000.0000.0000.1180.0090.3880.0130.0000.000

Note: S.A. = service area; Popn density = population density; RO = renter occupied; MHV = mean housing value;MCR = mean contract rent.


tenure ) than do those for the radius method .The consistency of results between methodsand across park type suggests that thosegroups considered most impor tant with re-gards to their levels of access to park andrecreation oppor tunities within the city ofBryan tend to be located closest to them andwithin relatively homogenous censusblocks .

CONCLUSION

This paper has illustrated to leisure serviceprofessionals the utility of GIS as a means ofvisualizing and measuring levels of accessi-bility and equity. In addition , it has demon-strated the impor tance of measuring

distance and access as accurately as possi-ble , using network rather than straight-linedistances , so as to provide more realisticrepresentations of the geographic extent ofservice areas and their populations .

The case study of Bryan parks clearlyindicated the differences in level of accessi-bility that the straight-line and network meas-ures of distance can produce. If the latter isaccepted , it appears that less than 40% ofBryan residents have good access to anyform of everyday open space, with only 12%being able to reach a neighbourhood parkwithin the distance speci� ed . However,based on the low acreages listed in Table 2and the even spread of facilities th roughoutthe populated area , this de� ciency can beattributed to a lack of suf� cient open space

Table 7 Mann–Whitney Analysis of Park Equity – Signi�cant Variables at 0.05

Type of Park Access Method Variable Median HigherWithin or OutsideService Area?

Indicates Equityor Inequity?

Neighbourhood Centroid Radii per cent Non-Whiteper cent BlackMHVMCR

WithinWithinOutsideOutside

EquityEquityEquityEquity

Network Analysis per cent Non-Whiteper cent Blackper cent Renter occupiedMHVMCR

WithinWithinWithinOutsideOutside

EquityEquityEquityEquityEquity

All Centroid Radii Population densityper cent Non-Whiteper cent Blackper cent Over 64MHVMCR

WithinWithinWithinWithinOutsideOutside

EquityEquityEquityEquityEquityEquity

Network Analysis Population densityper cent Non-Whiteper cent Blackper cent Under 18per cent Renter occupiedMHVMCR

WithinWithinWithinWithinWithinOutsideOutside

EquityEquityEquityEquityEquityEquityEquity

Note: MHV = mean housing value; MCR = mean contract rent

216 Nicholls

rather than its poor distribution relative tothe population .

Despite the paucity of access to leisureoppor tunities , however, the degree of equityassociated with them is high . Thus , althoughBryan is lacking in park and recreationresources in an absolute sense , those it hasare well distributed relative to the needs ofthe population . Less advantaged groups (es-pecially minorities and those whose housingtenure and value characteristics indicatelower incomes) do tend to have betteraccess , indicating the exis tence of need-based equity. In no case is inequity found .

Whether this situation is the result of goodplanning or good for tune is beyond the scopeof this ‘outcome’ paper. However, a ‘process’analysis may well highlight the populationdynamics of Bryan , with many of the moreadvantaged residents increasingly locating innew subdivisions with their own privateleisure oppor tunities (individual gardens aswell as communal facilities , such as play-grounds and swimming pools , funded by thelocal Homeowners Association) , but with fewpublicly provided leisure services . As houseprices in less desirable, older neighbour-hoods , where public parks have , never the-less , historically been located , increase lessrapidly than in newer areas , so less advan-taged residents are more able to affo rd them .Hence , an equitable distribution of publicleisure provision may emerge .

The methods described here constitute agreat advance over the use of ratios orcircular buffe rs to measure the accessibilityof park facilities . They could , however, bedeveloped fur ther to provide even moredetailed pictures of levels of access andequity. For example, the use of one buffe r ofone half of one mile could be replaced with aseries of concentric rings of access , in onequar ter or one half mile increments , coveringan entire area . This would create a surface ofaccessibility indicating the distance of allresidents from their nearest leisure facility.

Alternatively, maps might be generated toillustrate the number of facilities accessiblewithin a speci� ed travel time or distancefrom individual neighbourhoods or censusareas , providing a surface of leisure oppor-tunity for a city.

The methods employed in this case studydo not take into account the characteris ticsof parks , they are concerned only with whichareas and populations have access to themand which do not . Inclusion of information asto individual facilities’ levels of development ,number and types of facilities , and physicalcondition , in the fo rm of an index represent-ing their overall attractiveness , could add amore qualitative dimension to the analysis .Finally, it should be noted that there aremany means of measuring accessibility andequity other than those utilized here, all ofwhich are legitimate and capable of beingimplemented within a GIS environment.

It is evident that GIS offers leisure servicepractitioners a real oppor tunity to improvetheir methods of measuring accessibility andequity. This technology offe rs agencies a toolfor the more ef� cient and effective planningand management of facilities in other ways aswell. Potential applications include the dis-semination of information to the publicthrough interactive web sites in which theyare able to locate facilities , look up detailsabout them , and obtain directions to them;the automation of facilities management;and , the selection of optimum sites for newdevelopments so as to maximize accessi-bility and equity using location-allocationmodels incorporated into Spatial DecisionSuppor t Sys tems .

Finally, the highly visual maps that GIS canprovide may be useful tools for improvingusers’ perceptions of public authorities’ ac-countability and openness . Increased inter-action and understanding between serviceproviders and their clienteles are likely todecrease the perceptual gaps between them ,ultimately leading to more satis� ed users .


ACKNOWLEDGEMENTS

The author would like to thank the twoanonymous referees for their helpful sugges-tions and comments .

REFERENCES

Antunes , G.E. and Plumlee, J.P. (1977) The dis-tribution of an urban public service: ethnicity,socio-economic status , and bureaucracy asdeterminants of the quality of neighbourhoodstreets , Urban Affairs Quarterly, 12(3),313–333.

Beatley, T. and Brower, D.J. (1993) Sustainabilitycomes to Main Street, Planning, May, 16–19.

Cingranelli, D.L. (1981) Race, politics and elites:Testing alternative models of municipal serv-ice distribution , Am erican Journal of PoliticalScience , 25(4), 664–692.

City of Bryan Park and Recreation DepartmentStaff. (1996) The 1996 Bryan Com prehensivePark and Recreation Master Plan and OpenSpace Plan. Bryan , TX, Bryan Park and Recrea-tion Department.

Clarke, K.C. (1995) Analytical and Com puter Carto-graphy, second edition , Upper Saddle River,NJ, Prentice Hall.

Clift, R. (1994) Spatial analysis in public healthadminist ration: a demonstration from WIC, inProceedings of GIS/ LIS 1994, 164–173.

Crompton, J.L. and Wicks , B.E. (1988) Implement-ing a preferred equity model for the deliveryof leisure services in the US context , LeisureStudie s, 7, 287–304.

Gregory, D. (1986) Accessibility, in: R.J. Johnston ,D. Gregory, and D.R. Stoddart (eds) TheDictionary of Hum an Geography, second edi-tion , Oxford , England , Blackwell, p . 2.

Harvey, D. (1988) Social Justice and the City, secondedition, Oxford , England , Blackwell.

Hodgart , R.L. (1978) Optimising access to publicservices: a review of problems , models andmethods of locating central facilities , Progressin Human Geography, 2(1), 17–48.

Jones , B.D. and Kaufman, C. (1974) The distribu-tion of urban public services: A preliminarymodel, Administration and Society, 6(3),337–360.

Kit Campbell Associates (2001) Rethinking OpenSpace: Open Space Provision and Management,

A Way Forward, Edinburgh , Scotland , Scot tishExecutive Central Research Unit .

Koehler, D.H. and Wrightson , M.T. (1987) Inequal-ity in the delivery of urban services: Areconsideration of the Chicago parks , TheJournal of Politics, 49 , 80–99.

Lineberry, R.L. (1977) Equality and Urban Policy:The Distribution of Municipal Public Services,Beverly Hills, CA, Sage Publications .

Lineberry, R.L. and Welch , Jr., R.E. (1974) Who getswhat : Measuring the distribution of urbanpublic services . Social Science Quarterly, 54 ,700–712.

Lucy, W. (1981) Equity and planning for localservices , Journal of the Am erican PlanningAssociation, 47(4), 447–457.

McAllister, D.M. (1976) Equity and ef� ciency inpublic facility location. Geographical Analysis,8, 47–63.

Mer tes, J.D. and Hall, J.R. (1995) Park , Recreation,Open Space and Greenway Guidelines, Arling-ton , VA, National Recreation and Park Asso-ciation.

Mladenka , K.R. (1980) The urban bureaucracy andthe Chicago political machine : Who gets whatand the limits to political control. The Am er-ican Political Science Review, 74 , 991–998.

Mladenka , K.R. (1989) The distribution of an urbanpublic service: The changing role of race andpolitics , Urban Affairs Quarterly, 24(4),556–583.

Mladenka , K.R. and Hill, K. (1977) The distributionof bene� ts in an urban environment: Parksand libraries in Houston, Urban Affairs Quar-terly, 13(1), 73–94.

Morrill, R.L. (1974) Ef� ciency and equity of op ti-mum location models , Antipode , 6, 41–46.

Morrill, R.L. and Symons , J. (1977) Ef� ciency andequity aspects of optimum location , Geo-graphical Analysis, 9, 215–225.

Nedovic-Budic , Z., Knaap , G. and Scheidecker, B.(1999) Advancing the use of geographic in-formation sys tems for park and recreationmanagement , Journal of Park and RecreationAdministration, 17(4), 73–101.

Pred , A. (1977) City Systems in Advanced Econo-mies, London, England , Hutchinson.

Rawls , J. (1971) A Theory of Justice, Cambridge ,MA, The Belknap Press .

Scot t, M. and Cutter, S. (1996) GIS and environmen-

218 Nicholls

tal equity: An analysis of the assumptions ,available on line at: Http:/ / www.geo.wvu.edu/i19/ papers/ scott.html.

Smith , D.M. (1986) Equity, in: R.J. Johnston , D.Gregory, and D.R. Stoddart (eds) The Diction-ary of Human Geography, second edition ,Oxford , England , Blackwell, p . 139.

Symons , J.G. (1971) Some comments on equityand ef� ciency in public facility location mod-els . Antipode , 3(1), 54–67.

Talen , E. (1997) The social equity of urban servicedistribution: An exploration of park access inPueblo , Colorado , and Macon, Georgia , UrbanGeography, 18(6), 521–541.

Talen , E. (1998) Visualizing fairness: Equity mapsfor planners , Journal of the Am erican PlanningAssociation, 64(1), 22-38.

Talen , E. and Anselin, L. (1998) Assessing spatialequity: An evaluation of measures of access-ibility to public playground s , Environm ent andPlanning A, 30 , 595–613.

United States Bureau of the Census . (1990) Facts

about my community, available online at:Http:/ / factfinder.census.gov/ java_prod/ dads.ui.fac.Com munityFactsPage.

Werner, R.J. (1998) Equity in public t ransit: A tests tatistic for buffer problems, available onlineat : Http://www.esri.com /library/userconf/proc98/PROCEED/TO150 /PAP142/P142 .HTM

Wicks , B.E. and Crompton , J.L. (1986) Citizen andadminist rator perspectives of equity in thedelivery of park services . Leisure Sciences, 8,341–365.

Wicks , B.E. and Crompton , J.L. (1987) An analysisof the relationship between equity choicepreferences , service type and decision mak-ing group s in a US city, Journal of LeisureResearch 19(3), 189–204.

Wicks , B.E., Backman, K.F., Allen, J. and VanBlaricom , D. (1993) Geographic InformationSystems (GIS): A tool for marketing, manag-ing, and planning municipal park sys tems ,Journal of Park and Recreation Administration,11(1), 9–23.


Documents

Measuring the accessibility and equity of public parks: a ...staff.washington.edu/kwolf/Archive/Classes/ESRM304_SocSci/304 Soc Sci... · Measuring the accessibility and equity of