Digital Excellence in Chicago A Citywide View of Technology Use

Karen Mossberger, Ph.D., Public Administration, University of Illinois at Chicago Caroline J. Tolbert, Ph.D., Political Science, University of Iowa Commissioned by: City of Chicago Department of Innovation and Technology Supported by: John D. and Catherine T. MacArthur Foundation State of Illinois Department of Commerce and Economic Opportunity July 2009

2

TABLEOFCONTENTSExecutiveSummary 4PartI.Introduction:WhyDigitalExcellence? 10PartII.InternetAccessandUse 11

ReadingtheResults 13 InternetUseinAnyPlace 16 HomeInternetUse 17 NeighborhoodVariationforInternetUseAnywhereandatHome 18 BroadbandAccess 20 NeighborhoodVariationforHomeBroadband 22 SummarizingWhatMattersforUseandAccess 22 DescribingtheLess‐Connected 22PartIII.PublicandWirelessAccess 23 InternetUseatLibraries 25 NeighborhoodVariationinLibraryInternetUse 28 CommunityTechnologyCenters 30 WirelessAccess 31PartIV.BarrierstoAccessandPublicPolicy 33 NeighborhoodVariationinInterest 36 NeighborhoodVariationinCostConcerns 38 NeighborhoodVariationinDifficultyUsingtheInternet 40 PolicyImplications 42PartV. OnlineActivitiesthatInfluenceOpportunitiesforResidents 43 EmploymentandTraining 43 NewsandPoliticsOnline 46 DigitalGovernment 47 HealthCare 50 SummaryonInternetActivities 51PartVI.Conclusion:ChallengesandOpportunitiesforDigitalExcellenceinChicago 51 PublicOpinionasanOpportunity 52 ChallengesandOpportunitiesforAddressingDisparities 53References 54AppendixA.LogisticRegressionModels 56AppendixB.MultilevelModels 60AppendixC.Survey 76

TABLESWhatMattersTableA:WhoUsestheInternetinAnyPlaceandWhoHasHomeAccess? 15WhatMattersTableB:WhoHasaBroadbandConnectionComparedtoDial‐Up? 21WhatMattersTableC:WhoUsestheInternetatPublicLibraries? 26WhatMattersTableD:WhatAretheReasonsChicagoResidentsDoNotHaveHomeInternet? 34Table1.InternetUsewithNoHomeAccessby2007FamilyIncome 23Table2.ReasonsforUsingtheInternetattheChicagoPublicLibrary 25Table3.FrequencyofWirelessUseinPublicPlacebyAge 31Table4.FrequencyofWirelessUseinPublicPlacebyRaceandEthnicity 31

3

Table5.FrequencyofCellPhoneUsetoConnecttoInternetbyAge 32Table6.FrequencyofCellPhoneUsetoConnecttoInternetbyRaceandEthnicity 32Table7.ReasonsforNoInternetatHome 33Table8.MainReasonforNoInternetatHomebyRaceandEthnicity 34Table9.ActivitiesOnline 43Table10.FrequencyofInternetUseforJobbyEducation 44Table11.WhereShouldaWirelessAvailabilityProjectStart? 52Table12.WouldYouSupportaWirelessAvailabilityProjectforaSmallTaxorFeeIncrease? 52

MAPS

PercentWhoHaveInternetAccessatHome 19PecentWhoUsetheInternetattheLibrary 29PercentWhoDoNotUsetheInternetBecauseNotInterested 37CostastheReasonforNoInternetAccess 39PercentWhoHaveNoInternetAccessBecauseofDifficulty 41PercentWhoUsetheCity’sWebsite 49

4

EXECUTIVESUMMARY

InMay2007,theMayor’sAdvisoryCouncilonClosingtheDigitalDividedefinedasetofgoalstoachievedigitalexcellence—universal,meaningfulparticipationintechnology—

throughouttheCityofChicagoandidentifiedfivedriversnecessarytoachievedigitalexcellence:effectivenetworkaccess,affordablehardware,suitablesoftware,digitaleducation,andevolvingmind‐sets.Theadvisorycouncilannouncedthat“Closingthedigitaldividemustbeseenaspart

ofthelargeropportunityforChicagototransforminstitutions,theeconomyandcommunities.Thisisaninclusivevision,seekingtoprovideuniversalmeaningfulparticipation,expandedeconomicprosperity,strengthenedcommunitiesandmoreeffectivegovernmentforall.”1

Buildingonavibrantnetworkofcommunityorganizationsandpublicagenciesthathavebeenaddressingtechnologydisparitiesinthecity,theadvisorycouncilcalledforanewfocusondigitalexcellencethatwouldbeintegraltothecity’sabilitytocompeteeconomicallyinthe

twenty‐firstcentury.Therefore,theCityofChicago–incollaborationwithnumerouspartnersfromtheprivatesector,highereducationandnon‐profitsectors–hassinceidentifiedstrategiestohelpChicagoresidentsandbusinessesachievedigitalexcellenceandrealizethevisionofthe

advisorycommittee.2

InanefforttofullyunderstandanddeterminethebarrierstotechnologyusefortheunderservedinChicago,theCityofChicago,theJohnD.andCatherineT.MacArthurFoundationandtheStateofIllinoisDepartmentofCommerceandEconomicOpportunitycommissionedthis

digitalexcellencestudytoidentifythelevelsoftechnologyuseacrossChicago. ThestudywasdesignedbyresearchersfromtheUniversityofIllinoisatChicagoandtheUniversityofIowa,andisbasedonarandom‐sampletelephonesurveyof3453Chicagoresidentsaged18andolder,

conductedbytheUniversityofIowaHawkeyePollinJuneandJuly2008.TheresultingdatadefinetherelevantgapsintechnologyuseinChicagoandprovideabaselineforevaluatingprogressinthefuture.ThereportwillhelptheCityandthebroadercommunitytostrategically

targetdigitalexcellenceeffortsandchangeconditionsandawarenessintheChicagocommunitiesthateitherdonothaveaccesstotechnology,ordohaveaccessbuthavenotachieveddigitalexcellence.

Althoughstudiesofinternetuseorserviceavailabilityhavebeenconductedinothercities,this

studyisthefirstofitskindinseveralways.Itbreaksnewgroundbyshowinghowneighborhoodsacrossthecitydifferintheiruseofthetechnologyaswellasbarrierstotechnologyuse.Clearly,thisisimportantfordesigningprogramsthatmeetthevariedneedsofcommunitiesacrossChicago,andthe

resultsindicatethesignificanceofneighborhoodcharacteristicsforunderstandingtechnologyopportunities.Theneighborhoodestimatesinthisstudyarebasedonsophisticatedmultilevelmodels,buttheresultsdonotrequireanunderstandingofthestatisticalanalysisbehindthem.Thefindings

presentedherearebasedonatelephonesurveythathasalargesamplethatisrepresentativeofthe

1Mayor’sAdvisoryCouncilonClosingtheDigitalDivide2007,3.2CityofChicago.http://www.cityofchicago.org/digitalexcellence.

5

citypopulation.Somecitieshaverecentlyusedmailsurveystotracktechnologyuse.3Thesearepoorinstrumentsforreachingconclusionsaboutthepopulationofacity,becausetheyhavelowresponse

ratesandthosewhotakethetimetofilloutandreturnthesesurveysarelikelytobethosewhoaremostinterestedinthetopic,andnottypicalofcityresidents.ResidentswhoareleastlikelytoreturnmailsurveysareindividualswithlimitedliteracyorEnglishproficiency,whoarealsoamongthose

currentlylaggingbehindintechnologyuse.Whiletherehavebeenafewothercitieswithtelephonesurveysontechnologyusethatprovidereliablerandomsamples,theanalysisofferedhereisuniqueinitsabilitytoidentifywhichdisparitiesaresignificant–thatis,whetherrace,oreducation,or

neighborhoodpovertymatterforinternetuse.Thisrequiresstatisticalanalysisthathasnotbeenusedinothermunicipalstudiesoftechnologyuse.Togetherwiththeneighborhooddata,thisinformationiscriticalfordevelopingappropriatepoliciesandtargetedprograms.

Thefollowingtableshowssomekeyindicators–thepercentageofChicagoresidentswhomeet

criteriafordigitalexcellence.

IndicatorsofDigitalExcellence Chicago(citywide)

Internetaccess

%whousetheinternetinanyplace 75%

%whousetheinternetdaily 60%

%whousetheinternetathome 69%%withbroadbandaccessathome 61%%whouseacellphonetoconnecttotheinternet 26%

%whousewirelessinternetinapublicplace 35%

Hardware/Software

%whohaveusedaCommunityTechnologyCenter 16%

%whohaveusedalibraryforinternetaccess 33%

%withahomecomputer 77%

Skills/Education

%whoknowhowtouseasearchengine 70%

%whoknowhowtouseemail 72%

%whoknowhowtouploadimagesorfiles 61%

%whoknowhowtocreateawebsite 25%

%whousetheInternetforwork 48%Continuedonnextpage

3ScarboroughResearch.NYCComparativeComputer&InternetPenetrationData.Datawerecollectedthroughamail‐basedsurveyconductedbetweenFebruary2006andMarch2007;resultsrepresent211,468nationwiderespondentsand4,407NewYorkCityrespondents.

6

Awareness/Mindset

%whohavelookedforjobinformationonline 50%

%whohavelookedforhealthcareinformationonline 64%

%whohavereadonlinenews 67%

%whohaveusedtheCity'swebsite 49%

%whohavetakenanonlineclass/training 31%

%whodonotusetheinternetathomebecausetheyarenotinterested

16%

Thefindingssummarizedbelowincludetheresultsofstatisticalanalysisofinternetuseand

accessinChicago(describingless‐connectedaswellasdisconnectedresidents);useofpublicaccesssitessuchaslibraries,communitytechnologycentersandwirelessnetworks;barrierstoaccess;internetuseforeconomicopportunity,civicengagement,healthcare,ande‐government;andsupportforpublic

andprivateeffortstofosterdigitalexcellence.

AccessandUse

Achievingdigitalexcellencerequiresthatresidentsbeabletousetheinternetregularlyandeffectively,includingafull,media‐richexperience.Wethereforeneedtounderstandwhousestheinternetinanyplace,whousestheinternetathome,andwhohashigh‐speedbroadbandconnections

athome.Residentswhoareonlinedailyaremostlikelytohavehomeconnections,andalsotohavebroadband,whichencouragesfrequentuse,awiderrangeofactivitiesonline,andatleastsomebasiclevelofskill.Broadbandisneededforfullconnectivitytothewebaswellasfrequentuse.

Seventy‐fivepercentofChicagoresidentsusetheinternet,atleastoccasionally.Sixpercentof

thecity’spopulation,however,isonlineattimesbutdoesnotusetheinternetathome,andanothereightpercentrelyonslowdial‐upconnections.Thismeansthatnearly40percentofChicagoresidentsareeitherentirelyofflineorhavelimitedaccess.Topromotefullandmeaningfulparticipationonline,it

isnecessarytobetterunderstandwhocomprisesboththedisconnectedandtheless‐connected.

Chicagoanswhoarestatisticallymorelikelytobeofflineorless‐connectedareolder,Latino,African‐American,low‐incomeandless‐educated.ResidentsofneighborhoodswithahighpercentageofAfrican‐AmericansandLatinosareparticularlydisadvantagedintermsofinternetuse.Thereare

somedifferences,however,indisparitiesforuseanywhereversushomeuseandbroadband.

OlderandLatinoresidentsareleastlikelytousetheinternetanywhere(althoughincome,education,andracearesignificantpredictorsforbeingonlineinanyplaceaswell).Thegapsbasedonracealone,forAfrican‐Americans,arerelativelysmallforinternetuseanywhere,andAfrican‐Americans

arenodifferentfromwhiteswhenweaccountfordifferencesintheneighborhoodswheretheylive.But,disparitiesbetweenAfrican‐Americansandwhitesarelargerforhomeaccess,indicatingthatmany

7

low‐incomeAfrican‐Americansareamongtheless‐connectedwholackhomeaccess,butgoonlineelsewhere.Incontrast,Latinos(andolderresidents)lagbehindinuseanywhere,homeaccess,and

broadband.

Incomeisthemostimportantfactorinfluencinghomeaccess,incontrasttointernetuseanywhere.Incomealsoaccountsforthelargestdifferencesbetweenbroadbandanddial‐upinternetusers,althoughthesearemoremodest,andthelargerbarrierisacquiringhomeaccessofanykind.This

suggeststhatgreateraffordabilityofinternetservices(aswellasneededhardwareandsoftware)willhelptoclosesomegapsinhomeaccessandbroadbanduse.

PublicAccessandWirelessUse

MostChicagoresidentsareawareofpublicplacestousetheinternet,andmostperceivethemasfairlyaccessible.Thirty‐threepercentofChicagoresidentshaveusedtheinternetatapubliclibrary,

and16percentsaythattheyhaveusedacommunitytechnologycenter.Halfofthosewhouselibrarieshavesoughthelpinfindinginformationonline,andaround30‐40percentoflibraryinternetpatronshaveproblemswithorlackcomputersorinternetconnectionsathome.

ChicagoresidentswhoaremostlikelytousepublicaccessatlibrariesareyoungerandAfrican‐

American.Low‐incomeresidentsarestatisticallymorelikelytousetheinternetatlibraries,butsoarebetter‐educatedChicagoresidents.Latinosareabout8percentmorelikelythannon‐Hispanicwhitestousetechnologyatthelibrary,controllingforotherfactors,butthiscomparesto14percentforAfrican‐

Americans.Homeinternetusersarealsomorelikelytousetechnologyatlibraries,indicatingthatmanylibrarypatronsgothereforhelporconvenience.Still,publiclibrariesarereachinglow‐incomeandminorityresidents,amongothers.

Examininguseofcommunitytechnologycenters(CTCs)inlow‐incomeneighborhoods,wefind

somesimilaritieswithlibraries–youngerandbetter‐educatedresidentsareamongthemostlikelyvisitors.ParentsaremorelikelytouseaCTC,asareAfrican‐Americanresidents.Respondentsresiding

inthepoorestneighborhoodsarealsomostlikelytousetheinternetatacommunitytechnologycenter.

Useofwirelessnetworksinpublicplacesisfairlycommon–35percentofChicagoresidentshaveusedthemtogoonline.Higherpercentagesofyoungpeopleaged18‐29usewirelesshotspots,andthisisalsotheagegroupwiththehighestpercentageconnectingtotheinternetthroughcell

phones.

BarrierstoAccess

Todevisebetterpublicandprivatepoliciesforaddressingdigitaldisparities,weaskedresidentswhytheydidnothavetheinternetathome.Thethreemostcommonchoicesselectedasthemainreasonfornothavinginternetaccessathomewerelackofinterest,cost,anddifficultyofuse.There

werecleardifferencesbetweendemographicgroupsinthereasonsforbeingofflineorless‐connected.

Thosewhoarenotinterestedareolder;ageaccountsforthelargestinfluenceoninterest.Othersmorelikelytosaytheyarenotinterestedarehigher‐incomeresidentswhodon’thavethe

8

internet,andless‐educatedresidents.African‐Americansarelesslikelythanwhitestosaythattheyarenotinterested.

Incomeisthemajorfactorexplainingconcernsaboutcost;residentswhosaytheycan’tafford

theinternetarestatisticallymorelikelytobelow‐income.Latinos,andtoalesserextent,African‐Americansarealsoamongthosemorelikelytocitecostasabarrier,controllingforotherfactors.

Chicagoresidentswhoperceivetheinternetastoodifficultareolder,less‐educatedandLatino.African‐Americansarelesslikelythanwhitestosaythattheinternetisdifficulttouse.

GiventhatAfrican‐Americanswhodonothavetheinternetathomearemorelikelytouseit

elsewhere,thisindicatesthatskillandinterestarenottheproblemforthisgroup,butcostisanissueforlow‐incomeAfrican‐Americans.Latinos,incontrast,perceivebothcostandskillasbarriers,andwerealsomorelikelytocitesomeofthelesscommonreasonsfornotbeingonlineaswell.ForLatinos,who

lagfurtherbehindininternetuse,thereappeartobemultipleobstacles.

ActivitiesOnline

Chicagoresidentswhoareonlineengageinavarietyofactivitiesthatillustratethepotentialoftheinternetforimprovingeconomicopportunity,civicengagement,health,andaccesstogovernmentservices.About48percentofChicagoresidents(63percentofemployedresidents)haveusedthe

internetfortheirjobs.Whileinternetuseatworkincreaseswitheducationandincome,itisnotconfinedtohighly‐skilledjobs.ThirtythreepercentofemployedChicagoresidentswithahighschooleducationusetheinternetatworkeitherdailyorseveraltimesperweek.

Themostcommonactivitiesonlineincludedinthesurveyarefollowingthenews(67percentof

Chicagoresidents)andseekinghealthinformation(64percent).Nearlyalloftheactivitiesweaskedaboutwereinfactquitecommon,onceresidentsareonline;forexample,91percentofChicago

internetusersreadnewsonline.Thisshowshowthoroughlydailytaskshavemigratedtotheinternet,andhowintegralthetechnologyisforaccesstoinformationandservices.

Whileyounger,better‐educatedandhigher‐incomeresidentsaremostlikelytoengageinanyactivityonline,therearesomedemographicandneighborhooddifferencesacrossactivities.African‐

Americansarestatisticallymorelikelythanwhitestolookforjobinformationonline.UsersoftheCityofChicagowebsitearemorediversethane‐governmentusersingeneral;womenandparentsareamongthemostlikelyCityofChicagowebsiteusers,andtherearenodifferencesbyraceandethnicity.

Residentsofhigh‐povertyneighborhoodsareamongthosewhoaremostlikelytousepublictransitwebsites.Womenandparentsareamongthemostcommonusersofhealthinformationonline,andtherearenodifferencesbetweenAfrican‐Americansandwhitesforhealthinformationuse.Young

peopleareamongtheresidentswhoaremostlikelytofollowthepoliticsornewsonlineortoaccesse‐government,althoughtheyaretraditionallyleastlikelytobeinterestedinpoliticsorpublicaffairs.Theinternetpresentsthepossibilityofcounteringcurrentinequalitiesforthedisadvantagedorless

engaged.Onceresidentsareonline,theinternetopensnewpossibilitiesfordemocratizinginformationandaccesstoservices.

9

PublicPolicySupport

Manyactivitiesareneededtoaddressthevariousdimensionsofdigitalexcellence,butonepolicyinitiativethathasbeendiscussedinChicagoandothercitiesistheprovisionofwirelessinternet

service.Chicagoresidentsexpressedstrongsupportforextendingwirelessaccessinthecity.Atotalof89percentofsurveyrespondentsfavoredsometypeofwirelesspolicy,andwirelessaccessavailablethroughoutthecitywasthepreferredoption(chosenby50percentofrespondents).Therewasalso

supportforalternativessuchaswirelessaccessinschoolsandlibrariesonly(26percent)orinlow‐incomeneighborhoodsonly(13percent).Whenaskedwhethertheywouldbewillingtopayasmalltaxorfeetoprovidewirelessinternetservice,60percentofrespondentsansweredthattheywould.Public

opinionclearlyfavorswirelessinitiativesasonewayofachievingwidespreaddigitalexcellence.

Conclusion

Thesurveyindicatesthatmanygapsininternetuseandaccesspersist,butthatthereareopportunitiesforprogress.Mostresidentsoflow‐incomecommunitiesarenotsimplyuninterestedingoingonline,suggestingthateffortstoprovideaffordableaccess,training,andtechnicalsupportshould

helptonarrowthesegaps.Chicagoresidentsarealreadyengagedinmanyactivitiesonline,forwork,politicalparticipation,healthcareandgovernmentservices.ThissurveyisintendedtoprovidevaluableinsightsforChicagoresidents,businesses,communityorganizations,educationalinstitutionsand

policymakerswhosharethevisionofpromotingdigitalexcellencethroughoutthecity.AsinternetuseexpandsinChicagothroughtechnologicaladvances,inclusivemarketactionsandthoughtfulpublicpolicy,moreresidentswillenjoythebenefitsofparticipatingfullyinsocietyonline.

10

PARTI.INTRODUCTION:WHYDIGITALEXCELLENCE?

“Transformingsocietyandeconomythroughdigitalexcellence”istheambitiousgoalthatwassetinMayof2007bytheMayor’sAdvisoryCouncilonClosingtheDigitalDivide.4Theadvisorycouncil

recognizedthatanetworkedcity,withwidespreadinformationtechnologyuse,enjoysadvantagesforattractinghigh‐wagejobs,growingtheeconomythroughinnovation,enrichingcommunitylifethroughinformationandcivicengagement,andrunninggovernmentmoreeffectivelyandefficiently.Thereis

growingevidenceofthebenefitsofinformationtechnologyuseforindividualsaswellassociety.Internetuseatworkincreaseswages,evenwhenwetakeintoaccountdifferencesinotherfactors,suchaseducationandoccupation.5Thisistrueevenforworkerswhohaveahighschooleducationorless,

andthewageincreaseforinternetuseatworkisslightlygreaterforAfrican‐AmericansandLatinosthanforwhitenon‐Hispanicworkers.Internetusealsopromoteshigherlevelsofpoliticalknowledge,discussionandpoliticalinterest,andmanystudieshavelinkedtheinformationandmobilizationcapacity

oftheinternetwithanincreasedlikelihoodofvoting.6E‐governmentusershaveimprovedinteractionswithgovernmentandenjoymoreconvenientaccesstoservices.7Thosewhoareexcludedfromparticipationonlinesufferfromunequalopportunitiesinboththeeconomicandcivicarenas,andthese

disparitiesinturnpreventcommunitiesfromrealizingtheirpotential.

Inclusioninsocietyonlinemeansmorethansimplyhavingsomewayofaccessingtheinternetonanoccasionalbasis.Publicpolicyoftentracksthepercentageofpeoplewhoreportevergoingonline,andthiscertainlyhassomevalue,asweshowhere.But,occasionaluseisnotsufficientto

achievedigitalexcellence,ortoparticipatefullyintheopportunitiesaffordedbytheinternet.Thisrequiresregularandeffectiveaccessandtheskillstousethetechnology.Nationalsurveyshaveshownthathomeaccessisimportantforfrequentuse.Mobiledevicesareincreasinglyexpandingthewaysin

whichtheinternetcanbeaccessed,buthomeinternetconnectionsremainthemostfrequentwaytogoonlineformostinternetusers.Broadbandorhigh‐speedconnectionsarenecessarytotakefull

advantageofcontentonline,especiallythemulti‐mediaandinteractivedimensionsoftheinternet.Broadbandisalsoassociatedwithmorefrequentinternetuseandawiderrangeofuses,andagreaterlikelihoodthatuserswilldevelopnecessaryskills.8Theskillsrequiredforeffectiveinternetusecanbe

describedinavarietyofways,9butcanbethoughtofastechnicalcompetenceandinformationliteracy.Technicalcompetenceistheabilitytousehardwareandsoftware.Informationliteracyonlineincludestheabilitytounderstandwhatinformationmightbeneededtosolveaproblem,tofinditonline,to

evaluateitsutilityandvalidity,andtoapplyit.Obviously,thisrequiresbasicliteracy(orreadingcomprehension)aswellassomeothereducationalcompetenciessuchascriticalthinking.Theinternetisareading‐intensivemedium,andthisisparticularlytrueforfindinginformationaboutpolitics,

governmentservices,healthcare,andjobs.

4Mayor’sAdvisoryCouncilonClosingtheDigitalDivide.2007.TheCitythatNetWorks:TransformingSocietyandEconomyThroughDigitalExcellence.5Mossberger,TolbertandMcNeal2008,chapter2.6Bimber2003;Krueger2002;TolbertandMcNeal2003;Mossberger,TolbertandMcNeal2008.7West2004;Welch,HinnantandMoon2005;TolbertandMossberger2006.8Mossberger,TolbertandMcNeal2008,chapter6.9See,forexampletheliteraciesidentifiedbyMarkWarschauer(2003)andJanVanDijk(2005).

11

TheMayor’sAdvisoryCouncil,whichincludedawiderangeofrepresentativesfromfoundations,businesses,community‐basedorganizations,universitiesandgovernment,definedthegoalofachieving

digitalexcellencethroughoutthecity.Accordingtothe2007reportoftheadvisorycouncil,thedriversofdigitalexcellenceare:

1. Effectivenetworkaccessthatishigh‐speed,reliable,affordableandavailableeverywhere.2. Affordablehardwarewithcapacitytoconnecttotheinternetandtapintothefullrangeofits

visualandotherresources.3. Suitablesoftwarethatmeetstheneedsofindividuals,families,business,andcommunities.4. Digitaleducationthatprovidesthetrainingandtechnicalsupportforuserstobecome

comfortableandproficient.5. Evolvingmind‐setsthatvaluelearning,connectingandcommunicatingthroughtechnology,

andthatrecognizethebusinessandotheropportunitiesofexpandinginternetparticipation.

(ExecutiveSummary,p.2,TheCitythatNetWorks:TransformingSocietyandEconomyThroughDigitalExcellence)

ThiscurrentreportprovidesaninitialviewofinternetuseandattitudesaboutinformationtechnologyinChicagotoassessareasofneedforprograms,policiesandtechnologicalinfrastructure.

Thereportprovidesuniqueinsightintowhereandwhya“digitaldivide”existsinalargemunicipalityandthedatacanbeusedtoinformpublicandprivatesectordecision‐makersandcommunityleaders.Wediscussinternetuseinavarietyofsettings,fromhomeaccesstopublicaccess,andtrendstoward

theuseofwirelessdevices.Beyondaccess,weexamineskillsintermsoftheabilitytousetheinternetfrequentlyandtoengageinavarietyofactivitiesonline.WeexplorethereasonswhysomeChicagoresidentsarenotonline,andwhatmightmotivatethemtousetheinternetinthefuture.

Thisreportisbasedonarandom‐sampletelephonesurveyofCityofChicagoresidents,conductedinJuneandJuly2008.ThestudywasdesignedbyresearchersattheUniversityofIllinoisatChicagoandtheUniversityofIowa,andthesurveywasconductedbytheUniversityofIowaHawkeyePoll.The

surveyyieldedasampleof3453respondentsfromChicago’s77communityareas.ThesurveywasconductedinSpanishandEnglish,andthecooperationratewas27percent,whichistypicalfortelephonesurveys.Thesampleofresidents18yearsandolderwasfairlyrepresentativeofChicago’s

population.Ofsurveyrespondents,45percentwerewhitenon‐Hispanic,31percentwereAfrican‐American,3percentAsian‐American,19percentLatinoand3percentotherormixedrace.Thepercentagesreportedinthisstudyareweightedtocorrectfordifferencesbetweenthesampleandthe

population.

PARTII.INTERNETUSEANDACCESS

InternetuseintheCityofChicagolooksremarkablyliketherestofthenation.Chicagoasawholeiskeepingpacewithnationalaverages,butasadiversecity,italsoreflectsthedisparitiesininternetusethatpersistnationwide.Byaddressingthesegaps,Chicagohasthechancetocreate

modelsfora21stcenturycitythatareforwarding‐lookingbothtechnologicallyandsocially.

12

Asofsummer2008,75percentofChicagoansusedtheinternet,incomparisonwithnationalfiguresof73percentduringthesameperiod.10SomecautionshouldbeexercisedincomparingChicago

withnationaldata,particularlyinreachingconclusionsbasedondifferencesofafewpercentagepoints.11So,whiletherearesomesmalldifferencesbetweenChicagoandthenation,thesoundestconclusionisthattheyarequitesimilar.

75%ofChicagoresidentsusetheinternetatleastoccasionally

Broadbandconnectsuserstotheinternetathigherspeedsthandial‐upmodem,andismostcommonlyavailablethrougheitheracablemodemordigitalsubscriberline(DSL).12Beyondspeed,

however,broadbandencouragesfrequencyofusebecauseitpermits“alwayson”connectionsthatdon’toccupyaphoneline.InChicago,61percentofthecity’spopulationhasabroadbandconnectionathome,incomparisonwith55percentoftheU.S.population.Thismayreflecttheinfrastructure

advantagesofamajorcityoverruralareas,althoughthedifferencesmaybepartlyduetosamplingforeitherthenationalorChicagosurvey.

61%havebroadband;60%goonlinedaily

Although75percentofcityresidentshavesomeexperiencewiththeinternet,only60percentusetheinternetonadailybasis.Frequentuseisimportantfordigitalexcellence,becausethosewhoareonlinefrequentlyaremorelikelytohavebothregularaccessandatleastsomebasicinternet

knowledgeandskills.

Frequentuseoccursathomeandwithbroadband;4in10facebarrierstofullaccessinChicago

Overall,25percentofChicagoresidentsarecompletelyoffline,another6percentusethe

internetattimesbutlackhomeaccess,and8percenthavemorelimitedandslowdial‐upconnectionsratherthanhigh‐speedbroadband.Approximately60percentofChicagoresidentshaveadequateaccess,butnearly40percenthavesomewhatlimitedornointernetaccess.Thus,4in10face

technologybarriersofvaryingdegrees.

Boththosewholackhomeaccessandbroadbandarelessfrequentinternetusers,andboththedisconnectedandless‐connectedtendtobeeconomicallydisadvantagedorolderresidents.ThelargestgapsarebetweenSpanish‐speakingandEnglish‐speakingresidents,asonly39percentofthosewho

respondedtothesurveyinSpanishusetheinternet,comparedto79percentofthosewhoansweredin

10May2008trackingsurvey,PewInternetandAmericanLifeProject,pewinternet.org11Random‐samplesurveyshaveamarginoferrorofplusorminusafewpercentagepoints.Whatthismeansisthatsurveysmayreflectthepeculiaritiesofthepeoplewhohappenedtorespond,tosomeextent.TheChicagosurveyreportedherehasamarginoferrorofplusorminus1.7percentagepoints.12TheFederalCommunicationsCommissiondefinesbroadbandasanyconnectionwithtransmissionspeedofatleasttwohundredkilobitspersecondinonedirection.

13

English,a40percentdifferencebasedonlanguageuse.Latinosasawholestandoutasamongtheleast‐connectedresidents.

Only39%ofrespondentswhotookthesurveyinSpanishusetheinternet

Thenextsectiononinternetaccessoffersamorecompletepictureofdisparitiesininternetuseinanyplace,homeaccess,andbroadbandconnections.Thesethreeelementsareneededto

understandtheneedsoftheless‐connectedaswellasthedisconnectedinChicago.First,however,weexplainhowtocompareandusedifferenttypesofinformationdiscussedinthereport.

ReadingtheResults

Inmanycasesinthisreportweusesimplepercentages,suchasthoseinthesectionabove.Weexaminethekeypointsinmoredepthusingstatisticalanalysis,butpresentthefindingsinwaysthatdo

notrequirethereadertohaveanystatisticalbackground.Thesemoreanalyticalfindingsarereportedin“WhatMatters”boxesthataresetapart.Simplepercentagescantelluswhatproportionofthepopulationdoessomething‐forexample,whatpercentageofLatinosusestheinternet.But,they

cannotexplainwhetherethnicityitselfhasanyeffectoninternetuse,orwhetheritisreallyeducationandincomethatexplaindifferencesbetweenLatinosandnon‐Hispanicrespondents.Todisentangletheeffectsofoverlappinginfluences,weuseamethodcalledmultivariateregression(orlogisticregression)

toisolatethefactorsthatarestatisticallysignificantpredictorsofinternetuse.Whatthisallowsustosay,forexample,iswhetherLatinoethnicityisasignificantpredictoroftechnologyusewhenwecontrolfortheeffectsofincome,educationandage.Themultivariateregressionmodelsarereportedinthe

appendix.Whennumbersarepresentedoutsideofthedesignated“WhatMatters”boxestheyarebasedonsimplepercentages.

Abriefwordisnecessaryonhowtousetheinformationinthe“WhatMatters”tables.Withinthesehighlightedtables,weincludeprobabilities(basedonmultivariateregressionanalysis)that

representthedifferencethateducationmakes,forexample,takingotherinfluencesintoaccount.Thesecanbereadaspercentages–whatpercentagedifferenceahighschooleducationmakesincomparisonwithacollegedegreewhenwecontrolforotherfactors.But,thekeydifferencebetween

ourprobabilitiesandsimplepercentagesisthattheprobabilitiesshowtheestimatedinfluenceofeducationalone,ifwecontrolforotherinfluences.

Anexamplewillhelptoshowthedifferencethismakes.Ifwelookatthesimplepercentagesforeducation,forexample,weseethat58percentofhighschoolgraduatesusetheinternetand88

percentofcollegegraduatesareinternetusers(a30percentdifference).Whenweusepredictedprobabilities,weseethatcollegegraduatesare15percentmorelikelythanhighschoolgraduatestobeinternetusers(not30percent),takingintoaccounttheeffectsofincome,age,gender,race,ethnicity,

andsoon.Bothtypesofinformationareuseful,buttheytellusdifferentthings.Withsimplepercentages,weknowhowcommonitisforcollegegraduatestobeonlineincomparisonwithhighschoolgraduates.Withpredictedprobabilities,weknowthateducationisresponsibleforonlysomeof

14

thisdifference(15percent),because,forexample,collegegraduatesalsohavehigherincomes,whichalsocontributetothe30percentage‐pointgapbetweenthesetwogroups.

Oneofthestrengthsofusingpredictedprobabilitiesistheabilitytocomparetheimpactof

differentfactorssuchaseducation,age,orincome.Welookatthepercentagedifferencemovingoneortwo“standarddeviations”fromthemean.Thisdefinesstatistically(basedonthedistributionofresponsesinthesurvey)what“low”or“high”isforeachexplanatoryfactor,anditsimplyallowsusto

comparetheinfluenceofloweducationversuslowincome,forexample.

Thediscussionafterthe“WhatMatters”tablesincludestheresultsofmultilevelmodels(whichcanbefoundinAppendixB).Theseareanalysesthatincludetheinfluenceofcharacteristicsofthesurroundingneighborhood(thecensustract)inadditiontothefactorsthatappearinthe“What

Matters”tables.Wemergedourindividual‐levelsurveydataontechnologywith2000datafromtheU.S.CensusBureauinformationforChicago’s700censustracts.13Multilevelregressionmodelswereestimatedtakingintoaccountindividual‐levelcharacteristics(arespondent’sage,race,ethnicity,

income,education,genderandparentalstatus),aswellasneighborhoodorcensustractlevelcharacteristics(percentLatinopopulation,percentblackpopulation,percentAsianpopulation,percentresidinginpovertyandpercentofthepopulationwithahighschooldegreeorhigher).Thesemultilevel

models(includingneighborhoodcharacteristics)wereusedtoestimateinternetuseforeachofthe77communityareasandeachofthecensustractsinChicago.Communityareamapsincludedinthisreportarebasedonthesemodels,aswellasthediscussionoftheimpactofcommunitycharacteristics

inthetext.14

Theresultsoftheregressionanalysisarehighlightedinthe“WhatMatters”tables,and

discussedinthetextimmediatelyafterwards.WeconsiderthisthemostreliablewaytounderstanddisparitiesintechnologyuseintheCityofChicago.Thepredictedprobabilitiesinthe“WhatMatters”

tablesprovidethemostaccuratepicture,isolatingtheeffectsofoverlappinginfluences.Thisoffersbetterguidanceforpolicy,toknowthetrueimpactofeducationversusincome,forexample,onthedisparitiesthatareevident.

Topromotedigitalexcellence,policy‐makersandcommunityleadersneedtotargetboththose

whoneverusetheinternetandthosewhohavelimitedexperienceonline.Regularandeffectiveuse13Anewvariablewascreatedthatwasthepercentageofthecommunityarea’spopulationthatresidesinthecensustract(totalpopulation/communityareatotalpopulation).Foreachexplanatory/predictorvariable(let'ssaypercentLatino),thefollowingformulawasapplied:(percentLatino/100)*[multipliedby]theweightedpopulationvariable.Thesecensustractvalueswerethencollapsedandsummedtogetthetotalcommunityareavalue.Thenewcommunityareageographiccharacteristicsweremergedbackintotheindividuallevelsurveydata.14ThesemultilevelmodelsprovideaccuratepointestimatesofthepercentofChicago’spopulationwithhomeinternetaccessatthegeographiclevel.ThecensustractlevelinformationwasalsoaggregatedtothecommunityareatoprovideestimatesoftechnologyuseandaccessforChicago’s77communityareas.Foreachdependentvariable,twomultilevelmodelswereestimated:onecombiningthesurveydatawithneighborhood(censustract)information,andonecombiningthesurveydatawithcommunityareageographicinformation.Thesepointestimatescanbereadaspercentages,buttheytakeintoaccountmultipleindividuallevelandgeographiccharacteristicslistedabove.Weconsiderthesemodelsfullyspecified.

15

requireshomeaccessandhigh‐speedconnections,asthissectiondemonstrates.Wecompareinternetuse,homeaccess,andbroadbandinthefollowinganalysis,toshowwherethegreatestgapsexist.

WhatMattersTableA:WhoUsestheInternetinAnyPlaceandWhoHasHomeAccess?

Thefactorslistedbelowhaveastatisticallysignificantinfluenceoninternetuse.Aplussign(+)indicatesincreasedprobabilityofinternetuse,andaminussign(‐)indicatesdecreasedchancesofinternetuse.InternetUseinAnyPlace InternetUseatHomeAge(‐) Income(+)Latino(‐) Age(‐)Income(+) Education(+)Education(+) Latino(‐)African‐American(‐) African‐American(‐) Parent(+)ReadingPredictedProbabilities:Afemale,whitenon‐HispanicChicagoresidentwithnochildrenandaverageage,income,andeducationhasa90percentprobabilityofusingtheinternetinanyplaceandan81percentchanceof

usingtheinternetathome.Resultsshownbelowindicate,forexample,thatarespondentwhoisLatino,butotherwisethesame,hasonlya72percentchanceofusingtheinternetanywhere.Latinoethnicityalonemakesan

18percentdifference(holdingotherfactorsconstant).Readthenumbersnotinparenthesesaspercentages.

UseInternetinAnyPlace UseInternetatHomeWhitenon‐Hispanic(Baseline) .90(.01) .81(.02)Latino .72(.03) .71(.03)DifferenceLatinovs.White ‐.18 ‐.10Black .84(.02) .71(.02)DifferenceBlackvs.White ‐.06 ‐.10AnnualIncome VeryLow($0,‐2SD) .61(.04) .40(.04)Low($10,000‐$20,000,‐1SD) .79(.02) .62(.03)Mean/Average($40,000‐$50,000) .90(.01) .81(.02)High($75‐$100,000,+1SD) .96(.01) .91(.01)VeryHigh(morethan$150,000,+2SD) .98(.00) .95(.01)DifferenceLowtoHigh +.17 +.29EducationLevel LessthanHS .70(.03) .59(.03)HighSchoolGraduate .79(.02) .68(.03)SomeCollege .91(.01) .81(.01)CollegeGraduate .94(.01) .86(.01)GraduateDegree .96(.01) .90(.01)DifferenceHStoCollege +.15 +.18Ageofrespondent Veryyoung(18yrs,‐2SD) .99(.00) .95(.01)Young(31yrs,‐1SD) .97(.00) .91(.01)Mean/Average(49yrs) .90(.01) .81(.02)Old(,67yrs,+1SD) .70(.02) .63(.02)Veryold(85yrs,+2SD) .37(.03) .40(.03)DifferenceYoungtoOld(27‐67yrs) +.27 +.28

Note:PredictedprobabilitiescalculatedwithClarifySoftwarefromthelogisticregressionmodelsreportedinAppendixTables1and2.Probabilitiesestimatedwithcontrolvariablessetatmeanormodalvalues.Standarderrorsoftheprobabilityestimateinparentheses.

16

InternetUseinAnyPlace

“WhatMatters”TableA,column1,showsthatChicagoreflectsthesamegapsininternetusethatarepresentnationally,butthatracialdisparitiesareconsiderablysmallerthanthosebasedonLatinoethnicity.Respondentswereasked“DoyouusetheInternetinanyplace?”andcouldrespond

“yes”or“no.”Thosewhoarelesslikelytousetheinternetinanyplaceareolder,Latino,African‐American,lower‐income,andless‐educatedresidents.Asinnationalsurveyssince2000,womenarenolesslikelytobeonlinethanmen.15

• Thelargestgapsarebasedonage,consistentwithnationalsurveys.Withallotherfactorsheld

constant,ayoungperson(31yearsofage,whichisminusonestandarddeviationfromthemean)hasa97percentprobabilityofinternetuseingeneral,comparedtoa70percentprobabilityforanolderperson(67yearsofage,whichisplusonestandarddeviationfromthe

mean).Thisisa26percentdifferencebasedonagealone.

• Thenextlargestgapisbasedonethnicity.Holdingconstantallotherfactors,Latinosare18percentlesslikelytousetheinternetthannon‐Hispanicwhites.ThiseffectdoesnotreflectlowerincomesofLatinoscomparedtowhites,butisbasedonethnicityaloneandSpanish

languagebarriers.LatinoswhochosetoconductthesurveyinterviewinSpanishratherthanEnglishhaveparticularlylowratesofinternetuse.16

• Incomeranksnextininfluencinguse.Thepoor(withannualincomesofbetween$10,000and$20,000peryear,whichareminusonestandarddeviationfromthemean)are17percentless

likelytousetheinternetingeneralthanthosewithhigherincomes($75,000‐$100,000peryear,whichisplusonestandarddeviationfromthemean).

• Educationgapsarenearlyasimportantasincome.Holdingallotherfactorsconstant,ahighschoolgraduateis15percentlesslikelytousetheinternetthanacollegegraduate.

• African‐Americansare6percentlesslikelytousetheinternetthanwhites.Thisgapissmaller

thanhasbeenreportedbyearliernationalsurveys,showingeitherimprovementoverthepastfewyearsorgreaterprogressinnarrowingracialgapsinChicagoinparticular.17WhilemanyAfrican‐Americansarestilloffline,raceaccountsforasmallerpartoftheexplanationfortheir

lackofinternetusethandisparitiesinincomeandeducation.

ThoseleastlikelytobeonlineareolderandLatinoresidents

15KatzandRice2002;Mossberger,TolbertandStansbury2003.16NationalstudiesofLatinosshowthatlanguageisasignificantfactorforinternetuseamongLatinos,althougheducationisalsosignificant.SeeFoxandLivingston2007.17The2003CurrentPopulationSurveyconductedbytheU.S.BureauoftheCensusshowslargergapsbasedonrace,forexample,asdiscussedinMossberger,TolbertandMcNeal(2008),chapter5.

17

HomeInternetUse

Respondentswerealsoasked“Doyouusetheinternetathome?”18Column2ofWhatMattersTableAshowstheprobabilityofhavingtheinternetathome.Theresultsparallelthefindingsfor

internetuseingeneral,butwithsomeimportantdifferences.

Incomeismostimportantforhomeuse,andtheeffectsofincomearesubstantial

Thesesimulationsdemonstratethatincomecausesthelargestgapsinhomeinternetaccess,followedbyage,whileageandLatinoethnicitycausethelargestgapsinuseoftheinternetinanyplace.

• Thelargestgapinhomeinternetaccessisbasedonarespondent’sfamilyincome,notage.Thepoor(onestandarddeviationbelowthemean)are29percentlesslikelytohaveaccessthanhigher‐incomeresidents(plusonestandarddeviationabovethemean),allelseequal.Poor

Chicagorespondentshaveonlya62percentprobabilityofhavinghomeaccesscomparedtoa91percentprobabilityofhomeaccessforhigher‐incomeresidents.Low‐incomerespondentsare29percentlesslikelytohavehomeaccess,butonly17percentlesslikelytobeinternet

usersanywhere.Thismakessense,giventheinvestmentinhardwareandsoftwareandthemonthlyinternetbillsassociatedwithhomeinternetuse.

• Disparitiesinhomeaccessbasedonagearesignificant.Theyoungare28percentmorelikelytohavehomeinternetaccessthanolderrespondents;a31yearold(onestandarddeviationbelow

themean)hasa91percentprobabilityofhavinghomeaccesscomparedtotheanolderindividual(67years,onestandarddeviationabovethemean),whohasonlya63percentprobabilityofhavingtheinternetathome.

• Educationalattainmentisthenextlargestpredictorofgapsininternetuseathome.Acollege

graduateis18percentmorelikelytohavetheinternetathomethanaresidentwithonlyahighschooldiploma,allelseequal.

• BothAfrican‐AmericansandLatinosare10percentlesslikelytohavehomeInternetaccessthanwhitenon‐Hispanics.GapsforhomeaccessarewiderforAfrican‐Americansthanforinternet

useanywhere,suggestingblacksaremorelikelytotakeadvantageofinternetaccessoutsidethehome.

• Parentsarealsomorelikelytohavetheinternetathome,althoughtheyarenotmorelikelytobeinternetusers.

18Responseswerecoded1foryesand0forno.

18

NeighborhoodVariationforInternetUseAnywhereandatHome

Tounderstandhowneighborhoodsvary,multilevelmodels(AppendixB)canbeusedtoincludethefactorslistedabove,aswellascensusdataonneighborhoodcharacteristics:poverty;percentageof

highschoolgraduates;andpercentageofAfrican‐Americans,LatinosandAsian‐Americans.Neighborhoodshaveasignificantinfluenceonbothinternetuseanywhereandhomeaccess:

InternetUseinAnyPlace

• ResidentsofneighborhoodswithhighpercentagesofAfrican‐AmericansorLatinosarelesslikelytousetheinternetanywhere.

• Infact,thedifferencesbetweenAfrican‐Americansandwhitesinuseanywherecanbeexplainedbyneighborhoodfactors–byresidenceinhigh‐povertyminorityneighborhoods.Thisisconsistentwithnationalresearch.19

HomeInternetUse

• ResidentslivinginneighborhoodswithahigherpercentageofLatinosarealsolesslikelyto

havehomeaccess.

• ResidentsofcommunitieswithahighproportionofAsian‐Americanshavehigherratesofhomeuse.

Thereisconsiderablevariationacrossthecommunityareasforbothinternetuseanywhereand

alsointernetuseathome.Themultilevelmodelswereusedtoestimateinternetuseanywhereandhomeaccessforthe77CommunityAreasinChicago.Theseareasaretraditionallyusedformanyplanningpurposes.

Themaponthenextpagehighlightsthevariationinhomeinternetuserevealedbythemodels.

Redareashavethelowestratesofhomeinternetuseinthecity,andblueareashavethehighestrates.Communityareascoloredinblueareestimatedtohaveatleast65percentofresidentswhohaveinternetaccessathome.Thisisalittleunderthecity‐wideaverageforhomeaccessof69percent.

Areasinbluethereforeapproachorexceedthecityaverage.Areaswithlowerinternetuse(redoryellow)areprimarilylow‐incomeandAfrican‐AmericanorLatinoneighborhoods.Homeaccessisanimportantsteptowarddigitalexcellencebecauseitencouragesfrequentinternetuse,andmanyof

Chicago’ssouthandwestsidecommunityareasaredisadvantagedinthisrespect.

19Mossberger,TolbertandGilbert2006

19

20

BroadbandAccess

Broadbanduseisimportantforanumberofreasons.Today,manyapplicationsrequirebroadbandspeedsforfullconnectivity.Downloadinggraphicsanddocuments,submittingonlineforms,making

commercialtransactionsandpayments,accessingmanywebsites,orviewingvideosonlinecanbedifficultwithdial‐upaccess.Theslowerspeedsareoftenfrustrating,discouragingfrequentinternetuse.Nationalstudieshaveshownthatthosewhousebroadbandaremorelikelytobefrequentusers,to

engageinalargervarietyofactivitiesonline,andtohavehigherlevelsofinternetskill.20Broadbandisclearlythestandard,asmostChicagoresidents(61percent)havehigh‐speedconnectionsathome.

Thefactorsthatinfluencebroadbandusearesimilartothosethataffectinternetusemoregenerally,exceptthatdifferencesbetweenbroadbandanddial‐upusersarelesspronouncedthan

betweenthosewhodonothavehomeaccessatallandthosewhodo.Homeaccessisthelargerhurdle,butbroadbandcostsfurtherrestrictfullaccessforsome.

Thegapsbasedonincomearethelargest,althougholderresidents,less‐educatedresidents,andLatinosareamongtheleastlikelytohavehigh‐speedinternetratherthandial‐upathome.(Seemodelsonthenextpage.)

• Low‐incomeresidentsare12percentlesslikelytohavehigh‐speedconnectionsratherthandial‐up,butare29percentlesslikelytohavehomeinternetaccessofanykind(seeWhatMattersTableA).Giventhatcomparisonsarebetweendifferenttypesofhomeinternetusers,disparitiesaremoremodest.Butincomeisthemostimportantfactorexplainingdifferencesbetweendial‐upandbroadbandusers.

• Agerankssecondinitsinfluenceonbroadbanduse.Anolderrespondent(age67)is10percentlesslikelythanayoungChicagoresident(age31)tohavebroadbandratherthandial‐upathome.

• Latinosare6percentlesslikelythannon‐Hispanicwhitestohavebroadbandathome,controllingforfactorsotherthanethnicity.

• Educationissignificant;collegegraduatesare5percentmorelikelytohavehigh‐speedconnectionsthanhighschoolgraduates.

• OnenotabledifferencecomparingbroadbandwithinternetuseingeneralisthatAfrican‐Americansarejustaslikelytohavebroadbandasnon‐Hispanicwhites,controllingfordifferencesinincomeandeducation.Nationally,therewasadramaticincreaseinbroadbandadoptionsbyAfrican‐Americansin2006,ashigh‐speedsubscriptionratesfellslightlyinpriceandaslessexpensiveDSLbecamemorewidelyavailable.21Incontrast,broadbandusehasnotgrownsubstantiallyamongthepoor,evenwithpricedeclines.22

Incomemattersmostforbroadbandaccess

WHATMATTERSTABLEB.WhoHasaBroadbandConnectionComparedtoDial‐up?20Horrigan2005;Mossberger,TolbertandMcNeal2008,chapter6.21Horrigan2006.22Horrigan2008.

21

Thefactorslistedbelowarethestatisticallysignificantdifferencesbetweeninternetuserswithbroadbandanddial‐upconnectionsathome.Aplussign(+)indicatesincreasedprobabilityofbroadbanduse,andaminussign(‐)indicatesdecreasedchancesofbroadbanduse.BroadbandUsevs.Dial‐Up Income(+) Age(‐)Education(+)Latino(‐)Asian‐American(+)ReadingPredictedProbabilities:Afemale,whitenon‐HispanicChicagointernetuserwithnochildrenandaverageage,income,andeducationhasa92percentprobabilityofusingbroadband.Resultsshownbelowindicate,for

example,thataninternetuserwhoisLatino,butotherwisethesame,hasonlyan87percentchanceofusingbroadband.Latinoethnicityalonemakesa5percentdifference(holdingotherfactorsconstant).Readthe

numbersnotinparenthesesaspercentages.

BroadbandConnectionversusDial­upAccess

Baselinea:Whitenon‐Hispanic .92(.01)Latino .87(.02)DifferenceLatinovs.White ‐.05Black .92(.01)DifferenceBlackvs.White 0Income VeryLow($0,‐2SD) .74(.05)Low($10,000‐$20,000,‐1SD) .83(.02)Mean/Average($40,000‐$50,000) .92(.01)High($75‐$100,000,+1SD) .95(.01)VeryHigh(morethan$150,000,+2SD) .96(.01)DifferenceLowtoHigh +.12EducationLevel LessthanHS .84(.03)HighSchoolGraduate .87(.02)SomeCollege .91(.01)CollegeGraduate .93(.01)GraduateDegree .94(.01)DifferenceHStoCollege +.06Ageofrespondent Veryyoung(18yrs,‐2SD) .96(.01)Young(31yrs,‐1SD) .95(.01)Mean/Average(49yrs) .92(.01)Old(,67yrs,+1SD) .85(.02)Veryold(85yrs,+2SD) .75(.03)DifferenceYoungtoOld(27­67yrs) ‐.10

Note: Predicted probabilities calculated with Clarify Software from the logistic regression models reported in

AppendixTable3.Probabilitiesestimatedwithcontrolvariablessetatmeanormodalvalues.Standarderrorsoftheprobabilityestimateinparentheses.

NeighborhoodVariationforHomeBroadband

22

Multilevelmodelsthatintroduceneighborhoodcharacteristicsareconsistentwiththeaboveresults(seeAppendixB),exceptthat:

• Broadbanduseissignificantlymorelikelyforresidentsofneighborhoodswithhighereducationalattainment.

Themodelsestimatingbroadbanduseineachofthecommunityareasarenotmappedhere,butwecanseetherangeacrossChicagobyexaminingthecommunityareaswiththehighestandlowestratesofbroadbandpenetration.LincolnPark(CommunityArea7)ranksatthetopforbroadbanduse,with90percentofthepopulationestimatedtobehigh‐speedinternetusers.LincolnParkisacommunityareawithhighincomeandeducationalattainment.AttheotherendofthespectrumisSouthLawndale(CommunityArea30),whereonly25percentofthepopulationhasbroadbandaccordingtothemultilevelestimates.

SummingUpWhatMattersforUseandAccess

Therearesomecommoninfluencesoninternetuse,homeaccess,andbroadbandaccess,withage,income,education,race,andethnicityappearingassignificantfactorstovaryingdegrees.Latinoandolderresidentsareleastlikelytousetheinternetanywhere,andgapsbasedoneducationandincomeranknext.African‐Americansaresignificantlylesslikelythanwhitestousetheinternet,butraceaccountsforthesmallestdisparityininternetuse.Infact,whenwetakeneighborhoodfactorsintoaccount,African‐Americansarenolesslikelytobeonlinethanwhites(butresidentswholiveincommunitieswithhighpopulationsofAfrican‐AmericansorLatinosaredisadvantagedininternetuse).Thissuggeststhatcommunityenvironmenthassomeindependenteffect,beyondindividualcharacteristics.

Forhomeinternetuse,alloftheindividual‐levelfactorsaresignificant,buttheeffectsofincomearemagnified–accountingfora29percentdifferenceinhomeaccessincomparisonwitha17percentdifferenceforuse.African‐Americansexperiencewidergapsforhomeaccessthanforinternetuse,indicatingthatmanyrelyonuseoutsidethehome.LivinginaneighborhoodwithahighpercentageofLatinosalsodecreasesthelikelihoodofhomeinternetuse.Incomeismostimportantforexplainingthedifferencebetweendial‐upandbroadbandusers,althougholder,less‐educatedandLatinoresidentsarealsosignificantlylesslikelytohavebroadbandratherthandial‐up.

DescribingtheLess‐Connected

Policyattentionisoftenfocusedonlyoninternetuse,butthequalityofaccessisalsoimportantfordigitalexcellence.Thosewhoarenotonlineathomeusetheinternetlessfrequently;83percentofChicagoresidentswithhomeaccessareonlinedailyincontrastwith7percentofthosewhodonothavehomeaccess.Similarly,88percentofthosewhohavebroadbandathomeareonlineeveryday,comparedto54percentofthosewhohavedial‐upinternetaccess.

83%ofChicagoresidentswithhomeaccessareonlinedaily,butonly7%ofthosewithouthomeaccessare

Takingacloserlookatthe6percentofcityresidents(10percentofinternetusers)whousetheinternetbutdonothavehomeaccess,wecanseesomerevealingpatternsinthesimplepercentages

thatroundoutthepicturewegetfromtheprobabilities.Theseindividualsareamongtheless‐connectedinternetusersinthecity,fortheyrelysolelyonwork,school,publicaccess,orthehomesof

23

friendsandrelatives.Asthenumbersshow,theyarelow‐incomeresidents.ThesimplepercentagesinTable1showtheproportionsofinternetusersindifferentincomegroupswhodonothavehome

access.

TABLE1.INTERNETUSEWITHNOHOMEACCESSBY2007TOTALFAMILYINCOME

%ofinternetusers withouthomeaccess

Lessthan$5,000 28%5tounder$10,000 26%10tounder$20,000 17% 20tounder$30,000 10%30tounder$40,000 9%40tounder$50,000 5%50tounder$75,000 5% 75tounder$100,000 3% Over$100,000 2% Totalforinternetusers 10%

Amongthosewhodonothaveinternetaccessathome,therearesomedifferencesbasedonraceandethnicity.AhigherpercentageofAfrican‐AmericansandAsian‐Americanswhodonothavethe

internetathomegoonlineanyway.Only22percentofwhiteswithouthomeaccessareinternetusers,whereas27percentofAfrican‐Americanswithouthomeaccessare,and29percentofAsian‐Americanswithouthomeaccessgoonlineelsewhere.Incontrast,only16percentofLatinoswhodonothave

homeaccessareinternetusers.23ThefindingsforAfrican‐AmericansareconsistentwithastudyinnortheastOhiothatdiscoveredthatpoorAfrican‐Americanneighborhoodshadhigherproportionsofinternetuserswholackedhomeaccessthanpoorwhitecommunities.Thisdemonstratedefforttogo

onlinedespitealackofconvenientaccess,butitalsomeansthatpoorAfrican‐Americanstendtobeless‐connectedinternetusers.24Theseresultsalsosuggestthataffordabilityistheissueforlow‐incomeAfrican‐Americansratherthanalackofawarenessofthebenefitsofbeingonline.InChicago,itisclear

thatAsian‐AmericansarealsomorelikelytogoonlinewithouthomeaccessandthatLatinosareleastlikelytobeinternetusersawayfromhome.

Residentswhohaveinternetaccessathomehaverealadvantagesineaseofuse,andgreaterprospectsfordevelopingtheskillneededforinternetuseonthejob.High‐speedconnectionsfacilitate

frequentinternetuseandthemigrationoftasksonline,sothatbroadbandusersengageinagreaterrangeofactivitiesonlineandgaingreaterfamiliaritywiththeinternet.Convenient,qualityaccessathomeallowsresidentstofollowpoliticsorneighborhoodeventsonline,helpchildrenwithhomework

assignments,learnmoreabouthealthissues,searchforjobs,takeanonlineclass,orstartasmallbusiness.Manyoftheseactivitiesaredifficulttosustainwithonlyintermittentaccess.

23Thesearesimplepercentagesratherthanprobabilitiesbasedonregressionanalysis.24Mossberger,KaplanandGilbert2008

24

Still,publicaccessinlibrariesandcommunitytechnologycenterscontinuestobeanimportantresourceforthosewhowouldotherwisenotbeonlineatall.Inaddition,publicaccesssitescanprovide

trainingortechnicalsupporttoencourageinternetuseanddevelopdigitalskills.Thenextsectionexaminesinternetuseinpublicplaces,particularlylibrariesandcommunitytechnologycenters,aswellaspublicwirelessaccess.

PARTIII.PUBLICANDWIRELESSACCESS

Inadditiontobusiness‐sponsoredhotspots(i.e.cafesandrestaurants),theCityofChicago,its

sisteragencies,andanarrayofnonprofitorganizationsofferpublicinternetaccessinplacesacrossthecity.WirelessInternetZones,freeWi‐FihotspotsprovidedbytheCityasaservicetoresidentsandvisitors,areavailableatpublicplacessuchasMillenniumPark,theChicagoCulturalCenter,andRichard

J.DaleyPlaza.FreeWi‐Fiandcomputerandinternetaccessareavailableatall79ChicagoPublicLibrarybranches.LibrariansandCyberNavigatorsprovidevaluabletechnologyassistanceandtrainingtoresidents.TheCity’sSeniorCenters,YouthCareerDevelopmentCenters,andWorkforceDevelopment

Centersprovidecomputerandinternetaccessandtechnologytraining;and,theDepartmentofBusinessAffairsoffersfreemonthlytechnologytrainingthroughitsbusinesseducationworkshops.TheChicagoHousingAuthorityprovidesresidentswithcomputerandinternetaccess.SomeChicagoPublicSchools

provideparentswithcomputerandinternetaccessandtraining.

Nonprofitorganizationsalsooffertechnologyaccess,trainingandsupportatcommunitytechnologycenters(CTCs)locatedprimarilywithinlow‐incomeneighborhoods.Thereisnosinglelistingofsuchcenters,butCTCNetChicagohasapproximately40memberorganizations.25TheStateofIllinois

providedsupportforapproximately100communitytechnologycentersinChicagothroughtheDigitalDivideInitiativegrantprogramduring2008.Thisprogram,administeredbytheIllinoisDepartmentof

CommerceandEconomicOpportunity,supportstraininginbasiccomputerskills,vocationalskillsrelatedtotechnologyoccupations,literacyskills,computerapplicationsforsmallbusinesses,andassistivetechnologyforindividualswithdisabilities.

GiventheeffortsofCTCsandlibrariesinChicago,weaskedresidentswhetheritwaseasyor

difficulttogetto“placesinyourcommunitywithpublicaccesstotheinternet,likealibraryor

communitytechnologycenter.”Mostrespondentsbelievethatitisrelativelyeasytoreachpublicaccesssites,with76percentsayingthatitiseitherveryeasyorsomewhateasytogetpublicinternetaccesswithintheircommunities.Fourteenpercentfeltthatitwaseithersomewhatorverydifficultto

usepublicaccess,and10percentdidnotknow.Chicagoresidentsseemtofeelpositivelyabouttheavailabilityofpublicaccessintheircommunities.Howmanyhaveusedpublicaccesssites,andwhatarethecharacteristicsofpublicaccessusers?Wediscusslibrariesinmoredetailnext,andthenpresent

findingsoncommunitytechnologycentersandwirelessaccess.

InternetUseatLibraries

25Seehttp://www.connectchicago.net/ChicagoInitiatives.aspx

25

Useofpublicaccesstechnologyatlibrariesismostcommon,as33percentofthecity’sresidents(44percentofChicagointernetusers)havegoneonlineataChicagoPublicLibrarybranch.Because

librarieshavefunctionsotherthantechnologyuse,internetuserstheremaysimplycheckemailorotherinformationwhilevisitingthelibraryforbooksorothermedia.Toestablishwhatproportionoflibraryinternetusersdependeduponpublicaccess,weaskedaboutreasonsforusingtheinternetatthe

library.TheresultsaredisplayedinTable2below.

One‐thirdofChicagoresidentsusetheinternetatapubliclibrary

TABLE2.REASONSFORUSINGTHEINTERNETATTHECHICAGOPUBLICLIBRARY

Percentagesforlibraryinternetusersonly;multipleresponsespossible

Convenience 70%Needhelptofindinformation 46%Computerathomenotworking 39%Nocomputerathome,computerslow 33%Nointernetathome 29%Totakeaclass 25%Totakemychildrenforhomework 25%Needhelptousecomputer 17% Librariesclearlyservethosewhohavelimitedaccessorwhoneedhelp,buttheyhaveabroaderaudienceofcasualusersaswell.Convenienceisthemostimportantreasonforinternetuseatthe

library(at70percent),indicatingthatnotalllibrarypatronslackhomeinternetaccess.Still,between30and40percentciteproblemswithcomputersorconnectivityathomeasareasontoseekoutthelibrary.Obtaininghelpinfindinginformationisthemotivationfornearlyhalfofthosewhousethe

internetatpubliclibraries,althoughtheneedforhelpwithcomputerhardwareislessprevalent(at17percent).Twenty‐fivepercentoflibrarypatronsusethelibraryforformalinstructionthroughcomputerclassesorforchildren’shomework.

Regressionanalysisallowsustobetterunderstandthecharacteristicsoftechnologyusersat

libraries,introducingfactorssuchasparentalstatus,awarenessofpublicaccessintheneighborhood,easeofaccesstoapublicinternetfacilityintheneighborhood,andwhetherornottherespondentusestheinternetathome.TheresultsaredisplayedinWhatMattersTableCbelow.

26

WHATMATTERSTABLEC.WhoUsestheInternetatPublicLibraries?

Thefactorsbelowarestatisticallysignificantinfluencesontechnologyuseatpubliclibraries.Aplussign(+)indicatesincreasedprobabilityoflibraryinternetuse,andaminussign(‐)indicatesdecreasedchancesoflibraryinternetuse.LibraryInternetUse

Age(‐) AwarenessofOtherPublicAccess(+)HomeInternetUse(+) Education(+) African‐American(+) PerceivedEaseofUse(+)Income(‐) Latino(+)ReadingPredictedProbabilities:Afemale,whitenon‐HispanicChicagoresidentwithnochildren,internetaccessathome,whoisawareofpublicaccessintheneighborhoodandperceivesveryeasyaccesstothepublicinternetfacility,andwhohasaverageage,income,andeducationhasa39percentprobabilityofusingtheinternetatthelibrary.Resultsshownbelowindicate,forexample,thataresidentwhoisLatino,butotherwisethesame,hasa47percentchanceofusingtheinternetatthelibrary.Latinoethnicityalonemakesan8percentdifference(holdingotherfactorsconstant).Readnumbersnotinparenthesesaspercentages. UseInternetattheLibrary

Whitenon‐Hispanic(Baseline) .39(.02)Latino .47(.03)DifferenceLatinovs.White +.08Black .53(.03)DifferenceBlackvs.White +.14AnnualIncome VeryLow($0,‐2SD) .52(.04)Low($10,000‐$20,000,‐1SD) .46(.03)Mean/Average($40,000‐$50,000) .39(.02)High($75‐$100,000,+1SD) .33(.02)VeryHigh(morethan$150,000,+2SD) .29(.03)DifferenceLowtoHigh ‐.13EducationLevel LessthanHS .28(.03)HighSchoolGraduate .32(.03)SomeCollege .40(.02)CollegeGraduate .44(.02)GraduateDegree .48(.03)DifferenceHStoCollege +.12Ageofrespondent Veryyoung(18yrs,‐2SD) .66(.03)Young(31yrs,‐1SD) .55(.03)Mean/Average(49yrs) .39(.02)Old(,67yrs,+1SD) .25(.02)Veryold(85yrs,+2SD) .14(.02)DifferenceYoungtoOld(27­67yrs) ‐.30Donotuseinternetathome .17(.02)Useinternetathome .39(.02)DifferenceNoUseatHometoUse +.22

CONTINUEDONNEXTPAGE

27

WHATMATTERSTABLEC(CONTINUED)WhoUsestheInternetatPublicLibraries?

UseInternetattheLibrary

Notawareofpublicinternetfacility .26(.03)Awareofpublicinternetfacility .39(.02)DifferenceNotAwaretoAware +.13Verydifficulttoaccesspublicinternetfacility .27(.03)Veryeasytoaccesspublicinternetfacility .39(.02)DifferenceVeryDifficulttoVeryEasy +.12

Note: Predicted probabilities calculated with Clarify Software from the logistic regression models reported inAppendixTables4.Probabilitiesestimatedwithcontrolvariablessetatmeanormodalvalues.Standarderrorsoftheprobabilityestimateinparentheses.

YoungerresidentsandAfrican‐Americansareamongthemostlikelytousepublicaccessatlibraries

The results reinforce conclusions that libraries serve residentswho seek themoutbasedonbothneedandconvenience.Youngerandbetter‐educatedresidentsarebothmorelikelytousetheinternet

atthe library (aswellastousethe internet ingeneral).Atthesametime,African‐Americans,Latinos,and lower‐income residents are also significantly more likely to go online at the library. There aresubstantial gapsbetweenAfrican‐Americansand Latinos in libraryuse,however, again indicating that

Latinoresidentsareamongthemostdisadvantagedintermsofinternetuse.Particularlyhighratesoflibrary internet use among African‐Americans support findings that many go online in some settingdespitelowerratesofhomeaccess.

• Ageisthemost importantfactoraffectingtechnologyuseatChicagopublic libraries. Younger

residents (age31)are30percentmore likely tousethe internetatapublic library thanolderresidents(age67),evenwhenwecontrolforinternetuse.YoungerChicagoresidentsaremoreengagedintechnologyingeneral,andaremoreinclinedtousetheinternetatpubliclibraries.

• Home internet use is the next most significant influence on technology use at the library.

Although it is clear that somepatrons lackhome internet connections, residentswhouse theinternet at home are 22 percent more likely to go online at the library. This supports thefindingsonconvenience,butalsoindicatesthatlibrariesprovideassistanceforthosewhohave

the internet at home – including the 46 percent of library users who need help findinginformationorthe25percentwhotaketechnologyclasses.

• African‐AmericansinChicagoare14percentmorelikelythanwhiteresidentstousetheinternetat apublic library, controlling forother factors. National surveyshave indicated thatAfrican‐

Americans have more positive attitudes toward use of public access,26 but this study showssubstantial differences in actual use (not just attitudes). Given that a higher percentage of

26Mossberger,TolbertandStansbury2003,chapter3.

28

internetuserswithouthomeaccessareAfrican‐American,publiclibrariesarefillingpartoftheneedforaccessamongtheseinternetusers.

• Incomeisnearlyasimportantasrace,forlow‐incomeresidentsare13percentmorelikelythan

higher‐income respondents to use the internet at a public library. Again, this indicates thatlibrariesarereachingneedypopulationsinthecity.

• But,educationhastheoppositeeffect. Internetuseat libraries increaseswitheducation,andcollege‐educatedresidentsare12percentmorelikelythanhighschoolgraduatestouselibrary

technologyfacilities.

• Awarenessofcommunitytechnologycenters (CTCs) intheirneighborhoodandperceivedeaseofuseforpublicaccessaccountfor12and13percentincreasesintechnologyuseat libraries,holdingotherfactorsconstant.Libraryinternetusersaremorelikelytobeawareofotherpublic

accessinthecommunity,andtoalsoperceivepublicaccessasconvenienttouse.

• While Latinos alsouse technology at public librariesmore thannon‐Hispanicwhites, they areonlyabouthalfaslikelyasAfrican‐Americanstobelibraryinternetusers.Latinosare8percentmorelikelythannon‐Hispanicwhitestousetheinternetatthelibrary.

NeighborhoodVariationinLibraryInternetUse

The neighborhood results provide further support for the dual nature of library internet use:

convenience for residentswhoare frequent librarypatronsor frequent internetusers,andassistanceforindividualswithlimitedaccessorskills.Therearesomewhatopposingpatternsforneighborhoods:

• ResidentsofneighborhoodswithhigherpercentagesofAfrican‐AmericanandLatino residentsare less likely to use public access at the library than residents in neighborhoodswith higher

percentagesofnon‐Hispanicwhites.

• Residents of neighborhoodswith high poverty rates are less likely to use the internet at thelibrary.

• Yet,Chicagoans living inneighborhoodswitha lowerpercentageofhigh schoolgraduatesare

alsomorelikelytouselibraries.So,somelow‐incomecommunitieshaverelativelymorelibraryuse.

Themaponthenextpageshowslibraryuseacrossthecommunityareas.Thereisonecommunityareawithveryhighuse (ArmourSquare, inblue). Theyellowareasareestimated tohaveat least35

percentofthepopulationusingthe internetattheChicagoPublicLibrary‐somewhathigherthanthecity‐wideaverageof33percent. Some,butnotall low‐incomecommunitieshavehigherusageoftheinternetatthelibrary.Theredareasareestimatedtohavelessthan35percentofresidentswhouse

theinternetatthelibrary,andarearoundthecity‐wideaverageorlower.

29

30

Librariesareclearlyfulfillinganimportantneedforsomeoftheless‐connectedinChicago,atthesametimethattheyappeal tothosewhoare frequentlyonline. Towhatextentarecommunitytechnology

centersbeingusedbyChicagoresidents,andwhoismostlikelytousepublicaccessthere?

CommunityTechnologyCenters

Communitytechnologycenters(CTCs)arealsoimportantforpublicaccessinChicago,as16percentofcityresidents(21percentofChicagointernetusers)reporthavingusedacommunitytechnologycenter.Surveyrespondentswereasked,“Asfarasyouknow,isthereaplacewhereyou

cangoinyourneighborhoodwheretheinternetispubliclyavailabletoanyonewhowantstouseit?SuchplacesareoftencalledCommunityTechnologyCenters.”Respondentswerethenaskedafollow‐upquestionaboutwhethertheyhadeverusedtheinternetatsuchaplace.

Becausecommunitytechnologycenterstargetlow‐incomeneighborhoods,wefocusedthe

regressionanalysisonlow‐incomecommunitiesonly(censustractswithpovertyratesabovethemean).Giventhepotentialimportanceoftheneighborhoodcontext,themainmodelforinternetuseatcommunitytechnologycenters(CTCs)isamultilevelmodelthatincludesneighborhoodcharacteristics.

Accordingtothemultilevelmodel,whichincludesbothindividual‐levelandneighborhood

characteristics(AppendixB):

• ParentsaremorelikelytouseCTCsinthislow‐incomesample,andthiscontrastswiththefindingsforlibraries.

• African‐AmericansaremorelikelytovisitaCTCthanwhiteresidents.

• Similartothelibraryfindings,communitytechnologycenterusersarealsoyounger,better‐educated,andthosewhoperceivetheCTCtobeconvenient.

• CTCuseincreasesasneighborhoodpovertyincreases,eveninthissampleofpoor

neighborhoods.

ThissuggeststhatCTCsarecertainlyreachingsomeofthepoorestresidents.Again,African‐Americansarefrequentusersofpublicaccess,butLatinosarenotsignificantlymorelikelythanwhitestouseCTCs

inlow‐incomecommunities.

Lookingatlow‐incomeneighborhoodsonly,CTCuseishigherinthepoorestcommunities;parentsarealsomorelikelytouseCTCs.

Therearesomecommonalitiesinuseofpublicaccessacrossthecity.WhilelibrariesandCTCsseemtobereachingdisadvantagedresidents,usersalsotendtobebetter‐educatedandyounger.

Therearetwoco‐existingpatterns–userswhoaremorelikelytobeinterestedintechnology,butalsothosewhoareless‐connected.Low‐incomeresidentsareservedbybothlibrariesandCTCs.Thisisparticularlyevidentincommunitytechnologycenters,whereuseincreaseswithneighborhoodpoverty.

31

WirelessAccess

Wirelesscanprovidemorefrequentaccessforresidentsandbusinessesthroughnetworksthataccommodatemobileusewithlaptopsandavarietyofhandheldwirelessdevices.Wirelessnetworks

canalsoprovideconnectionsforindividualswithouthomeaccess.Currently,thisisavailableinlibrariesandotherpublicplaces,aswellas“hotspots”thataresponsoredbysomebusinesses,suchascoffeehouses.Thereisthepotentialtoprovidebothmobileandhomeaccessthroughwirelessnetworksthat

coverresidentialareasofthecity.

Whatdoeswirelessuselooklikecurrently?Usingsimplepercentages,itisclearthatwirelessuseisfairlycommonnowandpromisestogrowinthefuture.Morethanone‐thirdofcityresidents(35percent)andnearlyhalfofChicagointernetusers(46percent)havegoneonlineusingwirelessaccessin

apublicplace.YoungChicagoresidentsaged18‐29arethemostlikelyusersofwireless,as55percentusepublicwirelessnetworks,and35percentdosoatleastafewtimespermonth.Whetheritisthroughlaptops,cellphones,orothermobiledevices,thereisafairamountofuseofwirelessnetworks

inthecity.Table3showssimplepercentagesforfrequencyofwirelessusebyage.

TABLE3.FREQUENCYOFWIRELESSUSEINPUBLICPLACEBYAGEPercentofcitypopulation

18‐29 30‐59 60‐74 75andoverTotal,allagesDaily/fewtimespermonth 35% 25% 9% 1% 21%Rarely 20% 15% 10% 2% 14%Totalpercentforagegroup 55% 40% 19% 3% 35%

Table4,below,displayspercentagesforwirelessusebyraceandethnicity,demonstratingthat

Asian‐Americansareaheadofothergroups.Nearlyone‐thirdofAfrican‐AmericansandLatinoshaveusedwirelessinChicago,althoughthisislessthanothergroups.

TABLE4.FREQUENCYOFWIRELESSUSEINPUBLICPLACEBYRACEANDETHNICITYPercentofcitypopulation WhiteNon‐Hispanic Black Asian Latino TotalDaily/fewtimesperweek 25% 18% 38% 19% 21%Rarely 15% 12% 16% 12% 14%Totalpercentforgroup 40% 30% 54% 31% 35%

Hand‐heldwirelessdevicessuchasinternet‐enabledcellphonesor“smartphones”alsoprovideawaytoaccesstheinternet,andtodaytheycanaccomplisharangeofusesdespitetheirsmallerscreens.Thisposesthequestionofwhethersomeresidentswhodonothavehomeinternetaccessare

goingonlinethroughcellphonesratherthanthroughpersonalcomputers.Insomenewly‐industrializingcountries,forexample,cellphoneuseisthemostcommonwaytogoonline.CouldthisbetrueintheU.S.,atleastamongsomewhowerenotpreviouslyinternetusers?Table5belowshowssimple

percentagesdescribinginternetaccessthroughcellphonesbyageinChicago.

32

TABLE5.FREQUENCYOFCELLPHONEUSETOCONNECTTOINTERNETBYAGEPercentofcitypopulation

18‐29 30‐59 60‐74 75andoverTotal,allagesDaily/fewtimespermonth 26% 14% 4% 1% 13%Rarely 13% 7% 3% 1% 7%Totalpercentforagegroup 39% 21% 7% 2% 20%

One‐fifthofChicagointernetusershaveusedacellphoneforinternetaccess.Thisisclearlymoreimportantforyoungerresidents,with26percentof18‐29year‐oldsreportingthattheygoonlinethiswayatleastafewtimespermonth(doubletheaverageof13percentforthecity).Personal

computersremainthedominantwaytoaccesstheinternetinChicago,buttherelativelyhighratesofuseamongtheyoungindicatepotentialforfuturegrowth.PatternsofcellphoneusebyraceandethnicityareindicatedinsimplepercentagesinTable6below,whichalsoshowsthepercentageof

residentswhousecellphonestoconnecttotheinternetbuthavenohomeinternetaccess.Thisprovidesawaytoassesstheextenttowhichcellphoneuseissubstitutingforhomeinternetuse,especiallyamongpopulationsthathavetraditionallybeenless‐connected.

TABLE6.FREQUENCYOFCELLPHONEUSETOCONNECTTOINTERNETBYRACEANDETHNICITYPercentofcitypopulation WhiteNon‐Hispanic Black Asian Latino TotalDaily/fewtimespermonth 12% 12% 20% 12% 13%Rarely 6% 8% 10% 7% 7%Totalpercentforgroup 18% 20% 30% 19% 20%Percentofpopulationwhousedaily/fewtimespermonthandhavenohomeinternet 1% 3% 0% 2% 2%

CellphoneusetogoonlineishigheramongAsian‐Americans,butthereisclearlynosubstitutionforhomeaccessinthisgroup.African‐AmericansandLatinosappeartousecellphonestoconnectto

theinternetasasubstituteforhomecomputersattimes.However,suchsmalldifferencesmaybeduetosampling.Overall,cellphonesstilldonotreplacehomeaccess,accordingtotheseresults.

Insummary,wirelessnetworkspresentanumberofopportunitiestosupplementhomeuseortoprovidelow‐costinternetconnectionsforcityresidents.Theythereforeencouragemorewidespread

use,morefrequentuse,flexibility,andinnovativeapplicationsinnewsettings.OverathirdofChicagoresidentshaveaccessedtheinternetthroughsometypeofwirelessdevice,andtheconcentrationofsuchuseamongresidentsunder30suggeststhatthistrendislikelytoincreaseinthefuture,especially

withadvancesintechnology.Freeandpublicwirelessaccesscanencouragefrequencyofuse,andmayextendaccessforsome,especiallyifitisavailableinmoreareasofthecity.Willthismakemuchdifference,however?Towhatextentisthereasonthatpeopleareofflineamatterofcost,forexample,

orasimplelackofinterest?Towhatextentareskillsandtechnicalsupportalsoneeded?Weexaminethebarriersthatresidentsthemselvesperceiveforachievingdigitalexcellence.

33

PARTIV.BARRIERSTOACCESSANDPUBLICPOLICY

Homeaccessisanimportantresourceforachievingdigitalexcellence.Weaskedthosewhodonotusetheinternetatallaswellasthosewhodonotuseitathometochooseanyandallreasonsfor

notusingtheinternetathome,andthenaskedthemtoselectthemostimportantreasonfornothavinganinternetconnectionathome.Inthisway,wecouldbetterunderstandwhetherrespondentswhosaidthattheycan’taffordtheinternetmightsimplybeuninterestedaswell,andthereforenotvery

motivatedtospendmoneyonacomputeroramonthlyinternetbill.

TABLE7.REASONSFORNOINTERNETATHOMEPercentofrespondentswhodonotusetheinternetathome Mainreason Onereason Don’tneedit/notinterested 30% 48% Costistoohigh 27% 52% Canuseitelsewhere 5% 52%Don’thavetime 5% 24%Toodifficulttouse 9% 43%Iamworriedaboutprivacy 2% 57%Theinternetisdangerous 2% 46%HardtouseinformationinEnglish 1% 19%Physicalimpairment 3% 13%Other 16% ‐‐

Whenrespondentsareallowedtogivemultipleanswers,issuessuchasprivacyanddangeremergeassecondaryreasonsformanyrespondents,eventhoughfewresidentscitethemasthemain

reasonfornothavingtheinternetathome.Difficultyisalsomoreimportantasasecondaryreason–peoplewhodonothavetheinternetathomemaynotchoosethisastheonlyreasonfornotinvestingintheinternet,buttheyarelessconfidentoftheirskills.Only5percentsaythatuseoutsidethehomeis

theirmainreasonfornothavinghomeaccess,butoverhalfoftherespondentscanusetheinternetsomewhereelse.Still,thereislittlestatisticalrelationshipbetweenthereasonsfornotusingtheinternetathomeinTable7below,evenwhenrespondentscouldchoosemultipleanswers.27Inother

words,ouranalysisshowsthatthosewhoarenotinterestedinhavingtheinternetathome,forexample,arenotthesamerespondentswhosaythatcostistheissue.

Table7showsthatinterest,affordability,andskillstandoutasthemostimportantmainreasonsfornothavingahomeconnection,andthattherearesomeinterestingpatternsbyraceand

ethnicitywhenweexaminesimplepercentagesforthemainreasonfornothavingtheinternetathome.Table8,below,showsthesepatternsbyraceandethnicity.

27Thiswasexploredthroughfactoranalysisandthroughcorrelations.

34

TABLE8.MAINREASONFORNOINTERNETATHOMEBYRACEANDETHNICITYPercentofrespondentswhodonotusetheinternetathome WhiteNon‐Hispanic Black Asian Latino TotalDon’tneedit/notinterested 42% 29% 42% 19% 31% Costistoohigh 14% 30% 12% 37% 27%Toodifficulttouse 9% 8% 9% 13% 9% Thereisconsiderablevariationbyraceandethnicityinthemainreasonfornothavingtheinternetathome.MorewhiteandAsian‐Americanresidentswhodonotcurrentlyusetheinternetathomearenotinterested,andAfrican‐AmericansandLatinosarebyfarmoreconcernedaboutcost.

Latinosarethegroupmostlikelytosaythatdifficultyusingtheinternetisthemainreasonfornothavingitathome.

Tosortoutdifferencesinreasonsfornothavinghomeaccess,weconductedmultivariateregression.WhatMattersTableDonthenextpagepresentstheresultsforlackofinterest,cost,and

difficultyusingthetechnology.AppendixAalsoshowstheresultsoftheregressionanalysisforseveralotherreasonsthatwerelesscommonasmainreasonsfornotusingtheinternetathome:useelsewhere;lackoftime;andprivacyconcerns.Latinosandhigher‐incomeresidentsarestatistically

morelikelytobeamongthosewhosaytheydonothavetime.Latinosandwomenaremorelikelytohaveprivacyconcerns.Thosewhousetheinternetelsewherearealsomoreeducatedandyoungerthanotherswithouthomeaccess.Becauseofspaceconsiderationswedonotanalyzethesehere,but

readerscanconsulttheresultsintheappendix.

WhatMattersTableD:WhataretheReasonsChicagoResidentsDoNotHaveHomeInternet?

Thefactorslistedbelowarethestatisticallysignificantinfluencesonthefollowingreasonsfornotusingtheinternetathome.Aplussign(+)indicatesincreasedprobabilityforgivingthisreason,andaminussign(‐)indicatesadecreasedchanceofcitingthisreason.NotInterested Age(+)Income(+)Education(‐)African‐American(‐)CostisTooHigh Income(‐)Latino(+)Female(+)Education(‐)DifficulttoUseAge(+)Education(‐)Latino(+)African‐American(‐)**borderlinesignificance

35

WhatMattersTableDcontinued:WhataretheReasonsChicagoResidentsDoNotHaveHome

Internet?

ReadingPredictedProbabilities:Afemale,whitenon‐HispanicChicagoresidentwithnochildrenandaverageage,

income,andeducationwhodoesnotusetheinternetathomehasa50percentprobabilityofsayingthatthisisbecausesheisnotinterested.Resultsshownbelowindicate,forexample,thatarespondentwhoisLatino,but

otherwisethesame,hasa48percentchanceofsayingsheisnotinterested.Latinoethnicityalonemakesa2percentdifference,holdingotherfactorsconstant,andisnotsignificant.Readthenumbersnotinparenthesesas

percentages.

NotInterested

CostisToo

High

TooDifficult

toUseWhitenon‐Hispanic(Baseline) .50(.04) .54(.04) .43(.04)Latino .48(.04) .69(.05) .57(.04)DifferenceLatinovs.White ‐.02 +.15 +.14Black .43(.04) .57(.03) .36(.03)DifferenceBlackvs.White ‐.07 +.03 ‐.07Male .55(.04) .39(.04) .37(.04)DifferenceFemalevs.Male ‐.05 +.15 +.06AnnualIncome VeryLow($0,‐2SD) .40(.05) .72(.04) .50(.05)Low($10,000‐$20,000,‐1SD) .47(.04) .59(.03) .45(.04)Mean/Average($40,000‐$50,000)

.50(.04) .54(.04) .43(.04)

High($75‐$100,000,+1SD) .62(.05) .29(.04) .35(.05)VeryHigh(morethan$150,000,+2SD)

.66(.05) .21(.04) .31(.05)

DifferenceLowtoHigh +.15 ‐.30 ‐.10EducationLevel LessthanHS .54(.04) .58(.04) .52(.04)HighSchoolGraduate .52(.04) .56(.04) .47(.04)SomeCollege .46(.04) .52(.04) .37(.04)CollegeGraduate .43(.04) .50(.04) .32(.04)GraduateDegree .40(.05) .47(.05) .28(.04)DifferenceHStoCollege ‐.09 ‐.06 ‐.15Ageofrespondent Veryyoung(18yrs,‐2SD) .24(.05) .50(.06) .15(.04)Young(31yrs,‐1SD) .32(.05) .51(.05) .22(.04)Mean/Average(49yrs) .50(.04) .54(.04) .43(.04)Old(,67yrs,+1SD) .56(.03) .55(.03) .52(.03)Veryold(85yrs,+2SD) .68(.03) .57(.03) .67(.03)DifferenceYoungtoOld(27­67yrs)

+.24 +.04 +.30

Note: Predicted probabilities calculated with Clarify Software from the logistic regression models reported inAppendixTable5.Probabilitiesestimatedwithcontrolvariablessetatmeanormodalvalues.Standarderrorsof

theprobabilityestimateinparentheses.

36

Respondentswereaskedwhytheydidnothaveinternetaccessathomeandcouldgivemultiplereasons.Themostfrequentlycitedanswersforthemainreasonwere“Idon’tneedit/notinterested,”

“thecostistoohigh,”“It’stoodifficulttouse.”Columns1,2and3ofWhatMattersCshowthepredictedprobabilityofcitingoneoftheaboveresponses,respectively,bydemographicattributesoftherespondents.

Theanalysisshowsthatamongthosewhodonothavetheinternetathome,olderandhigher

incomerespondentsareuninterested,andAfrican‐Americansaresignificantlylesslikelythanotherracialandethnicgroupstosaythattheyhavenointerestintheinternet.

Olderandmoreaffluentrespondentswithouthomeaccesscitelackofinterest

• Olderrespondentsare24percentmorelikelytocitealackofinterestasthereasontheyareofflinecomparedtoyoungrespondents;a31year‐old(onestandarddeviationbelowthemean)hasonlya32percentprobabilityofsayingheorsheisnotinterested,comparedtoanolder

individual(67years,onestandarddeviationabovethemean),whohasa56percentprobabilityofcitingthisreason.

• Higher‐incomeresidentsarealsomorelikelytosaythattheyareuninterested.Residentswithannualfamilyincomesbetween$75,000and$100,000are15percentmorelikelytocitelackof

interestthanrespondentswithincomesbetween$10,000and$20,000.

• Incomparison,educationmakesasmallerdifferencethanageandincome.Residentswithahighschooldiplomaare9percentmorelikelythancollegegraduatestosaytheyarenotinterestedintheinternet.

• African‐Americansare7percentlesslikelythanwhitestocitealackofinterestingoingonline.

NeighborhoodVariationinInterest

Multilevelmodels(seeAppendixB)showthatincomemattersatthecommunitylevelaswellasat

theindividuallevel:

• Residentsofmoreaffluentneighborhoodswithouthomeinternetaccessaremorelikelytosaythattheyarenotinterestedingoingonline.

Themaponthenextpageshowsclearlythispatternforincome.Communityareasinblueareestimatedtohave50percentormoreofresidentswithouthomeaccesswholackinterestinthe

internet.Communityareasinredareestimatedtohavebetween25and35percentofthosewithouthomeaccesswhogivethisreason.Redareastendtobeamonglow‐incomeareasinChicago.

37

38

Low‐incomerespondentsandLatinosareamongthosewhocitecostasthemainfactor,asseenincolumn2.

Residentscitingcostareinfactlow‐income;Latinosareamongthemostlikelytoviewcostasabarrier

• Thelargestfactorinfluencingthosewhosaythatcostistoohighis,notsurprisingly,familyincome.Thepoor(withincomesbetween$10,000‐$20,000onestandarddeviationbelowthemean)are30percentmorelikelytoperceivecostasabarriertohomeaccessthantheaffluent

(incomesbetween$75,000‐$100,000,plusonestandarddeviationabovethemean),allelseequal.PoorChicagoresidentshavea60percentprobabilityofcitingcostbarriers,comparedtohigher‐incomeresidents,whohavelessthana30percentchanceofsayingthis.

• Holdingarespondent’sincome,educationandageconstant,Latinoswere15percentmorelikelytosaycostisaproblemforinternetaccessthannon‐Hispanics.

• African‐Americans,incontrast,wereonly3percentmorelikelythanwhitestosaycostisanissueforhomeaccess,controllingforotherfactors.

• Interestingly,womenwere15percentmorelikelythanmentomentioncostasareasonfornot

havinghomeaccess,allelseequal.

NeighborhoodVariationinCostConcerns

Accordingtothemultilevelmodels(AppendixB),thereisvariationacrossChicagoneighborhoodsincostasabarriertohomeaccess:

• ResidentsofcommunitieswithhighAfrican‐Americanpopulationsaremorelikelytostatethatcostisthemainreasonfornothavingtheinternetathome.

• ResidentsinneighborhoodswithhighproportionsofLatinosarealsomorelikelytocitecost.

• Costconcernsaremorelikelyinneighborhoodswithahigherlevelofhighschoolgraduationas

well.Moreeducatedenvironmentsmayincreaseinterest,absentworriesaboutcost.

Themaponthenextpageshowscommunityareasmarkedinredwhere39percentormoreofthepopulationwithouthomeinternetconnectionscitecostbarriers.Costconcernsarefairlyimportantoverall,buttheredareasclearlycovermanyneighborhoodswithhighproportionsofLatinosand

African‐Americans.

CostismorelikelythemainreasonfornothavinginternetathomeinneighborhoodswithhighpercentagesofAfrican‐AmericansandLatinos

39

40

ThelastcolumnofWhatMattersTableDshowsthatless‐educated,olderandLatinorespondentsaremorelikelytosaythattheyhavedifficultywiththeinternet.

Olderandless‐educatedresidentsfindtheinternetdifficult,asdoLatinos

• Olderrespondents(onestandarddeviationabovethemean)were30percentmorelikelytociteskillbarrierscomparedtotheyoung(onestandarddeviationbelowthemean).

• Respondentswithonlyahighschooldegreewere15percentmorelikelytosaytheinternetis

“toodifficulttouse”comparedtothosewithacollegedegree.

• Latinosare14percentmorelikelytocitealackofskillsordifficultygoingonlineasabarriertousethanwhitenon‐Hispanics,againindicatinggreaterdisparitiesforLatinos.

• Incontrast,African‐Americansare7percentlesslikelytociteskillsasabarriertousecomparedtowhiteswhodonothavehomeaccess.Thismayreflectinternetuseoutsidethehomeamong

African‐Americans.

NeighborhoodVariationinDifficultyUsingtheInternet

Whenneighborhoodcharacteristicsareintroducedinmultilevelmodels,therearetwoapparentlycontradictoryfindings:

• Individualsresidinginhigher‐povertycensustractsarelesslikelytocitealackofskillsasareasonfornothavinghomeaccess,controllingforotherfactors.Thismayreflecttheinfluence

ofotherreasons,suchascost.

• Yet,residentsinneighborhoodswithahighpercentageofAfrican‐Americansaremorelikelytomentiondifficultyinuse(althoughattheindividuallevelAfrican‐Americansarenot).

Thismaysuggestsomeskilldeficitsconcentratedintheseareasnotcapturedbytheotherfactorsexaminedhere.Residentsofsuchareasmayhaveexperiencedunequalqualityofeducation.

Themaponthenextpageshowsdiversepatternsaswell.Inthiscase,thecommunityareas

coloredinredareestimatedtohavehigherpercentagesofresidentswithouthomeaccesswhofindinternetusedifficult(between30and45percent).Itisclearthatmanylargely‐African‐Americancommunityareasinthesouthofthecityareonthislist,butothersarealsocoloredinblue,meaning

thattheyhavethelowestratesofresidentswithoutaccesswhohavedifficultyonline(between10and20percent).Themapsarebasedonmultilevelmodelsthatcombineneighborhoodandindividualcharacteristics,andfactorssuchasageorLatinoethnicityofrespondentsarereflectedintheresultsas

well.

41

42

PolicyImplications

Summarizingtheaboveresults,wecanseethattherearesomedistinctdifferencesinreasonsfornothavingtheinternetathome.Thosewhosaytheycan’taffordtheinternetareindeedlower‐

incomeincomparisontootherrespondentswithouthomeaccess,andsomeresidentswhohavedifficultywiththeinternetareinfactless‐educated,controllingforotherfactors.Theseindividualsaredifferentfromthosewhoaresimplyuninterested,andvariedpolicysolutionsareneededtoaddress

dissimilarbarriers.Low‐costinternetconnectionsandhardwarewillhelptobringmorelow‐incomeresidentsonline,buttrainingandsupportareneededforthosewhoareless‐educatedandlessconfidentoftheirskills.

Residentswhoareuninterestedintechnologyaredifferentfromotherswithouthomeaccess

becausetheyareolderandhavehigherincomes.Olderrespondentsaremorelikelytosaythattheyarenotinterestedorthatthetechnologyistoodifficulttouse.Thosewhohavehigherincomes(withinthisgroupwithouthomeaccess)alsolackinterestorsaythattheyhavenotime.28Greaterawarenessof

whatcanbedoneonlineandtheprovisionofsupportcouldchangethemindsofsome.

ThereareclearopportunitiesforexpandinghomeaccessamongAfrican‐Americans,whohavefewernegativeattitudestowardtechnologythanwhites,consistentwithearlierresearchandwithuseoftechnologyoutsidethehome.29Theyarelesslikelytosaythattheyarenotinterested,orthatthe

internetistoodifficulttouse.Theserelativelypositiveattitudestowardtheinternetmightbetranslatedintogreaterhomeaccessifitisaffordable.WhileAfrican‐Americansarenomorelikelythanwhitestocitecostswhenwecontrolforfactorslikeincomeandeducation,thesimplepercentagesshowa

tendencyforAfrican‐Americanstobeamongthelower‐incomeresidentswhoareconcernedwithaffordability.ThemultilevelmodelsalsoshowthatresidentsofpredominantlyAfrican‐Americanand

Latinoneighborhoodsaremorelikelytoseecostasanissue.

PositiveattitudesamongAfrican‐Americanspresentanopportunity;Latinos,however,perceivemanybarriers

Latinosstandoutasperceivingmanybarrierstohomeinternetaccess:costanddifficultywereanalyzedhere,butresultsinAppendixAshowthatLatinosarealsosignificantlymorelikelythannon‐

Hispanicwhitestocitelackoftimeandconcernsaboutprivacy.Latinosarealsoprevalentinthe19percentofrespondentswithouthomeaccesswhomentionlanguagebarriersonline.Affordability,technicalsupportandtrainingareallneededtoaddressdisparitiesforLatinos.Recentimmigrants,in

particular,arelikelytohavealackofexperiencewiththeinternetaswellaslanguagebarriers.

28Theonlymodestcorrelationthatwefoundbetweenanswersinthemultipleresponsesectionwasbetweenlackofinterestandlackoftime.29SeeMossberger,TolbertandStansbury2003.AnationalsurveyshowedthatAfrican‐Americanshadsignificantlymorepositiveattitudestowardpublicaccess,technologytraining,onlineeducation,anduseoftheinternetforeconomicopportunity.

43

Anotherwayofthinkingaboutpossiblemotivationstogoonlineistoexaminepatternsofinternetuseamongthosewhoareonline.Aretheresomeactivitiesontheinternetthataremore

frequentlyengagedinbylow‐incomeresidentsorbyminorities,forexample?Couldthistellussomethingaboutpossiblemotivationsforthosewhoarenotonline,butmayhavesimilarinformationneedsorpreferences?

PARTV.ONLINEACTIVITIESTHATINFLUENCEOPPORTUNITIESFORRESIDENTS

OnceChicagoresidentshavebridgedthedigitalaccessdivide,therangeofactivitiesonlineis

almostinfinite.Thissectionfocusesonactivitiesthatpublicpolicymayhavesomeinterestinpromoting,becausetheyhavethepotentialtoenhanceoutcomessuchaseconomicadvancement,civicparticipation,accesstogovernmentservices,orhealthcare.Table8belowshowsthesimple

percentagesofChicagoresidentsandinternetusersengaginginthefollowingactivitiesonline.

TABLE9.ACTIVITIESONLINEHaveeverusedtheinternetto... CityPopulation InternetUsersReadonlinenews 67% 91%Findhealthinformation 64% 86%Findinformationongovernment 57% 76%Getinformationaboutpublictransport 56% 74%Getinformationonpolitics 53% 71%Lookforjobinformation 50% 67%UseCityofChicagowebsite 49% 65%Doworkforyourjob 48% 64%Takeaclassortraining 31% 41%

Internetusersperformmanytasksonline;thewebisreplacingotherwaysoffindinginformationorconductingtransactionsinChicago

Thereisamigrationtotheinternetofactivitiesthatcanbedoneoffline(suchasreadingthe

news,contactinggovernment,orfindinghealthinformation)becauseoftheconvenienceandinformationcapacityonline.Whiletwo‐thirdsormoreofinternetusershaveengagedinmostoftheaboveactivities,cityresidentswhoarenotonlineareexcludedfromthebenefitsoftheinternetforeconomicopportunityandforinformation.Thebalanceofthissectionfocusesonpatternsofuseforthecitypopulation,usingmultilevelmodelsthatincludeneighborhoodcharacteristics.EmploymentandTraining

Internetandcomputerskillsareincreasinglyrequiredforjobsthroughoutthelabormarket,and

thedemandisnotlimitedtothetechnologyindustryortoprofessionaloccupations.Whiletherearecertainlysomelow‐skillpositionsthatdonotrequireinternetuse,occupationsdemandinginternetusepaymore,evenforless‐skilledworkerswhohaveahighschooleducationorless.Onestudyconcluded

44

thatin2003,anaverageworkerwhousedtheinternetonthejobearned$118perweekmoreforinternetuse,controllingforotherfactors,includingeducationandoccupation.Wage‐earnerswitha

highschooleducationorlessgainednearlyasmuchfrominternetuseonthejob‐$111perweek.African‐AmericanandLatinoworkerswithahigh‐schooleducationorlessreceivedaslightlylargerpercentagewageincreasefrominternetusethanwhiteworkers,helpingtonarrowracialandethnic

wagegaps.30Howrelevantisinternetuseforless‐educatedworkersinChicago?Table10belowshowsinternetuseonthejobinChicagobyeducationalattainment,includingsimplepercentagesforfrequencyofuse.ThefiguresareforemployedChicagoresidentsonly,ratherthanallresidents.

TABLE10.FREQUENCYOFINTERNETUSEFORJOBBYEDUCATIONEmployedChicagoresidentsonlyEducationalAttainment %ofEmployedWhoUsetheInternetforWork DailyoraFewTimesPerWeek0‐8Years 9%9‐11Years 13%HighSchoolGraduate 33%Vocational/TechnicalEducation 35%SomeCollege 54%4‐YearCollegeDegree 74%Post‐GraduateStudy 88%TotalEmployed 63%

33percentofemployedChicagoresidentswithahighschooleducationusetheinternetforworkdailyorseveraltimesperweek

Chicagoresidentswhousetheinternetatworkaremostprevalentinhigh‐skilloccupationsheld

byworkerswithabaccalaureateorpost‐graduatedegree.Thereisasteepincreaseininternetuseatworkbyeducationalattainment.Amajorityoftherespondentswithsomepost‐secondaryeducationusetheinternetatleastoccasionallyonthejob.But,itisalsosignificantthat33percentofemployed

respondentswhohaveahighschooleducationusetheinternetforworkonaregularbasis.Thesefiguresshowthatinternetuseisfairlycommoneveninlower‐skilledoccupationsthatdonotrequirecollegedegrees.Internetusethroughoutavarietyofindustriesispredictedtogrownationallyovera

numberofyears,31anddigitalskillswillbeincreasinglyimportantfortheeconomicprospectsofeven

30Mossberger,TolbertandMcNeal2008,41.Controllingforotherfactors,African‐Americanmenearnanaverage18.36%wage“premium”forinternetuseatwork,African‐Americanwomenearn17.31%more,Latinosearn16.99%more,andLatinasearna16.11%increase,whilewhitemenearn14.77%more,andwhitewomengain13.56%.ThebenefitsforinternetuseatworkmeanalargerincreaseforAfrican‐AmericansandLatinosrelativetowhiteworkers,becausetheytendtohavelowerwages.31LitanandRivlin2002(BrookingsInstitution)

45

less‐educatedworkers,andforthecity’sabilitytoattractorcultivateinnovativefirmsseekingtechnology‐skilledemployees.

Howdoesinternetuseatworkvarybyage,andbyraceandethnicity?Multilevelmodels(Appendix

B)wereusedtopredictinternetuseatworkforthosewhoareemployed(ratherthanthewholecitypopulation).

• ChicagoworkerswhoarelesslikelytousetheinternetatworkareLatino,African‐American,low‐income,andless‐educated.

• Residenceinaneighborhoodwithloweducationalattainmentdiscouragesinternetuseatwork.

Internetuseatworktracksdisparitiesininternetusegenerally

Internetuseonthejobcanincreasewages,butinternetuseforjobsearchandtrainingmaycontributetoeconomicadvancementaswell.Nationalresearchshowsthatdespitehavinglowerrates

ofinternetuse,African‐Americansaremorelikelythanothergroupstosearchforjobsonline.32ThisistrueinChicagoaswell.

Multilevelmodelsforthe(AppendixB)demonstratethat:

• Chicagoresidentswhoaremorelikelytosearchforjobsonlineareyounger,African‐American,andresidentswithhigherincomesandhighereducation.

• Latinosarelesslikelytosearchforjobsonlinethanwhitenon‐Hispanics.

• Neighborhoodfactorsarenotsignificant.

Thosewhousetheinternetforjobsearcharegenerallyresidentsmostlikelytousetheinternet–

withtheexceptionofAfrican‐Americans.PriorresearchhasshownthatAfrican‐Americans,inparticular,associateinternetusewitheconomicopportunity.Highuseofonlinejobsearchmaybeperceivedasastrategytocounterdiscriminationinthejobmarket.33

African‐Americansaremorelikelytosearchforjobsonline;Latinosarelesslikely

Distanceeducation,onlinetrainingprovidedbyemployers,andotheronlinecoursesprovidenewpossibilitiesforeconomicadvancementfordisadvantagedworkers.Multilevelmodelsforonline

education(AppendixB)demonstratethat:

• Residentswhoaremorelikelytotakeclassesortrainingovertheinternetareyounger,higher‐incomeandmoreeducatedChicagoresidents.

32PewInternetandAmericanLifeProject,May2008,internettrendsovertimeatpewinternet.org;Mossberger,TolbertandStansbury2003.33Mossberger,TolbertandStansbury2003.

46

• Therearenostatisticaldifferencesbasedonrace,ethnicityorgender.

• Neighborhoodcharacteristicsarenotsignificant.

ClearlytheaspirationtosucceedeconomicallypresentsanopportunitytopromotedigitalexcellenceamongmanyChicagoresidents.Neighborhoodfactorsaresignificantforinternetuseforthe

job,butdonotaffectonlinejobsearchortraining.Researchonlow‐incomecommunitieshasconcludedthatresidentsareoftenisolatedfrombetter‐payingjobsbecausetheylacksufficientinformationaboutsuchopportunitiesintheirinformalinformationnetworks.34Theinternetcanpossiblyextendthose

networks.Internetuseismoreprevalentin,butnotconfinedtojobsrequiringthehighestlevelsofeducation,andjobsrequiringtheinternetofferbettercompensationevenforless‐educatedworkers.Assistancewithjobsearch,onlinetrainingoreducation,anddigitalskillsfortheworkplacehave

particularrelevanceinlow‐incomeandminoritycommunities,andshouldbeanimportantpartofpublicandcommunity‐basedprogramsfordigitalexcellence.

NewsandPoliticsOnline

Surveyrespondentsfrequentlyreportedreadingonlinenews,lookingatpoliticalinformationonline,andusinggovernmentwebsitesforinformationandservices.Followingonlinenewsisthemost

commonactivityincludedinthesurvey,as91percentofChicagointernetusershaveeverreadthenewsonline,and74percentpursuethenewsonlineatleastafewtimesperweek.

Nationalresearchhasdemonstratedthatuseofonlinenewsisrelatedtohigherlevelsofcivicengagement–politicalknowledge,interest,anddiscussion.35Anumberofstudieshavealsoestablished

apositivelinkbetweenuseofonlinenewsandvoting,36andthisrelationshipissignificanteventakingintoaccounttheuseofnewspapersandtelevisionfornews.Whilethosewhofollowthenews(fromanysource)alsotendtohavehigherlevelsofcivicengagementandvoting,theresearchindicatesthat

onlinenewshasincreasedbenefits–perhapsbecauseofitsconvenientavailability,in‐depthcoverage,multi‐mediacapacity,orthevarietyofsourcesthatcanbeaccessed.

Internetusersalsohavemanyoptionsforfindinginformationaboutpoliticsapartfromonline

news.Thesourceshaveproliferatedinrecentyears,withpoliticalblogs,campaignwebsites,interestgroupwebsites,YouTubevideosandFacebookentriesallcontributingtotheflowofpoliticalinformation,especiallyaroundelectiontime.Alittleoverhalfofthecity’sresidentshavegoneonline

forpoliticalinformationaccordingtoourJuneandJuly2008survey.

Multilevelmodels,includingneighborhoodcharacteristics,wereestimatedforuseofpoliticalinformationonlineforthecitypopulation(seeAppendixB).Theyrevealthat:

34TheclassicstudyisGranovetter(1973).35SeeTolbertandMcNeal2003,andchapter3ofMossberger,TolbertandMcNeal(2008)foradescriptionoftheuseoftwo‐stagemodelsinthisresearch,aswellasmoredetailsontheresults.36Bimber2003;TolbertandMcNeal2003,amongothers.

47

• Chicagoresidentsmostlikelytobeinterestedinpoliticsonlineareyounger,whitenon‐Hispanic,higher‐income,better‐educated,andmale.Thisisconsistentwithpublishedresearch.37

• Parents,however,arelesslikelythanthosewithoutchildrentolookforpoliticalinformationon

theinternet.Thismaysuggesttimeconstraintsforeitherpoliticsorinternetuse.

• ResidentsofneighborhoodswithhighpercentagesofLatinosreportmoreinternetuseforpolitics.Thisisanintriguingfindingthatcouldmeritfurtherinvestigation.

• Residentsinneighborhoodswithhigherpercentagesofhighschoolgraduatesaremorelikelytoparticipateinpoliticsonline.Education(attheindividuallevel)isoneofthestrongestpredictors

ofpoliticalparticipationmoregenerally,andfindingsoneducationintheneighborhoodcontextunderscorethispoint.

Politicsonlineengagestheyoung,whoareotherwiselesslikelytobeinvolved

Theinternetischangingpoliticsbecauseofitsattractionfortheyoung.Inotherways,however,itcontinuesmoretraditionaldivisionsinpoliticalparticipation,especiallythosebasedonincomeandeducation.Thefindingsforeducationalattainmentinneighborhoodsreinforcethispattern.Livingina

Latinoneighborhoodisalsoassociatedwithhigheruseoftheinternettofolloworengageinpolitics.ThisrunscountertomoregeneralpatternsoflowerinternetuseinLatinoneighborhoods.Otherwise,disparitiesininternetusebasedonrace,ethnicity,educationandincomethreatentowidengapsin

politicalparticipation.

DigitalGovernment

Governmentwebsitesareanothersourceofinformationonpoliticsandpublicpolicy,buttheyalsocontainvaluableinformationaboutservicesandonlineservicetransactions.E‐governmentusershavegenerallypositiveattitudestowardtheironlineexperiences,includingfeelingsthatgovernmentis

moreresponsive,moreeffective,andefficient.38Governmentonlineincreasestheaccessibilityofgovernmentservices.Paradoxically,low‐incomeresidentsdependmostonmasstransitandotherpublicservices,yetareamongthosewhoareleastlikelytobeonlineandtobenefitfromthe

convenienceandaccessprovidedbye‐government.Nationalstudiesindicatethate‐governmentusersaremorelikelytobeyoung,higher‐income,educatedandmale.39Somenationalsurveyshaveshownthatlocalgovernmentmaybedifferent;higherpercentagesofAfrican‐Americansandwomenuselocal

governmentwebsites.40Itisunclear,however,whetherAfrican‐Americansandwomenaremorelikelytouselocalwebsitescontrollingforfactorssuchasincomeandeducation.

Thesurveycontainedquestionsaboutuseofgovernmentwebsites(foranylevelofgovernment),theCityofChicagowebsite,andpublictransitwebsitesfortheChicagoTransitAuthority

37Krueger2002;Mossberger,TolbertandStansbury2008amongothers.38West2003;Welch,HinnantandMoon2006;TolbertandMossberger2006.39West2005;Mossberger,TolbertandStansbury2003.40LarsenandRainie2002

48

(CTA)orRegionalTransitAuthority(RTA).Multilevelmodelsthatincludeneighborhoodcharacteristicswereestimatedforallthreetypesofgovernmentwebsites(seeAppendixB).Themodelsareforthe

citypopulationasawhole(ratherthaninternetusersonly).

Governmentwebsites(anylevelofgovernment).Useofe‐governmentingeneralbearssomesimilaritiestouseofonlinepoliticalinformationinChicago,andresultsforgenerale‐governmentuseinChicagofitthenationalpatterns,involvingfrequentinternetusers.

• Younger,whitenon‐Hispanic,higher‐incomeandbetter‐educatedresidentsaremorelikelyto

visitgovernmentwebsites.

• ResidentsofneighborhoodswithhigherpercentagesofAsian‐AmericansandAfrican‐Americansarealsomoreattunedtodigitalgovernment.Thismaysuggestsomethingabouttheneighborhoodsthatisnototherwisecapturedintheanalysis.

CityofChicagowebsite.NearlyhalfofChicagoresidents(49percent)haveusedthecity’swebsite–

slightlylessthanthe57percentwhohaveusedanygovernmentwebsite.Resultsforthelocalwebsiteconfirmsomeofthenationalpatternsforlocale‐governmentuseapparentinearlierstudiesthatdidnotusestatisticalanalysis.

• Parentsandfemaleresidentsaremorelikelytousethecity’swebsite.

• Younger,moreaffluent,andmoreeducatedresidentsaresignificantlymorelikelytousethe

city’swebsite.

• Therearenostatisticaldifferencesbyraceorethnicity.

Localgovernmentwebsitesmaybeparticularlyrelevantforthedailyroutinesofresidents,attractingparentsandwomen.Whiletherearenoracialorethnicdifferencesinlocale‐governmentuseoncewecontrolforotherfactors,thiscontrastswiththefindingsforalllevelsofgovernment.In

general,usersofthecity’swebsitearemorediversethangenerale‐governmentusers.

Themaponthenextpageshowscommunityareasinbluewhereuseofthecity’swebsiteisestimatedtobe50percentormore(abovethecity’s49percentaverage).Neighborhoodfactorsaren’t

significantpredictorsofcitywebsiteuse,althoughitisclearthatareaswithhigher‐incomeindividualsaccountforsomeofthehigh‐useareasinblue.Still,somelowerincomeareasareshadedinblueandmanylow‐incomeareasareincludedinthecommunitiescoloredinyellow,wherebetween35and49

percentofresidentsareestimatedtousethecity’swebsite.

UseofChicago’swebsiteismoreinclusivethane‐governmentuseingeneral;womenandparentsaremorelikelytouseit,andtherearenodifferencesby

raceandethnicity

49

50

Masstransituse.PublictransitplaysanimportantroleinChicago.TheChicagoTransitAuthority(CTA)andtheRegionalTransitAuthority(RTA)haveonlinetripplanners,schedules,andonlinetransactions

forfarecards.Seventy‐fourpercentofinternetusers(56percentofcityresidents)haveusedpublictransitwebsitesforinformationortransactions,slightlyhigherthanthepercentageusingthecity’swebsite.

Multilevelmodels(AppendixB)indicatethatChicagotransitwebsiteusersaremostlythosewho

areonlinefrequently,exceptthatneighborhoodpovertyplaysaroleaswell.

• Young,whitenon‐Hispanic,higher‐incomeandbetter‐educatedChicagoansareamongthosewholookuptransitinformationonline.

• Additionally,residentsofneighborhoodswithhighpovertyratesarealsosignificantlymorelikelytousetransitwebsites.Needmattersaswellasinternetuse.

• Controllingforotherfactors(suchasneighborhoodpoverty),African‐AmericanandLatino

neighborhoodsaresomewhatlesslikelytousemasstransitinformationonline.Thisindicatesthatnotallpoorneighborhoodsareequallylikelytohaveresidentswhousetheinternetfortransitinformation.

Residentsofpoorneighborhoodsareamongthemostlikelytousepublictransitwebsites

Thefindingsforthecityandtransitwebsitessuggestthattheneedforgreateraccesstolocalservicesmayalsobeamotivatingfactorforresidentstogoonline,particularlyincommunitieswhere

thereisrelianceonmasstransitandotherpublicservices.Atthesametime,somepoorcommunitiesarelessconnectedtoonlinemasstransitinformation.Forcityservicesmoregenerally,thereismorediversityofuse.

HealthCare

Amongtheonlineactivitiesincludedinthesurvey,lookingforhealthinformationisoneofthe

mostcommon,with64percentofthepopulation(86percentofinternetusers)whohavedonethisatsometime.Healthinformationcanbechallengingtounderstand,assomewebsitesareorientedtopractitioners,andothershavequestionablecredentials.Informationliteracyisparticularlycriticalinthis

area,forinternetusersneedtomakejudgmentsaboutthecredibilityofsourcesandtopayattentiontohowrecentlyinformationhasbeenpostedorupdated.

Multilevelmodelsforuseofonlinehealthinformation(AppendixB)showsomeinterestingpatterns:

• Younger,moreaffluent,andmoreeducatedresidentsaremorelikelytoturntotheinternetfor

healthinformation.

51

• Womenandparentsarealsomorelikelytousetheinternetforthispurpose.Thisfitswithpreviousresearchdemonstratingthatwomenandcaretakersarethemostfrequentusersof

healthinformationontheweb.41

• Latinosaresignificantlylesslikelythannon‐Hispanicwhitestofindhealthinformationonline.

• But,therearenosignificantdifferencesbetweenAfrican‐Americansandwhites,controllingforotherfactorssuchasincome,education,andneighborhood.

• RespondentswholiveinneighborhoodswithahighpercentageofAfrican‐AmericansandLatinosarelesslikelytoresearchhealthonline,indicatingsomespatialpatternstohealth

disparitiesonline.

Womenandparents,inparticular,valueonlinehealthinformation;therearenodifferencesbetweenAfrican‐Americansandwhites

SummaryonInternetActivities

Lookingacrossthesemanyactivities,theimpactofdisparitiesininternetusearevisible,

especiallyforLatinos,low‐income,andless‐educatedresidents.Low‐incomeandminorityneighborhoodsaccountforsomedisparitiesaswell.Buttherearealsoindicationsthatsomeusesoftheinternetmayprovideaparticularmotivationtogoonlineinlow‐incomecommunities.Theinternetcan

beanequalizingforceinprovidingaccesstoinformationandservices.African‐Americansuseonlinejobinformationtoagreaterextentthanwhites.Youngresidentsareamongthemostfrequentusersofpoliticsandnewsonlineaswellase‐government,eventhoughtraditionallytheyaremostapathetic

aboutpoliticsandcivicaffairs.Residentsofpoorcommunitiesusepublictransitwebsitesmore.Locale‐governmentattractsmorewomenandparentsthanothergovernmentsites,andforlocalgovernmenttherearenorealdifferencesbasedonraceorethnicity.African‐Americansarejustaslikelyaswhitesto

lookforhealthinformationontheweb.Embeddedinthesefindingsaresomeindicationsofhowtoengagemoreresidentswhoarenowunconnectedorless‐connected.

VI.CONCLUSION:CHALLENGESANDOPPORTUNITIESFORDIGITALEXCELLENCEINCHICAGO

Although75percentofChicagoresidentshavesomeexperiencewiththeinternet,almost40percentareeitherofflinecompletelyorhavelimitedaccess.Regularandeffectiveuseisbuiltona

foundationofhomeaccessandhigh‐speedconnections,whichfostermorefrequentuse,knowledgeofactivitiesonline,anddigitalskill.Aretheresomesolutionsthatmightbeemployed?Istherepublicsupportforaddressingtheseissues?

PublicOpinionasanOpportunity

Chicagoresidentsfavorpoliciestoclosethesegapsthroughgreateravailabilityofinternet

access.Oneofthequestionsincludedonthesurveyintroducedthetopicofwirelessnetworksand

41Fox2005

52

askedresidentstochooseamongseveraloptions.Respondentsweretold:“There'sbeentalkaboutbuildingawirelessnetworkinneighborhoodsinChicago.Whichofthefollowingshouldbethefocusin

doingthisproject?”Table11comparestheresponsesforthecitypopulation,lower‐incomeneighborhoods,andhigher‐incomeneighborhoods.Lower‐incomeneighborhoodshadmedianincomeslowerthanthecity’smean,andhigher‐incomeneighborhoodswereabovethemean.

TABLE11.WhereShouldaWirelessAvailabilityProjectStart? Choices Respondents CityPopulationLower‐incomeHigher‐income NeighborhoodsNeighborhoodsAlloverthecity 50% 50% 47%Inlow‐incomeneighborhoods 13% 15% 9%Inpublicschools,libraries,publicplaces 26% 24% 29%Shouldn’tworkonthisproject 7% 5% 11%Don’tknow 3% 3% 3% Forthecityasawhole,89percentsupportedsometypeofinitiative.Themostpopularalternativewastoprovidewirelessacrossthecity–supportedbyhalfofChicagoresidents.Approximatelyone‐quarteroftherespondentschosewirelessinpublicplaces,andabouthalfasmany(13percent)favoredprovidingwirelessfirstinlow‐incomeneighborhoods.Therearesmalldifferencesbyneighborhood,withresidentsoflower‐incomeareasslightlymorelikelytochooselow‐incomeareasfirstandslightlylesslikelytosupportwirelessinpublicplacesfirst.Residentsofhigher‐incomeneighborhoodsareabitmorecriticaloftheideaoverall,althoughthedifferencesaremodest–atmost5‐6percentagepointsdifferentfromlow‐incomeareas. Residentswerealsoaskedwhethertheywouldsupportawirelessprojectifitinvolvedasmalltaxorfeeincrease.Asexpected,supportforwirelessdropped,butthereisstillmajoritybackingfortheidea–61percentforthecityasawhole,60percentforlow‐incomeneighborhoods,and56percentforhigher‐incomeneighborhoods.Overall,Chicagoanshaveapositiveviewofwirelessprograms.Thismayindicatemoregeneralsupportfortechnologyinitiativesinthefuture.TABLE12.WouldYouSupportaWirelessAvailabilityProjectforaSmallTaxorFeeIncrease? Respondents CityPopulation Lower‐income Higher‐income Neighborhoods NeighborhoodsYes 61% 60% 56%No 32% 28% 36%Don’tknow 7% 7% 6%Refused/missing ‐‐ 3% 1%

ChallengesandOpportunitiesforAddressingDisparities

Atthesametimethatthisreportprovidesastraightforwardassessmentofexistinginequalities,therearereasonstobeoptimisticaboutthefutureifcurrentchallengesaremet.Digitalgapsarepatternedalongfamiliarlines–age,income,education,raceandethnicity.Latinosstandoutasthe

groupinChicagothatisleast‐connectedtotheinternet,especiallyamongthosewhopredominantly

53

speakSpanish.African‐Americansarestrivingtowardequityonline,asraceaccountsforarelativelysmallgapininternetuseanywhere,butalargergapinhomeaccess.Inturn,African‐Americanshave

higheruseofpublicaccess,andAfrican‐Americanswholackhomeaccesshavemorepositiveattitudestowardtechnologythansimilarly‐situatedwhites.Latinos,whoperceivemorebarrierstotechnologyusethanothergroups,aremuchmorelikelytocitecostratherthanalackofinterestfornotusingthe

internetathome.Thereisanopportunitytoreachdisconnectedorless‐connectedAfrican‐AmericanandLatinoresidentswithaffordableaccessandappropriatetrainingandsupport.Thesignificanceofincome–especiallyforhomeaccessandforbroadbandaswell–demonstratesthatitispoorresidents

whoareoftenexcludedfromthebenefitsoftheinternet.Ourfindingshighlightdifferencesinattitudesandneedsacrosslow‐incomegroups,andcommunity‐basedeffortsarelikelytobemostsuccessfulinaddressingthesevariedneedsforoutreachandassistance.

Ageaccountsforthelargestdisparitiesininternetuse,andisoneofthemostimportantfactors

inhomeaccessandbroadbanduse.Olderresidentsareleastlikelytousepublicaccess,andarealsoleastlikelytoexpressinterestintheinternet.Whiletheagingofcurrentinternetuserswillcontinuetochangethispicturesomewhat,theinitialchallengewithmanyolderresidentswillbeinterestand

awarenessofthepossibilitiesonline.

Publicaccessprovidedbylibrariesandcommunitytechnologycentershasmadeimportantinroads.Amongthosewhousetheseresourcesmostarelow‐income,African‐American,andLatinoresidents.CTCs,inparticular,servethepoorestresidentsofthecity.Wirelessaccessinpublicplaces

accommodatesinternetusersonthemove,andoverone‐thirdofChicagoresidentsgoonlinethisway.But,asfrequentuseismostlikelywithhomeaccessandbroadband,thereisaneedtoencouragehigh‐speedconnectionsathomethroughoutthecity.ExperimentsinseveralChicagoneighborhoodswith

implementingbroadbandaccesssolutionstoreachlocalresidencesandbusinessesareanimportantstepforextendingaccessanddrawinglessonsforthefuture.Criticalelementsintheseeffortsarethe

participationofcommunityorganizationsandtheprovisionoftrainingandsupport,aswellaslow‐costhardwareandsoftware.Publicaccessprovidersintheseneighborhoods,andthroughoutthecity,areimportantpartnersinsupportingthiseffortaswell.

Theinternethasbecomeacriticalresourceforwork,information,civicengagement,accessto

governmentservicesandhealth.Asmoreinformationandservicesmoveonline,thecostsincreaseforresidentswhoareexcludedfromthismedium.Someactivitiesclearlyprovidemotivationtogoonlineamonggroupsthatgenerallylagbehindininternetuse.African‐Americansaremorelikelythanwhites

tosearchforjobsonlineandresidentsofpoorneighborhoodsusetransitwebsitesmore.E‐governmentuseshowsnodifferencesbyraceandethnicity,andAfrican‐Americansarejustaslikelyaswhitestosearchforhealthinformationontheinternet.

WithbetterinformationaboutthestateofinternetuseinChicago,communityorganizations,

residents,nonprofits,businesses,educationalprovidersandpublicinstitutionscanaddressboththechallengesandopportunitiesfordigitalexcellence.

54

REFERENCES

Bimber,Bruce.2003.InformationandAmericanDemocracy:TechnologyintheEvolutionofPoliticalPower.Cambridge:CambridgeUniversityPress.

Bertot,J.C.,P.T.Jaeger,L.A.Langa,C.R.McClure.2006.PublicAccessComputingandInternetAccessin

PublicLibraries:TheRoleofPublicLibrariesinE‐GovernmentandEmergencySituations.FirstMonday,11(9),September4,2006.Availableat:http://firstmonday.org.

Fox,S..2005.HealthInformationOnline.PewInternetandAmericanLifeProject.Availableat:http://www.pewinternet.org.AccessedDecember31,2008.

Fox,S.andG.Livingston.2007.LatinosOnline.PewInternetandAmericanLifeProjectandPew

HispanicCenter.Availableat:http://www.pewinternet.org.AccessedDecember31,2008.

Granovetter,M.S.1973.Thestrengthofweakties.AmericanJournalofSociology78(6):1360‐80.

Katz,J.E.andR.E.Rice.2002.SocialConsequencesofInternetUse:Access,Involvement,and

Interaction.Cambridge,MA:MITPress.

Krueger,2002.AssessingthePotentialofInternetPoliticalParticipationintheUnitedStates.AmericanPoliticsResearch30:476‐98.

Larsen,E.andL.Rainie.2002.TheRiseoftheE‐citizen:HowPeopleUseGovernmentAgencies’WebSites.PewInternetandAmericanLifeProject.Available[online]:

www.pewinternet.org/reports/pdfs/PIP_Govt_Website_Rpt.pdfLitan,R.E.andA.M.Rivlin,eds.2002.TheEconomicPayofffromtheInternetRevolution.Washington,D.C.:BrookingsInstitutionPress.

Mayor’sAdvisoryCouncilonClosingtheDigitalDivide.2007.TheCitythatNetWorks:TransformingSocietyandEconomythroughDigitalExcellence.May2007.Availableat:

http://egov.cityofchicago.org/webportal/COCWebPortal/COC_EDITORIAL/DigitalDivide.pdf

Mossberger,K.,D.KaplanandM.Gilbert.2008.GoingOnlineWithoutConvenientAccess:ATaleofThreeCities.JournalofUrbanAffairs

Mossberger,K.,C.J.TolbertandR.S.McNeal2008.DigitalCitizenship:TheInternet,SocietyandParticipation.Cambridge,MA:MITUniversityPress.

Mossberger,K.,C.J.TolbertandM.Stansbury2003.VirtualInequality:BeyondtheDigitalDivide.

Washington,D.C.:GeorgetownUniversityPress.

Tolbert,C.andR.McNeal.2003.UnravelingtheEffectsoftheInternetonPoliticalParticipation.PoliticalResearchQuarterly56(2):175‐85.

Tolbert,C.andK.Mossberger.2006.TheEffectsofE‐GovernmentonTrustandConfidenceinGovernment.PublicAdministrationReview66(3):354‐69.

55

VanDijk,J.A.G.M.2005.TheDeepeningDivide:InequalityintheInformationSociety.London:SagePublications.

Warschauer,M.2003.TechnologyandSocialInclusion:RethinkingtheDigitalDivide.Cambridge,MA:

MITPress.

Welch,E.W.,C.HinnantandM.J.Moon.2005.LinkingCitizenSatisfactionwithE‐GovernmentwithTrustinGovernment.JournalofPublicAdministrationResearchandTheory15(1):271‐91.

West,D.M.2004.E‐governmentandtheTransformationofServiceDeliveryandCitizenAttitudes.PublicAdministrationReview64(1):15‐27.

West,D.M.2005.DigitalGovernment:TechnologyandPublicSectorPerformance.Princeton,NJ:

PrincetonUniversityPress.

56

APPENDIXA

Table1:InternetUseinGeneral(LogisticRegression)IndependentVariables

Coef. RobustStd.Err. z P>|z|

Age ‐.077 .004 ‐19.40 .000Latino ‐1.263 .177 ‐7.15 .000Black ‐.525 .134 ‐3.93 .000Asian .296 .464 0.64 .524Income .363 .031 11.74 .000Education .472 .037 12.65 .000Parent .073 .140 0.52 .603Female ‐.144 .112 ‐1.27 .205Constant 2.077 .302 6.89 .000Numberofobs=3259Waldchi2(8)=738.45Prob>chi2=0.0000PseudoR2=0.4042Logpseudolikelihood=‐1097.8317

Table2:InternetUseatHome(LogisticRegression)IndependentVariables

Coef. RobustStd.Err. z P>|z|

Age ‐.051 .003 ‐16.56 .000Latino ‐.546 .154 ‐3.56 .000Black ‐.559 .117 ‐4.77 .000Asian .447 .375 1.19 .233Income .379 .027 13.96 .000Education .371 .034 10.79 .000Parent .228 .119 1.92 .055Female ‐.043 .100 ‐0.42 .672Constant .348 .248 1.40 .161Numberofobs=3259Waldchi2(8)=778.28Prob>chi2=0.0000PseudoR2=0.3283Logpseudolikelihood=‐1355.5936

57

Table3:BroadbandInternetConnectionatHome(LogisticRegression)(1‐Broadband,0‐Dial‐upAccess)

Coef. RobustStd.Err. z P>|z|Age ‐.033 .004 ‐7.58 .000Latino ‐.553 .192 ‐2.87 .004Black .004 .179 0.02 .984Asian 1.337 .766 1.75 .081Income .247 .037 6.63 .000Educate .212 .051 4.13 .000Parent ‐.201 .155 ‐1.30 .193Female ‐.130 .146 ‐0.88 .376Constant 1.522 .365 4.17 .000Numberofobs=2226Waldchi2(8)=186.15Prob>chi2=0.0000PseudoR2=0.1242Logpseudolikelihood=‐693.87635

Table4:InternetUseatthePublicLibrary(LogisticRegression)IndependentVariables Coef. RobustStd.Err. z P>|z|Useinternetathome 1.143 .125 9.14 .000Awarenessofpublicinternetfacilityinneighborhood .187 .055 3.37 .001Easeofaccesstopublicinternetfacilityinneighborhood .608 .141 4.31 .000Age ‐.037 .003 ‐13.15 .000Latino .318 .138 2.30 .021Black .563 .111 5.06 .000Asian .077 .281 0.28 .783Income ‐.108 .025 ‐4.39 .000Education .176 .034 5.17 .000Parent .104 .096 1.08 .281Female ‐.021 .090 ‐0.23 .818Constant ‐1.480 .302 ‐4.91 .000Numberofobs=2815Waldchi2(11)=440.46Prob>chi2=0.0000PseudoR2=0.1424Logpseudolikelihood=‐1582.1751

58

Table5:ReasonsforNotUsingInternetatHome(LogisticRegression)

IamNotInterested TheCostIsTooHighIndependentVariables Coef. RobustStd.

Err.P>|z| Coef. RobustStd.

Err.P>|z|

Age .029 .004 .000 .005 .004 .263Latino ‐.079 .225 .725 .647 .225 .004Black ‐.280 .161 .082 .104 .166 .529Asian .784 .746 .293 ‐.879 .815 .281Income .120 .041 .004 ‐.256 .043 .000Female ‐.158 .145 .275 .607 .146 .000Education ‐.115 .045 .012 ‐.084 .047 .073Parent ‐.168 .196 .392 ‐.176 .197 .370Constant ‐1.45 .391 .000 .405 .371 .275 Numberofobs=1011

Waldchi2(8)=90.17Prob>chi2=0.0000PseudoR2=0.0763Logpseudolikelihood=‐645.9321

Numberofobs=1011Waldchi2(8)=103.14Prob>chi2=0.0000PseudoR2=0.0876Logpseudolikelihood=‐637.9946

Table5continued:ReasonsforNotUsingInternetatHome(LogisticRegression)

IcanUseItSomewhereElse IDon'tHaveTimetoUsetheInternetIndependent

Variables Coef. RobustStd.Err.

P>|z| RobustStd.Err.

RobustStd.Err.

P>|z|

Age ‐.032 .004 .000 ‐.005 .005 .273Latino .156 .220 .478 .703 .241 .004Black .184 .165 .264 ‐.402 .201 .046Asian ‐.357 .657 .587 .503 .689 .465Income ‐.025 .041 .541 .076 .044 .082Female .067 .142 .639 ‐.522 .161 .001Education .142 .047 .002 ‐.073 .053 .167Parent ‐.314 .192 .101 ‐.124 .219 .573Constant 1.35 .377 .000 ‐.575 .422 .173 Numberofobs=1011

Waldchi2(8)=74.28Prob>chi2=0.0000PseudoR2=0.0596Logpseudolikelihood=‐658.65633

Numberofobs=1010Waldchi2(8)=54.60Prob>chi2=0.0000PseudoR2=0.0540Logpseudolikelihood=‐514.93941

59

Table5continued:ReasonsforNotUsingInternetatHome(LogisticRegression)

It'sTooDifficulttoUse IamWorriedAboutPrivacyIndependentVariables Coef. RobustStd.

Err.P>|z| Coef. RobustStd.

Err.P>|z|

Age .037 .005 .000 .004 .004 .312Latino .573 .228 .012 1.15 .233 .000Black ‐.273 .169 .107 .034 .163 .835Asian ‐.412 .648 .525 ‐.339 .618 .584Income ‐.087 .041 .033 .019 .040 .628Female .250 .147 .089 .592 .143 .000Education ‐.201 .047 .000 ‐.063 .046 .164Parent .260 .191 .173 .064 .191 .739Constant ‐1.62 .382 .000 ‐.379 .377 .315 Numberofobs=1011

Waldchi2(8)=103.36Prob>chi2=0.0000PseudoR2=0.0930Logpseudolikelihood=‐627.89249

Numberofobs=1010Waldchi2(8)=68.29Prob>chi2=0.0000PseudoR2=0.0515Logpseudolikelihood=‐651.02475

60

APPENDIXB.MULTILEVELMODELSTable 1: Probability of Internet Use in Any Place: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.078 0.000 -0.077 0.000 (0.004) (0.004) Latino -0.964 0.000 -1.041 0.000 (0.201) (0.211) Black -0.099 0.630 -0.281 0.112 (0.205) (0.177) Asian 0.374 0.426 0.335 0.438 (0.470) (0.431) Income 0.359 0.000 0.366 0.000 (0.032) (0.034) Education 0.470 0.000 0.464 0.000 (0.038) (0.042) Parent 0.118 0.425 0.112 0.435 (0.147) (0.143) Female -0.117 0.294 -0.115 0.309 (0.111) (0.113) Geographic Level Variables Pct. Latino -0.010 0.039 -0.006 0.348 (0.005) (0.006) Pct. Black -0.009 0.007 -0.009 0.013 (0.003) (0.004) Pct. Asian -0.001 0.911 -0.011 0.267 (0.010) (0.010) Pct. Below Poverty Line 0.006 0.367 0.023 0.004 (0.007) (0.008) Pct. High School Graduate 0.002 0.762 0.014 0.158 (0.007) (0.010) Constant 2.208 0.002 1.024 0.285 (0.718) (0.959) Observations 3117 3117 Pseudo R-squared 0.4107 0.4102 Log-likelihood -1045.5454 -1046.4632 Wald Chi2 776.7520 821.2099 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

61

Table 2: Probability of Home Internet Access: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.051 0.000 -0.051 0.000 (0.003) (0.003) Latino -0.276 0.117 -0.327 0.067 (0.176) (0.178) Black -0.333 0.082 -0.482 0.001 (0.192) (0.151) Asian 0.410 0.327 0.417 0.155 (0.419) (0.293) Income 0.373 0.000 0.378 0.000 (0.028) (0.026) Education 0.364 0.000 0.365 0.000 (0.036) (0.033) Parent 0.284 0.025 0.273 0.052 (0.126) (0.140) Female -0.011 0.915 -0.011 0.907 (0.099) (0.096) Geographic Level Variables Pct. Latino -0.007 0.031 -0.009 0.003 (0.003) (0.003) Pct. Black -0.003 0.221 -0.005 0.083 (0.003) (0.003) Pct. Asian 0.015 0.099 0.003 0.657 (0.009) (0.007) Pct. Below Poverty Line -0.002 0.708 0.009 0.132 (0.005) (0.006) Constant 0.474 0.112 0.445 0.180 (0.298) (0.332) Observations 3117 3117 Pseudo R-squared 0.3318 0.3308 Log-likelihood -1298.6818 -1300.7910 Wald Chi2 780.9496 827.6910 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

62

Table 3: Probability of Internet Use at the Public Library: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Ease of Access (self reported) 0.303 0.000 0.298 0.000 (0.055) (0.050) Age -0.044 0.000 -0.044 0.000 (0.003) (0.003) Latino 0.278 0.064 0.212 0.144 (0.150) (0.146) Black 0.647 0.000 0.677 0.000 (0.163) (0.142) Asian 0.153 0.584 0.179 0.467 (0.280) (0.246) Income -0.053 0.027 -0.055 0.038 (0.024) (0.026) Education 0.259 0.000 0.258 0.000 (0.032) (0.033) Parent 0.136 0.160 0.127 0.167 (0.097) (0.092) Female 0.030 0.730 0.024 0.781 (0.087) (0.088) Geographic Level Variables Pct. Latino -0.017 0.000 -0.016 0.004 (0.004) (0.006) Pct. Black -0.007 0.006 -0.007 0.031 (0.003) (0.003) Pct. Asian 0.002 0.748 0.003 0.621 (0.006) (0.006) Pct. Below Poverty Line -0.011 0.033 -0.020 0.023 (0.005) (0.009) Pct. High School Graduate -0.027 0.000 -0.029 0.001 (0.007) (0.009) Constant 1.899 0.006 2.178 0.007 (0.684) (0.809) Observations 2794 2794 Pseudo R-squared 0.1251 0.1233 Log-likelihood -1596.4057 -1599.5373 Wald Chi2 359.2470 412.7878 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

63

Table 4: Probability of Internet Use at a CTC: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Ease of Use (self reported) 0.316 0.009 0.290 0.002 (0.121) (0.095) Age -0.031 0.000 -0.031 0.000 (0.005) (0.005) Latino 0.179 0.598 0.309 0.410 (0.340) (0.376) Black 0.145 0.646 0.513 0.031 (0.316) (0.238) Asian 0.195 0.732 0.659 0.213 (0.569) (0.530) Income 0.012 0.788 0.007 0.874 (0.045) (0.046) Education 0.117 0.065 0.094 0.201 (0.063) (0.074) Parent 0.414 0.049 0.427 0.011 (0.210) (0.167) Female 0.069 0.722 0.045 0.811 (0.194) (0.190) Geographic Level Variables Pct. Latino 0.015 0.265 0.007 0.716 (0.013) (0.018) Pct. Black 0.009 0.394 -0.001 0.961 (0.011) (0.011) Pct. Asian 0.009 0.744 -0.002 0.940 (0.028) (0.028) Pct. Below Poverty Line 0.015 0.074 0.010 0.462 (0.009) (0.014) Pct. High School Graduate 0.008 0.582 0.009 0.659 (0.014) (0.020) Constant -3.981 0.024 -3.086 0.189 (1.765) (2.351) Observations 886 1102 Pseudo R-squared 0.0747 0.0680 Log-likelihood -400.8104 -493.4263 Wald Chi2 64.9925 232.4717 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval. Subsample of respondents from census tracts with above average poverty levels.

64

Table 5: Probability of High Speed (Broadband) versus Dial-up Access: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.034 0.000 -0.034 0.000 (0.005) (0.005) Latino -0.481 0.021 -0.498 0.021 (0.208) (0.215) Black -0.021 0.936 -0.136 0.627 (0.258) (0.279) Asian 1.371 0.077 1.327 0.088 (0.775) (0.778) Income 0.234 0.000 0.244 0.000 (0.037) (0.036) Education 0.207 0.000 0.202 0.000 (0.051) (0.053) Parent -0.232 0.171 -0.239 0.165 (0.169) (0.172) Female -0.142 0.335 -0.140 0.283 (0.147) (0.130) Geographic Level Variables Pct. Latino 0.002 0.723 0.001 0.932 (0.006) (0.008) Pct. Black 0.003 0.465 0.001 0.894 (0.004) (0.005) Pct. Asian -0.008 0.432 -0.014 0.197 (0.010) (0.011) Pct. Below Poverty Line -0.009 0.356 0.011 0.453 (0.010) (0.015) Pct. High School Graduate 0.007 0.472 0.011 0.001 (0.010) 0.001 Constant 1.236 0.202 0.726 0.594 (0.968) (1.363) Observations 2113 2113 Pseudo R-squared 0.1270 0.1262 Log-likelihood -652.9279 -653.4911 Wald Chi2 190.0820 246.7003 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

65

Table 6: Probability of Citing Cost as a Reason for No Home Internet Access: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age 0.006 0.143 0.005 0.313 (0.004) (0.005) Latino 0.310 0.212 0.509 0.009 (0.248) (0.196) Black -0.020 0.946 0.052 0.804 (0.299) (0.210) Asian -0.951 0.215 -0.906 0.228 (0.767) (0.752) Income -0.253 0.000 -0.253 0.000 (0.047) (0.046) Education -0.091 0.065 -0.099 0.029 (0.049) (0.046) Parent -0.181 0.387 -0.199 0.309 (0.209) (0.196) Female 0.585 0.000 0.587 0.000 (0.147) (0.126) Geographic Level Variables Pct. Latino 0.020 0.002 0.020 0.007 (0.006) (0.007) Pct. Black 0.008 0.084 0.010 0.013 (0.004) (0.004) Pct. Asian 0.011 0.371 0.018 0.038 (0.013) (0.009) Pct. Below Poverty Line 0.006 0.382 -0.008 0.473 (0.007) (0.011) Pct. High School Graduate 0.019 0.047 0.020 0.114 (0.010) (0.013) Constant -1.807 0.076 -1.737 0.172 (1.018) (1.273) Observations 984 984 Pseudo R-squared 0.0959 0.0924 Log-likelihood -615.4857 -617.8867 Wald Chi2 100.6470 101.4212 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

66

Table 7: Probability of Citing Too Difficulty as a Reason for No Home Internet Access: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age 0.038 0.000 0.038 0.000 (0.005) (0.005) Latino 0.603 0.022 0.586 0.051 (0.264) (0.300) Black -0.231 0.435 -0.303 0.248 (0.296) (0.262) Asian -0.352 0.557 -0.326 0.610 (0.598) (0.638) Income -0.094 0.027 -0.096 0.021 (0.043) (0.042) Education -0.203 0.000 -0.210 0.000 (0.049) (0.049) Parent 0.256 0.197 0.259 0.210 (0.198) (0.207) Female 0.229 0.144 0.218 0.185 (0.157) (0.165) Geographic Level Variables Pct. Latino 0.003 0.659 0.012 0.107 (0.006) (0.007) Pct. Black 0.005 0.189 0.010 0.036 (0.004) (0.005) Pct. Asian 0.007 0.569 0.010 0.388 (0.013) (0.011) Pct. Below Poverty Line -0.025 0.001 -0.027 0.009 (0.008) (0.010) Pct. High School Graduate -0.007 0.477 0.008 0.538 (0.009) (0.013) Constant -1.034 0.273 -2.408 0.045 (0.943) (1.198) Observations 984 984 Pseudo R-squared 0.1043 0.1039 Log-likelihood -602.4645 -602.7304 Wald Chi2 120.5170 125.6407 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

67

Table 8: Probability of Citing a Lack of Interest as a Reason for No Internet Access: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age 0.028 0.000 0.029 0.000 (0.005) (0.005) Latino -0.084 0.727 -0.127 0.603 (0.240) (0.243) Black 0.176 0.512 -0.021 0.926 (0.269) (0.229) Asian 0.828 0.288 0.824 0.298 (0.780) (0.791) Income 0.110 0.015 0.119 0.008 (0.045) (0.045) Education -0.123 0.010 -0.130 0.010 (0.048) (0.050) Parent -0.190 0.326 -0.216 0.257 (0.193) (0.191) Female -0.154 0.311 -0.138 0.358 (0.152) (0.150) Geographic Level Variables Pct. Latino 0.003 0.538 0.009 0.189 (0.004) (0.007) Pct. Black -0.003 0.500 0.004 0.453 (0.004) (0.005) Pct. Asian 0.010 0.465 0.020 0.166 (0.014) (0.014) Median Income 0.000 0.099 0.000 0.033 (0.000) (0.000) Constant -1.962 0.000 -2.728 0.000 (0.540) (0.749) Observations 984 984 Pseudo R-squared 0.0812 0.0816 Log-likelihood -625.1008 -624.8473 Wald Chi2 90.5790 86.2566 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

68

Table 9: Probability of Internet Use at Work: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.013 0.008 -0.013 0.002 (0.005) (0.004) Latino -0.833 0.000 -0.754 0.000 (0.190) (0.192) Black -0.495 0.042 -0.521 0.033 (0.243) (0.245) Asian -0.443 0.176 -0.413 0.292 (0.327) (0.392) Income 0.226 0.000 0.225 0.000 (0.036) (0.032) Education 0.499 0.000 0.494 0.000 (0.050) (0.052) Parent -0.070 0.615 -0.047 0.703 (0.139) (0.124) Female -0.035 0.775 -0.049 0.626 (0.124) (0.101) Geographic Level Variables Pct. Latino 0.025 0.000 0.030 0.000 (0.006) (0.007) Pct. Black 0.009 0.014 0.010 0.022 (0.004) (0.004) Pct. Asian 0.001 0.924 -0.003 0.829 (0.008) (0.013) Pct. Below Poverty Line 0.006 0.435 0.013 0.187 (0.007) (0.010) Pct. High School Graduate 0.030 0.001 0.041 0.000 (0.009) (0.011) Constant -5.513 0.000 -6.596 0.000 (0.905) (1.115) Observations 1546 1546 Pseudo R-squared 0.2332 0.2323 Log-likelihood -784.3304 -785.3172 Wald Chi2 328.0555 293.6138 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval. Subsample of employed respondents (full or part-time) only.

69

Table 10: Probability of Internet Use for Health Information: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.048 0.000 -0.048 0.000 (0.003) (0.003) Latino -0.529 0.001 -0.566 0.000 (0.159) (0.155) Black -0.038 0.828 -0.121 0.471 (0.174) (0.168) Asian 0.203 0.509 0.200 0.444 (0.308) (0.261) Income 0.304 0.000 0.308 0.000 (0.026) (0.031) Education 0.366 0.000 0.362 0.000 (0.032) (0.030) Parent 0.250 0.033 0.244 0.028 (0.117) (0.111) Female 0.283 0.002 0.283 0.004 (0.091) (0.098) Geographic Level Variables Pct. Latino -0.007 0.095 -0.006 0.170 (0.004) (0.004) Pct. Black -0.005 0.059 -0.007 0.011 (0.003) (0.003) Pct. Asian 0.014 0.083 0.006 0.277 (0.008) (0.006) Pct. Below Poverty Line 0.005 0.398 0.018 0.003 (0.006) (0.006) Pct. High School Graduate 0.002 0.719 0.008 0.285 (0.006) (0.008) Constant -0.139 0.830 -0.761 0.324 (0.649) (0.772) Observations 3116 3116 Pseudo R-squared 0.2923 0.2922 Log-likelihood -1438.2169 -1438.3387 Wald Chi2 742.3475 935.8000 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

70

Table 11: Probability of Online Job Search: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.078 0.000 -0.078 0.000 (0.003) (0.003) Latino -0.286 0.078 -0.329 0.045 (0.162) (0.164) Black 0.473 0.003 0.381 0.029 (0.161) (0.174) Asian 0.329 0.203 0.335 0.123 (0.258) (0.217) Income 0.054 0.020 0.057 0.040 (0.023) (0.028) Education 0.314 0.000 0.319 0.000 (0.032) (0.032) Parent 0.015 0.891 0.011 0.937 (0.109) (0.134) Female 0.011 0.905 0.008 0.936 (0.095) (0.096) Geographic Level Variables Pct. Latino -0.006 0.173 -0.008 0.191 (0.004) (0.006) Pct. Black -0.002 0.338 -0.003 0.358 (0.003) (0.003) Pct. Asian 0.001 0.907 -0.004 0.699 (0.009) (0.011) Pct. Below Poverty Line 0.001 0.906 0.001 0.939 (0.005) (0.007) Pct. High School Graduate -0.004 0.587 -0.010 0.336 (0.007) (0.010) Constant 2.313 0.001 2.844 0.005 (0.675) (1.019) Observations 3115 3115 Pseudo R-squared 0.2715 0.2715 Log-likelihood -1572.6548 -1572.6726 Wald Chi2 889.7194 738.0208 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

71

Table 12: Probability of Internet Use for Classes or Training: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.034 0.000 -0.034 0.000 (0.003) (0.003) Latino -0.147 0.324 -0.158 0.249 (0.149) (0.137) Black 0.049 0.739 0.009 0.950 (0.146) (0.149) Asian 0.214 0.384 0.210 0.398 (0.246) (0.249) Income 0.106 0.000 0.106 0.000 (0.023) (0.024) Education 0.361 0.000 0.362 0.000 (0.036) (0.037) Parent 0.060 0.546 0.060 0.592 (0.100) (0.111) Female -0.005 0.957 -0.010 0.909 (0.090) (0.088) Individual Level Variables Pct. Latino -0.001 0.746 -0.003 0.603 (0.004) (0.005) Pct. Black -0.000 0.886 -0.000 0.996 (0.003) (0.003) Pct. Asian 0.003 0.671 0.002 0.768 (0.006) (0.007) Pct. Below Poverty Line 0.003 0.525 0.002 0.797 (0.005) (0.008) Pct. High School Graduate -0.005 0.463 -0.009 0.332 (0.007) (0.009) Constant -1.285 0.054 -0.956 0.264 (0.666) (0.856) Observations 3115 3115 Pseudo R-squared 0.1322 0.1323 Log-likelihood -1662.6053 -1662.4013 Wald Chi2 431.8355 510.0489 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

72

Table 13: Probability of Internet Use for Public Transportation Information: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.053 0.000 -0.052 0.000 (0.003) (0.003) Latino -0.783 0.000 -0.840 0.000 (0.155) (0.137) Black -0.319 0.064 -0.342 0.028 (0.172) (0.156) Asian 0.128 0.687 0.089 0.748 (0.318) (0.278) Income 0.169 0.000 0.176 0.000 (0.023) (0.023) Education 0.263 0.000 0.263 0.000 (0.030) (0.027) Parent 0.119 0.257 0.116 0.260 (0.105) (0.103) Female 0.040 0.633 0.047 0.581 (0.085) (0.085) Geographic Level Variables Pct. Latino -0.005 0.154 -0.004 0.354 (0.004) (0.004) Pct. Black -0.005 0.051 -0.007 0.006 (0.002) (0.003) Pct. Asian -0.003 0.683 -0.008 0.124 (0.007) (0.005) Pct. Below Poverty Line 0.011 0.045 0.026 0.000 (0.005) (0.007) Pct. High School Graduate 0.010 0.071 0.015 0.097 (0.006) (0.009) Constant 0.261 0.645 -0.298 0.691 (0.568) (0.751) Observations 3115 3115 Pseudo R-squared 0.2289 0.2292 Log-likelihood -1652.4466 -1651.7189 Wald Chi2 680.1880 775.0732 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

73

Table 13: Probability of Internet Use for Information about Politics: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.043 0.000 -0.042 0.000 (0.003) (0.003) Latino -0.585 0.000 -0.538 0.000 (0.151) (0.142) Black -0.411 0.011 -0.388 0.018 (0.163) (0.164) Asian 0.067 0.798 0.020 0.934 (0.263) (0.240) Income 0.238 0.000 0.240 0.000 (0.025) (0.029) Education 0.394 0.000 0.381 0.000 (0.033) (0.032) Parent -0.298 0.007 -0.282 0.004 (0.110) (0.098) Female -0.166 0.059 -0.157 0.066 (0.088) (0.085) Geographic Level Variables Pct. Latino 0.007 0.059 0.013 0.021 (0.004) (0.006) Pct. Black 0.003 0.238 0.003 0.397 (0.002) (0.003) Pct. Asian 0.004 0.615 0.007 0.341 (0.007) (0.008) Pct. Below Poverty Line 0.006 0.252 0.024 0.007 (0.006) (0.009) Pct. High School Graduate 0.018 0.004 0.036 0.000 (0.006) (0.010) Constant -2.155 0.000 -3.879 0.000 (0.618) (0.892) Observations 3115 3115 Pseudo R-squared 0.2609 0.2640 Log-likelihood -1592.5272 -1585.8280 Wald Chi2 832.9049 783.1396 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

74

Table 14: Probability of Internet Use for Information about Government: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.038 0.000 -0.037 0.000 (0.003) (0.003) Latino -0.554 0.000 -0.550 0.000 (0.143) (0.142) Black -0.267 0.110 -0.215 0.163 (0.167) (0.154) Asian -0.035 0.896 0.018 0.936 (0.270) (0.219) Income 0.209 0.000 0.211 0.000 (0.023) (0.028) Education 0.367 0.000 0.368 0.000 (0.031) (0.029) Parent 0.150 0.162 0.148 0.199 (0.107) (0.115) Female -0.116 0.183 -0.115 0.122 (0.087) (0.074) Geographic Level Variables Pct. Latino 0.005 0.187 0.005 0.388 (0.004) (0.006) Pct. Black 0.006 0.026 0.003 0.277 (0.003) (0.003) Pct. Asian 0.017 0.017 0.009 0.147 (0.007) (0.006) Pct. Below Poverty Line -0.002 0.681 0.007 0.302 (0.005) (0.007) Pct. High School Graduate 0.009 0.122 0.012 0.147 (0.006) (0.009) Constant -1.418 0.016 -1.751 0.039 (0.586) (0.849) Observations 3116 3116 Pseudo R-squared 0.2250 0.2239 Log-likelihood -1653.2776 -1655.7325 Wald Chi2 704.9155 740.5732 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

75

Table 15: Probability of Using the City of Chicago’s Website: Multilevel Logistic Regression Estimates, Clustering by Census Tract or Chicago Community Area

Model 1: Census Tract Model 2: Community Area Coef. (S.E.) p>|z| Coef. (S.E.) p>|z| Individual Level Variables Age -0.027 0.000 -0.027 0.000 (0.002) (0.003) Latino -0.167 0.226 -0.196 0.138 (0.138) (0.132) Black -0.080 0.628 -0.059 0.670 (0.166) (0.138) Asian -0.202 0.426 -0.169 0.500 (0.254) (0.250) Income 0.208 0.000 0.209 0.000 (0.024) (0.025) Education 0.314 0.000 0.321 0.000 (0.030) (0.031) Parent 0.311 0.002 0.305 0.001 (0.102) (0.092) Female 0.184 0.031 0.176 0.035 (0.085) (0.084) Geographic Level Variables Pct. Latino -0.002 0.649 -0.004 0.351 (0.003) (0.005) Pct. Black 0.002 0.513 -0.000 0.968 (0.002) (0.002) Pct. Asian 0.008 0.230 0.003 0.559 (0.006) (0.005) Pct. Below Poverty Line -0.002 0.595 -0.003 0.648 (0.004) (0.006) Pct. High School Graduate -0.008 0.119 -0.016 0.049 (0.005) (0.008) Constant -0.827 0.123 -0.203 0.797 (0.536) (0.788) Observations 3112 3112 Pseudo R-squared 0.1573 0.1578 Log-likelihood -1816.7383 -1815.7021 Wald Chi2 541.2771 634.1826 Prob. > chi2 0.0000 0.0000

Unstandardized logistic regression coefficients with robust standard errors in parentheses. Standard errors adjusted by clustering cases by geographic area (census tract or Chicago community area). Probabilities based on two-tailed significance tests. Variables with a p-value of .10 or lower are considered statistically significant with a 90% confidence interval; a p-value of .05 or lower is considered statistically significant with a 95% confidence interval.

76

APPENDIXC

UNIVERSITYOFIOWAHAWKEYEPOLLDEPARTMENTOFPOLITICALSCIENCE

CHICAGOINTERNETSURVEYCONDUCTEDJUNE23‐AUGUST7,2008QUESTIONNAIRE

INTRODUCTION:Hello,Iam_________________,callingfromtheUniversityofIowa.WearestudyingtheroleoftheinternetinChicago.Yourphonenumberwasselectedatrandomtorepresentyourneighborhoodinthisstudy.Iamnotsellinganythingandjustneedafewminutes.ATTEMPTTOIMPROVEYOUNGMALERESPONSERATESYNGMALE:I’dliketoasksomequestionsoftheyoungestmalewhois18yearsorolderandnowathome.[IFRISMALE]Wouldthatbeyou?IFRESP:YESCONTINUEWITH[AGESCREEN]IFRESP:LETMEGETHIMWAITFORNEWPERSON,GOTO[REINTRO]IFRESP:NOMALE,ASK:

Isthereanotherpersonover18Icanspeakwith?CouldIspeakwithyou?IFCURRENTR.GOTO[AGESCREEN]IFWILLGETSOMEONE,WAITFORNEWPERSON,GOTO[REINTRO]IFNO,GOTO[SCHEDULE].REINTRO:Hello,Iam_________________,callingfromtheUniversityofIowa.WearestudyingtheroleoftheinternetinChicago.Yourphonenumberwasselectedatrandomtorepresentyourneighborhoodinthisstudy.Iamnotsellinganythingandjustneedafewminutes.AGESCREEN:SCREENAGEFOR18andOVERQ1AAGE First,Ineedtomakesurewearereachingpeopleofallages18orover.Wouldyoutell

meyourage? ________years 97 97orolder 99 Don’tknow/Refused[VOL.]IFNOT18orOVER GOTOEND[INELIGIBLE]

77

CONSENT: WeinviteyoutoparticipateinastudyabouttechnologyaccessinChicagobeingconductedby

researchersfromtheUniversityofIowa.Yourphonenumberwaschosenatrandomtorepresentyourneighborhood.Ifyouagree,wewouldliketoaskyouaseriesofquestions.Youmayskipanyquestionsthatyouprefernottoanswer.Thiswilltakeabout12minutes.Yourresponsesareconfidentialanditwillnotbepossibletolinkyoutothem.Thissurveyisvoluntary.Yourwillingnesstoanswermyquestionswillindicateyourconsenttouseyouranswersinourresearchproject.Q1BAreyouwillingtoparticipateinthissurvey?

0 NOGOTOEND[ATTEMPTCONVERT]1 YES

Q2INTUSE OK,thanks!First,doyoueverusetheInternetinanyplace(home,work,school,

anywhereelse)?

0 No1 Yes8 Don’tKnow9 Refused

Q3INFO Weareinterestedintheinformationpeoplefeeltheyneedintheirdailyliveswhether

ornotitcomesfromtheinternet.Wouldyousaythatitisveryimportant,important,notveryimportant,ornotimportantatallforyoutogetinformationon:[PROMPTWITHRESPONSEOPTIONSASNEEDED]

Q3A JobsorbetterjobopportunitiesQ3B EducationortrainingformyselfQ3C Mychild'sschoolQ3D HealthcareorhealthissuesQ3E MyneighborhoodQ3F GovernmentorservicesprovidedbygovernmentQ3G Placestolive

RESPONSEOPTIONS

1 Veryimportant2 Important3 Notveryimportant4 Notatallimportant

78

8 Don’tKnow9 Refused

IFQ2ISNOTYES(1)GOTOQ6Q4FREQUSE AbouthowoftendoyouusetheInternet?[Readoptions]

1 Severaltimesaday2 Aboutonceaday3 3‐5daysaweek4 1‐2daysaweek5 Everyfewweeks6 Lessoften9 Refused

Q5HOWLONG AbouthowmanyyearshaveyoubeenanInternetuser?[ENTERYEARS]

_______years8 Don’tKnow9 Refused

Q6HCOMPDoyouhaveacomputerathome?

0 NOGOTOQ81 YES8 Don’tKnowGOTOQ89 RefusedGOTOQ8

Q7INETHOM DoyoueverusetheInternetathome?

0 NO1 YESGOTOQ108 Don’tKnow9 Refused

Q8NOACCESS Iamgoingtoreadalistofreasonswhysomepeopledon’tusetheInternetathome.

Foreach,justtellmewhetheritappliestoyoubysayingyesifitdoes,ornoifitdoesnot.

79

Q8A Idon’tneedit,I’mnotinterestedQ8B ThecostistoohighformeQ8C IcanuseitsomewhereelseQ8D Idon’thavetimetousetheInternetQ8E It’stoodifficulttouseQ8F IamworriedaboutprivacyandpersonalinformationonlineQ8G TheInternetisdangerousQ8H It’shardformetousetheinformationinEnglishQ8I IhaveaphysicalimpairmentthatmakesitdifficulttousetheInternet

RESPONSEOPTIONS

0 NO1 YES

8 Don’tKnow9 Refused

Q9MAIN Now,pleasetellmeinacouplewordstheMAINreasonyoudon’tusetheInternetat

home?[DON’TREAD,CODEANSWERTOBESTFIT]

1 Idon’tneedit,I’mnotinterested2 Thecostistoohighforme3 Icanuseitsomewhereelse4 Idon’thavetimetousetheInternet5 It’stoodifficulttouse6 Iamworriedaboutprivacyandpersonalinformationonline7 TheInternetisdangerous8 It’shardformetousetheinformationinEnglish9 IhaveaphysicalimpairmentthatmakesitdifficulttousetheInternet10 Other11 Don’tKnow12 Refused

Q9AINTFUT Isthereanythingthatmightmakeyouinterestedinusingtheinternetinthefuture?If

so,justtellmeinacouplewordswhatitis.Ifnot,justtellmeno.[OPENENDED,RECORDVERBATIM]

IFQ7ISNOT1GOTOQ14Q10HCONTYP DoesthecomputeryouuseatHOMEconnecttotheInternetthroughadial‐up

telephoneline,ordoyouhavesometypeofhighspeedconnection,?

1 Dial‐uptelephone

80

2 HighSpeedConnectionGOTOQ128 Don’tKnowGOTOQ129 RefusedGOTOQ12

Q11NOBBND WhatistheMAINreasonyoudonothavehigh‐speed(thatis,fasterthandial‐up)

Internetaccessathome?[DON’TREAD,CODEANSWERTOBESTFIT]

1 Don’tneeditornotinterested2 Costsaretoohighforme3 Canuseitsomewhereelse4 Idon’thavetimetousetheInternet5 Toodifficulttouseordon’tnowhowtouse6 Nocomputerorcomputerinadequate7 Privacyandsecurity8 Notavailableinarea9 Other10 Don’tKnow11 Refused

Q12WHEREINT WherewouldyousaythatyouusetheInternetmostoften?[DON’TREAD,CODEANSWERTO

BESTFIT]

1 Home2 Work3 School4 Alibraryorpublicplace5 Friendorrelative’shouse6 CoffeeShoporInternetCafe7 Other

8 Don’tKnow(Vol.)9 Refused(Vol.)

Q13WHERESEC WherewouldyousaythatyouusetheInternetmostoftenafterthat?[DON’TREAD,

CODEANSWERTOBESTFIT]

1 Home2 Work3 School4 Alibraryorpublicplace5 Friendorrelative’shouse

81

6 CoffeeShoporInternetCafe7 Other

8 Don’tKnow(Vol.)9 Refused(Vol.)

Q14CTCAWAR Asfarasyouknow,isthereaplaceyoucangoinyourneighborhoodwherethe

Internetispubliclyavailabletoanyonewhowantstouseit?SuchplacesareoftencalledCommunityTechnologyCenters.

0 NO1 YES

8 Don’tKnow(Vol.)9 Refused(Vol.)

IFQ2ISNOTYES(1)SKIPTOQ16Q15CTCHELP HaveyoueverusedtheInternetorgottenhelpusingtheInternetataCommunity

TechnologyCenter?

0 NO1 YES

8 Don’tKnow(Vol.)9 Refused(Vol.)

Q16PUBLICACC Wouldyousaythatitiseasyordifficulttogettoplacesinyourcommunitywithpublic

accesstotheInternet,likealibraryoracommunitytechnologycenter?Wouldyousaythatitisveryeasy,somewhateasy,somewhatdifficultorverydifficult?

1 Veryeasy2 Somewhateasy3 Somewhatdifficult4 Verydifficult

8 Don’tknow(Vol)9 Refused(Vol)

IFQ2ISNOTYES(1)SKIPTOQ24

82

Q17LIBRARY HaveyouusedtheInternetattheChicagoPublicLibrary?

0 NO1 YES

8 Don’tKnow(Vol.)9 Refused(Vol.)

IFQ15ISNOTYES(1)ANDQ17ISNOTYES(1)GOTOQ19Q18WHYLIB IamgoingtoreadanumberofstatementsaboutwhyyouusetheInternetatthe

libraryoratacommunitytechnologycenter.Pleaserespondyesornotoeachstatement.

Q18A Idon’thaveacomputerathomeormycomputeritslowQ18B Idon’thaveanInternetconnectionathomeQ18C IneededhelptofindinformationQ18D IneededhelptousethecomputerQ18E MycomputerorInternetconnectionsathomearen’tworkingQ18F ItisconvenientQ18G TotakeaclassQ18H Totakemychildrentodotheirhomework

RESPONSEOPTIONS

0 NO1 YES

8 Don’tKnow(Vol.)9 Refused(Vol.)

Q19ACTIVITIES Iamgoingtoreadalistofthingsyoumightdoontheinternet.Pleasetellmehow

frequentlyyoudoeachbysayingifyoudothesethingsdaily,afewtimesperweek,afewtimespermonth,rarely,ornever.[PROMPTWITHOPTIONSASNEEDED]

Q19A GetnewsonlineQ19B DoworkforyourjobQ19C UseasocialnetworkingsitelikeFacebookQ19D SendorreceiveemailQ19E UseacellphonetoconnecttotheInternetQ19F ReadablogQ19G UsewirelessaccesstoconnecttotheInternetinapublicplace

83

RESPONSEOPTIONS

1 Daily2 Afewtimesperweek3 Afewtimespermonth4 Rarely5 Never

8 Don’tKnow(Vol.)9 Refused(Vol.)

Q20ONLINE I’mgoingtoreadanotherlist.ForeachitempleasetellmeifyoueverusetheInternet

todoanyofthefollowingthingsbyjustsayingyesorno.DoyoueverusetheInternetto;[PROMPTASNECESSARY–JUSTTELLMEYESORNO]

Q20A FindhealthinformationQ20B LookforajoborinformationonjobsQ20C TakeaclassortrainingonlineQ20D GetinformationaboutpoliticsQ20E GetinformationabouttrainsorbusesusingtheCTAorRTAwebsiteQ20F Find information on government Q20G UsetheCityofChicagowebsite

RESPONSEOPTIONS

0 NO1 YES

8 Don’tKnow(Vol.)9 Refused(Vol.)

IFQ20GISYES(1)ASKQ21OTHERWISESKIPTOQ23Q21CHICAGO PleasetellmeifyouhaveeverusedtheCityofChicagowebsitetodoanyofthe

following.Justtellmeyesorno.[PROMPTASNECESSARY–JUSTTELLMEYESORNO]

Q21A GetanaddressorphonenumberQ21B ContactofficialsQ21C GettouristorrecreationinformationQ21D Getinformationaboutservices(otherthanrecreationortourism)Q21E Completeatransactiononline,suchaspayingabillorfine,orfilingaformonlineQ21F Lookforgovernmentpoliciesordocuments

84

RESPONSEOPTIONS

0 NO 1 YES 8 Don’tKnow(Vol.)9 Refused(Vol.)

Q22EVALCHI IamgoingtoreadyousomestatementsabouttheCityofChicagowebsite.Pleasetell

mewhetheryoustronglyagree,agree,disagreeorstronglydisagreewitheachstatement.[PROMPTASNECESSARYWITHRESPONSEOPTIONS]

Q22A ThewebsitehadtheinformationIneeded.Q22B Thewebsitewaseasytouseandfindinformation.Q22C Thewebsitewasdifficulttouseandcomplex.

RESPONSEOPTIONS

1 StronglyAgree2 Agree3 Disagree4 StronglyDisagree8 Don’tKnow(Vol.)9 Refused(Vol.)

Q23SKILLS Iamgoingtoreadsomethingspeoplesometimesdoonline.Pleasetellmeifyou

alreadyknowhowtodoeachone,orifyouwouldneedsomeoneelsetohelpyou.

Q23A UseasearchenginetofindinformationonlineQ23B SendandreceiveemailQ23C DownloadandfilloutaformQ23D UploadimagesorfilestoawebsiteoremailQ23E Createawebsite

RESPONSEOPTIONS1Knowhow2Needhelp8Don’tKnow(Vol.)9Refused(Vol.)

Q24

85

POLICY1 There'sbeentalkaboutbuildingawirelessnetworkinneighborhoodsinChicago.Whichofthefollowingshouldbethefocusindoingthisproject?Shoulditbeonmakingwirelessavailable:[RANDOMIZEORDEROFFIRSTTHREEOPTIONS;READINORDER]

1 alloverthecity 2 inlow‐incomeneighborhoods 3 inpublicschools,librariesandotherpublicplaces 4 ordoyouthinktheyshouldnotworkonthisproject?

8 Don’tKnow(Vol.) 9 Refused(Vol.)

Q25POLICY2 Wouldyousupportaprojecttoprovidefreewirelessinternetaccessifitcausedasmall

increaseinfeesortaxes?

0 NO1 YES8 Don’tKnow(Vol.)9 Refused(Vol.)

DemographicInformation–ALLRESPONDENTSNow,justafewlastquestionsforstatisticalpurposesonly.We’realmostdone.Iappreciatethetimeyou’vegivenme.Q26EDUC Whatisthelastgradeorclassthatyoucompletedinschool?

[DONOTREAD;MARKCLOSEST] 1 None,orgrade1‐8 2 Highschoolincomplete(Grades9‐11) 3 Highschoolgraduate(Grade12orGEDcertificate) 4 Technical,trade,orvocationalschoolAFTERhighschool 5 Somecollege,no4‐yeardegree(includingassociatedegree) 6 Collegegraduate(B.S.,B.A.,orother4‐yeardegree) 7 Post‐graduatetrainingorprofessionalschoolingaftercollege (e.g.,towardamaster'sDegreeorPh.D.;lawormedicalschool)

8 Don’tknow(Vol.)9 Refused(Vol.)

Q27RACE Whatisyourrace?Areyouwhite,black,Asian,orsomeother? 1 White

86

2 Black 3 Asian 4 Otherormixedrace

8 Don’tknow(Vol.)9 Refused(Vol.)

Q28HISP Areyou,yourself,ofHispanicoriginordescent,suchasMexican,PuertoRican,Cuban,

orsomeotherSpanishbackground? 0 NO 1 YES

8 Don’tknow(Vol)9 Refused(Vol)

Q29MARITAL Whatisyourmaritalstatus?Areyou…[READ]

1. Married,orwithacommittedpartner 2 Divorced 3 Separated 4 Widowed 5 Neverbeenmarried

8 Don’tknow(Vol.)9 Refused(Vol.)

Q31INCOME Lastyear,thatisin2007,whatwasyourtotalfamilyincomefromallsources,before

taxes?JuststopmewhenIgettotherightcategory.[READ] 1 Lessthan$5,000

1 5tounder$10,0002 10tounder$20,0003 20tounder$30,0004 30tounder$40,0005 40tounder$50,0006 50tounder$75,0007 75tounder$100,0008 100tounder$150,0009 $150,000ormore10 Don’tknow(Vol.)11 Refused(Vol.)

87

IFQ31ISREFUSED(11)ASK:Q31AINCOME2 Justforstatisticalpurposesitwouldbereallyhelpfulifyouwouldtellmeifyourfamily

incomeisabove$20,000.Isit:[readoptions]

1 Above$20,0002 AtorBelow$20,000

8 Don’tKnow9 Refused

Q32CHILD Areyoutheparentorguardianofanychildrenunder18nowlivinginyourhousehold?

0 NO 1 YES 8 Don’tknow 9 Refused

Q33JOB Whatisyouremploymentstatus?Areyou:[READ]

1 Employedfulltime 2 Employedparttime 3 Ahomemakerorstayathomeparent 4 Retired 5 Astudent 6 Unemployed 7 Laidoff 8 Disabled 9 Don’tknow 10 Refused

Q34ZIPCODE Whatisyourzipcode?

_____ EnterZipcode

8 Don’tknow(Vol.)9 Refused(Vol.)

Q35OCCUP Whatisyouroccupation?[OPENENDED,RECORDVERBATIM]Q36

88

CHA AreyoucurrentlyaCHA[ChicagoHousingAuthority]residentorareyouaformerresidentwhowillbereturningtoCHAhousinginthefuture?

0 NO1 YES8 Don’tknow(Vol.)9 Refused(Vol.)

Q37STREETS Whatarethecross‐streetsnearestyourresidence?[OPENENDED,RECORDVERBATIM]Q38SEX [DONOTASK;ENTERRESPONDENT'SAPPARENTSEX]

1 Male2 Female

ENDOFINTERVIEW.THANKRESPONDENTGOTO[COMPLETE][COMPLETE]OK,that’sallIhaveforyourtoday.Thankyouagainforyourtime.Haveaniceday/evening.[END;complete][ATTEMPTCONVERT]Iunderstandwhyyoumightnotwanttotakethetimerightnowtotalkwithus.ButwhatwearedoingisimportanttoChicagoandyouranswerswillhelpthecitybetterunderstandwhatkindoftechnologypeopleneed.Itwillonlytakeabout12minutes.Couldyouhelpusout?

1 YESRETURNTOQ22 NOGOTO[SCHEDULE]

[SCHEDULE]Woulditbepossibletoscheduleanothertimetotalkwithyouorsomeoneelseinyourhousehold?I’dbehappytosetupaspecificdayandtimetocall.

1 YES2 NOOK,thanksforyourtime.[END;Refusal]

Great,thanks.Iamcallingyouat[readphonenumber].Whenwouldyoulikemetocallback?[ENTERDAYANDTIMEFORCALLBACK]CouldyougivemeyourfirstnamesoIknowwhotoaskforwhenIcall?[RECORDFIRSTNAME].Thanks,we’lltalktoyousoon.[END,Callbackscheduled]

89

[PARTIAL]I’msorrythisistakingsolongrightnow,andIknowyouarebusy.CouldIscheduleatimetocallyoubacktofinishthesurvey?Weonlyhaveafewmoreminutestogoandyouranswersareveryimportanttothestudysinceyou’vebeenrandomlyselectedtorepresentyourneighborhood.Woulditbepossibleforustocallyoubackatanothertimeordaytofinishthissurvey?

1 YES2 NOOK,thanksforyourtime.[END;PartialRefusal]

Great,thanks.Iamcallingyouat[readphonenumber].Whenwouldyoulikemetocallback?[ENTERDAYANDTIMEFORCALLBACK]CouldyougivemeyourfirstnamesoIknowwhotoaskforwhenIcall?[RECORDFIRSTNAME].Thanks,talktoyousoon.[END,PartialCallback][INELIGIBLE]OK,we’reonlytalkingtopeople18orovertoday.Thanksforyourtime.[END,ineligible]OTHERCODESTORECORDASNEEDEDOUTOFSAMPLE–BusinessLineDISCONNECT–NumbernotinserviceLANGUAGE–RespondentdoesnotspeakEnglishorSpanish