The document discusses how media system dependency theory can help explain why people engage in certain online activities like shopping, chatting, and reading news. It proposes that people develop dependency relationships with the Internet to meet personal and social goals. A survey found that on average people bought 8 products online in the last 6 months, spent 21 minutes daily reading online news, and 10 minutes daily chatting online. The study aims to test if these online activities can be predicted by people's overall and specific goal-oriented dependency relationships with the Internet.
The document discusses how media system dependency theory can help explain why people engage in certain online activities like shopping, chatting, and reading news. It proposes that people develop dependency relationships with the Internet to meet personal and social goals. A survey found that on average people bought 8 products online in the last 6 months, spent 21 minutes daily reading online news, and 10 minutes daily chatting online. The study aims to test if these online activities can be predicted by people's overall and specific goal-oriented dependency relationships with the Internet.
Original Description:
Internet Dependency Relations and Online Consumer Behavior
The document discusses how media system dependency theory can help explain why people engage in certain online activities like shopping, chatting, and reading news. It proposes that people develop dependency relationships with the Internet to meet personal and social goals. A survey found that on average people bought 8 products online in the last 6 months, spent 21 minutes daily reading online news, and 10 minutes daily chatting online. The study aims to test if these online activities can be predicted by people's overall and specific goal-oriented dependency relationships with the Internet.
The document discusses how media system dependency theory can help explain why people engage in certain online activities like shopping, chatting, and reading news. It proposes that people develop dependency relationships with the Internet to meet personal and social goals. A survey found that on average people bought 8 products online in the last 6 months, spent 21 minutes daily reading online news, and 10 minutes daily chatting online. The study aims to test if these online activities can be predicted by people's overall and specific goal-oriented dependency relationships with the Internet.
Journal of Interactive Advertising, Volume 3, Number 2, Spring 2003
Internet Dependency Relations and Online Consumer Behavior: A
Media System Dependency Theory Perspective on Why People Shop Chat and Read !e"s Online Padmini Patwardan Scool of !ass "ommunications #e$as #ec %niversit& Jin 'ang "ollege of !ass "ommunication ( !edia Arts Soutern Illinois %niversit& "arbondale Ta#le o$ Contents Abstract Introduction !edia S&stem )ependenc& #eor& Internet Activities *nline "atting *nline News +eading !easurement !etod ,indings )iscussion "ontributions and -imitations +eferences Appendi$ A#stract #is stud& introduces Internet )ependenc& +elations .I)+/ as a predictor of online consumer activities0 I)+ is based on te teoretical perspective of !edia S&stem )ependenc& teor&, wic postulates dependenc& relations between individuals and media based on te perceived elpfulness of media in meeting understanding .social1self/, orientation .action1interaction/ and pla& .social1solitar&/ goals0 %sing a cross2sectional email surve& of 344 respondents randoml& drawn from te facult&, staff, and student population at a large mid2western universit& in te %nited States, te predictive influence of I)+ on online sopping, catting, and news reading was empiricall& tested0 *n average, consumers in te surve& ad bougt eigt products online in te last si$ monts, spent twent&2one minutes dail& reading news online, and catted ten minutes dail& on te Internet0 #e& also displa&ed moderate, toug positive dependenc& relations wit te Internet0 I)+ significantl& e$plained online sopping activities and online news reading, but did not predict online catting0 In terms of specific I)+ goal dimensions, te predictive influence of action orientation on online sopping, solitar& pla& on online catting, and social understanding on online news reading was confirmed0 Introduction 5eter dealing wit te consumption of goods, news, or oter t&pes of online content, it as been suggested tat consumer activities in online environments indicate a more instrumental tan ritualistic use of media0 6ven more so tan an& oter medium, te Internet anticipates an active rater tan passive audience, impl&ing tat, at te present time, its use is more purposive and goal2directed0 #erefore, it is possible tat te personal and social goals tat people see7 to meet troug te Internet ma& be important motivating factors in te activities tat te& pursue online0 In tis stud& we attempt to tie goal2directed motivations of Internet users wit online sopping, catting, and news reading0 Sopping, catting, and news reading are fast proliferating activities among %0S0 users in toda&8s online environment0 In !arc 2003 alone, more tan 300 million %0S0 consumers sopped online, collectivel& spending over 930: billion .Nielsen1Net+atings and ;arris Interactive 2003/0 Similarl&, tousands of cat rooms of ever& nature report osting over a million catters dail& .Palm "oast1,lagler Internet 2000/, testif&ing to te growing popularit& of instant messaging and related cat forms .Pastore 2003/0 And recent researc from Pew Internet ( American -ife .2000/ rated online news reading as te tird most popular dail& Internet activit& in te %nited States, after sending email and surfing te 5eb for fun0 #wo teoretical approaces available to stud& ow individual goals are met troug media .including Internet/ resources are %ses and <ratifications and !edia S&stem )ependenc& teor&0 %nli7e %ses and <ratifications researc, wic is premised on consumer control over accessing media content according to teir goals1needs, we focus on consumer dependenc& on Internet resources to satisf& goals0 5e believe tat suc a dependenc& on te Internet leads, over time, to te development of a consumer2Internet dependenc& relationsip, wic, in turn, ma& li7el& affect te nature and e$tent of consumers= online activities0 In tis stud&, we propose 22 and test 22 te multidimensional construct of Internet )ependenc& +elations .I)+/ as a possible predictor of online activities0 "onceptuali>ed as te e$tent to wic people depend on te Internet to meet teir social and personal goals, I)+ is derived from !edia S&stem )ependenc& teor& .?all2+o7eac 3@A:, 3@@AB ?all2+o7eac and )e,leur 3@C4B )e,leur and ?all2 +o7eac 3@A2, 3@A@/, wic defines individual2media relations in terms of bot overall intensit& of te dependenc& relationsip, as well as te e$tent to wic individuals relate to a medium to meet specific goals0 !S) goal dimensions include understanding .self and social/, orientation .action and interaction/, and pla& .solitar& and social/ goals tat individuals see7 to meet troug media resources .,igure 3/0 %i&ure '(oal Dimensions o$ Media System Dependency Relations #e central issue tis paper addresses is te e$tent to wic Internet beaviors can be e$plained b& I)+ bot as a summed intensit&, and as te intensit& of si$ specific goal dimensions0 5e argue tat overall I)+ intensit& will significantl& influence Internet users= online sopping, news reading, and catting e$periences0 #ese activities were selected bot due to teir growing popularit& among Internet users, and teir intuitive corresponding matc wit !S) goal dimensionsD sopping wit orientation, news reading wit understanding, and catting wit pla&0 5e furter &potesi>e connections between te intensit& of specific goal dimensions and specific online activities, and e$amine predictive lin7ages between action orientation and online sopping, social understanding and online news reading, and solitar& pla& and online catting0 Media System Dependency Theory According to !S) teor&, a media dependenc& relationsip is one Ein wic te satisfaction of needs or te attainment of goals b& individuals is contingent upon te resources of te oter part&F .?all2+o7eac and )e,leur 3@C4, p0 4/0 !S) suggests tat in toda&=s societ& individuals ave to rel& on media information resources in order to attain teir various goals0 Information resources include all media products .-oges and ?all2+o7eac 3@@3/, including commercial and advertising information0 #e intensit& of media dependenc& relations depends on te perceived elpfulness of te media in meeting goals0 #e goal scope .dimensions/ of tese relations .,igure 3/ covers a wide range of individual goals 22 understanding .social and self/, orientation .interaction and action/ and pla& .social and solitar&/ 22 tat ma& be met troug media resources .-oges 3@@G/0 %nderstanding goals deal wit people=s needs to understand te world and temselvesB orientation goals focus on te need to beave effectivel& in interactions wit oters as well as in personal beavioral decisionsB and pla& goals deal wit te need for entertainment and escapism .!orton and )uc7 2000/0 5ile tese goal dimensions are e$austive, te& are not mutuall& e$clusive 22 and more tan one 7ind of goal can be activated .and satisfied/ b& te same medium .)e,leur and ?all2+o7eac 3@A@/0 ?ot intensit& and goal scope ma& be determined b& ow e$clusive media resources are perceived to be in attaining tese goals, and var& for different individuals as well as for te same individual over time .?all2+o7eac 3@A:, 3@@AB ?all2+o7eac and )e,leur 3@C4B )e,leur and ?all2 +o7eac 3@A2, 3@A@/0 Internet Activities #e incorporation of te Internet into dail& lives is reflected in te 7inds of activities man& Americans pursue online0 *n a t&pical da& in !arc 2000, :A million Americans logged on to te Internet .Pew Internet ( American -ife 2000/ to send email, surf for fun, get news, bu& a product, or cat in a cat room or a discussion forum, among oter tings0 Internet users surve&ed in a recent stud& said te Internet ad improved teir connection to famil& and friends, te wa& te& pursue obbies, and teir abilit& to learn new tings0 !an& found te Internet elpful in doing Hobs, getting information on ealt care, sopping and managing personal finances .;oward, +ainie, and Jones 2003/0 #e diversit& and intensit& of online activities point to te need to investigate wat factors migt intervene in te activities0 +ecent models of media selection and use ave suggested tat, in addition to demograpics and media attributes, factors suc as assessment of needs fulfillment, appropriateness, social norms, and peer evaluations are important in determining te nature of media use .,lanagin and !et>ger 2003/0 #erefore, from an !S) perspective one migt argue tat individual goals 22and te Internet8s abilit& to meet tem 22 ma& e$ert some influence on consumer activities in te online environment0 Online Shoppin& A few &ears ago, sopping or IpurcaseF was rated among te least prolific uses of te 5eb .Jat> and Aspden 3@@CB Poinde$ter 3@@@/0 ;owever, most mar7eters believe it is onl& a matter of time before te maHorit& of consumers sop in teir virtual storefronts0 A !arc 2003 surve& of %0S0 users found tat e2commerce as it mainstream, wit GA02K of all Americans over 3A &ears old 2230002 million people 22 aving bougt products online0 )espite downturns in te dotcom boom, consumers= online spending as steadil& increased0 In !arc 2003 alone, more tan 930: billion was spent online, a Hump of 3:04K from 9204 billion in April 2000 .Nielsen Netratings1;arris Interactive 2003/0 A 'aooL1A"Nielsen Internet "onfidence Inde$ report found tat %S consumers planned to spend at least 930 billion online between Jul&2September, 2003 ."&beratlas 2003/0 6ven so, generall& spea7ing, bu&ing online still does not appear to be one of te primar& reasons w& people visit web sites, despite te overall increase in commercial activities on te Internet0 Poinde$ter=s 3@@@ stud& found tis to be true of bot ?ab& ?oomers and <eneration Mers, even toug &oungsters spent more tan 30K of teir disposable income on purcasing diverse products troug te 5eb .,orrester +esearc +eport 2000/0 5ile <V%=s 30t 555 %ser Surve& .3@@A/ found tat Nualit& information, eas& ordering, and reliabilit& were more important to respondents tan securit&, Jorgaon7ar and 5olin .3@@@/ found tat, among oter tings, securit& concerns and transaction an$iet& appeared to be te most prevalent causes for not bu&ing on te 5eb0 Studies using demograpic variables to e$plain online sopping beavior ave often reported conflicting or confusing results0 5ile -i, Juo, and +ussell .3@@@/ found age and education level pla&ed an important role in online sopping, as did consumers= sopping orientation, ?ellman, -ose, and Jonson .3@@@/ considered demograpics an imperfect surrogate to e$plain online purcasing0 #e& found tat wile demograpics e$plained w& people were online in te first place wen compared to te national %0S0 population, te& did not significantl& predict online purcase beavior0 )ontu .3@@@/ observed tat distinction was often not made between online users and online soppers0 ;is stud& found online soppers to be older, more affluent, wit a positive attitude towards advertising and direct mar7eting, less price and brand conscious and largel& convenience see7ers0 A ,orrester +esearc +eport .3@@@/ suggested oterwiseD &ounger consumers .G0K/ bougt more freNuentl& on te Internet as compared to more mature adults .30K/B and full& 42K of all &oung %0S0 consumers were li7el& to sop online b& 20030 A variet& of studies ave also pointed out te increasing online sopping sopistication of toda&=s 34222 &ear olds, as evidenced b& teir use of various aids suc as price comparison web sites and online coupons to bu& a wide variet& of products on te Internet0 In terms of motivational variables found to influence online sopping, te 5arton Virtual #est !ar7et results reported 5ired -ifest&le22caracteri>ed b& &ears= e$perience wit te Internet, reception of large amounts of emails and wor7 on te Internet in te office ever& wee7, and #ime Starvation22a result of te increasing number of ours wor7ed b& members of a ouseold especiall& in dual2income ouseolds, as predictive of online sopping .?ellman, -ose, and Jonson 3@@@/0 5ile !edia )ependenc& +elations as been previousl& used in purcase conte$ts .<rant, <utrie, and ?all2+o7eac 3@@3B S7umanic and Jintsfater 3@@A/, it as been studied onl& in television sopping environments0 <rant, <utrie, and ?all2 +o7eac .3@@3/ modeled relationsips between viewers1bu&ers, te television sopping program, and te television medium b& e$tending !S) teor& to dependenc& on te television sopping genre0 S7umanic and Jintsfater=s .3@@A/ stud& found viewer relationsip wit te medium, te genre and te genre personae .i0e0 te tele2sopping ost/ igl& predictive of purcase beavior0 *ur researc e$amines te relationsip between Internet )ependenc& +elations and online sopping0 Since it is evident tat te online sopping e$perience involves a range of diverse activities li7e conducting product information searces, price and brand comparisons, searcing for discounts, as well as actual online product purcase, tis stud& conceptuali>es online sopping as bot a range of activities, as well as te actual number of products bougt online0 In view of te mi$ed findings related to te use of demograpics as online sopping predictors, we include demograpic variables0 Specificall&, we as7 te following researc NuestionsD R)': #o wat e$tent do intensit& of Internet )ependenc& +elations, age, gender, and income influence consumers= online sopping activitiesO R)*: #o wat e$tent do intensit& of Internet )ependenc& +elations, age, gender, and income influence te number of products bougt onlineO ?ased on definitions of goals dimensions provided in !S) teor&, we furter &potesi>e a connection between individuals8 action oriented goals and online soppingD +': Stronger action orientation goal dimension will positivel& predict consumers= overall online sopping activities0 +*: Stronger action orientation goal dimension will also positivel& predict actual online product purcase0 Online Chattin& As an eas& wa& to get instant answers to messages and to carr& on conversations wit friends, colleagues, and strangers around te globe, online catting is one of te fastest growing activities on te Internet .Pastore 3@@@/0 Almost ever& portal or online communit& on te Internet toda& osts some t&pe of cat activit&0 #e cat statistics released b& America *nline indicated tat tere were more tan G0 million registered users of its ?udd& -ist and Instant !essenger services, and more tan C:0 million dail& messages were sent troug ?udd& -ist and I"P services0 A NetValue surve& in 2003 found tat online catters as a group were among te eaviest users of te InternetB te& generated twice as man& online sessions as non2catters .Pastore 2003/0 #e stud& found tat women were more li7el& to cat online tan men, and spent two more da&s per mont on online catting tan did males .Pastore 2003/0 #e paucit& of researc on tis fast proliferating online consumer activit&, and industr&2reported evidence of its growing popularit& among Internet users, leads us to investigate te connection between Internet )ependenc& +elations and time spent catting on te Internet0 "onsidering tat demograpics provided some understanding of online catting, we as7D R),: #o wat e$tent do intensit& of Internet )ependenc& +elations, age, gender, and income influence time spent catting onlineO Again, drawing upon te goal scope of !S) teor&, we propose te following predictive relationsipD +,: Stronger solitar& pla& goal dimension will positivel& predict time spent catting online0 Online !e"s Readin& Practicall& ever& maHor mainstream newspaper or maga>ine in te %nited States is available in an online editionB te same olds true for broadcast news networ7s0 In addition, most Internet portals temselves incorporate online news services0 !illions of oter official and unofficial news2based sites are online, providing collective information resources tat appear to be virtuall& limitless0 !ar7et ,acts1!SN?" reported in 3@@A tat 2003 million %0S0 residents used te Internet as a source for news .-evins 3@@A/0 A Pew +esearc "enter biennial news consumption surve& revealed tat tere was a Hump in online news activities between 3@@423@@A, from 4K of Americans to 20K searcing for news at least once a wee70 ,or tese users, science, ealt, finance and tecnolog& were big news draws .Pew +esearc "enter for People and Press 3@@A/0 !ore recentl&, Scarboroug +esearc=s first National Internet Stud&, surve&ing more tan 2000 %0S0 adult Internet users, found tat more tan two out of five Internet users .G:K/ ad read an online newspaper in te last 30 da&s0 ;alf .::K/ ad logged on to a national newspaper web site li7e te New 'or7 #imes, 5all Street Journal, and %SA #*)A' .Scarboroug +esearc 2003/0 Scarboroug +esearc .2003/ also indicated tat, generall&, online news readers tend to be &ounger .G3K were between te ages of 3A23G/ as compared to traditional newspaper readers .onl& 23K in te same age categor&/0 ;ence, it appears tat online editions of newspapers intentionall& or unintentionall& target a new &ounger audience0 6ven toug it is too earl& to claim tat online news as entered te mainstream, Noac7 .3@@@/ argued tat te obvious advantages of Internet2based news .accessibilit&, convenience, in2dept researc, and information/ would be 7e& in attracting readers, encouraging tem to spend more time reading news online0 Along wit increased usage of online news, researc as also found an increasingl& positive attitude toward news as well0 ,or instance, among ::0 Internet users polled b& Screaming!edia, more tan alf believed tat te Internet ad te most interesting information and provided in2dept, accurate, up2to2date information .Astor 2000/0 6$amining electronic newspaper usage, 5eir .3@@@/ concluded tat media consumption was purposeful and adopters of electronic newspapers used tem to get information important to tem0 Prior researc e$ploring te connections between media dependenc& relations and news in a print newspaper conte$t indicated tat intensit& of dependenc& relations added a significant amount of e$planation in newspaper reading variance wen demograpic variables were controlledB social understanding, self understanding and action orientation were important dimensions of newspaper dependenc& relations .-oges and ?all2+o7eac 3@@3/0 In anoter stud& on television media dependenc& relations, a lin7age between dependenc& and news was also investigated and confirmed .?all2+o7eac, +o7eac, and <rube 3@AG/0 No prior studies ave e$amined te connections between online news beavior and media dependenc& relations0 ,ollowing -oges and ?all2+o7eac=s .3@@3/ suggestion to consider bot media dependenc& relations as well as demograpic factors in anal&>ing media use, we as7 te following NuestionD +PGD #o wat e$tent do intensit& of Internet )ependenc& +elations, age, gender, and income influence time spend reading news onlineO ?ased on !S)=s identification of social understanding goals as leading to individuals= information see7ing beavior, we also &potesi>e its predictive relationsip wit online news reading0 ;GD Stronger social understanding goal dimension will positivel& predict time spent reading news online0 Measurement #o measure intensit& of Internet )ependenc& +elations .I)+/, te 3A item !S) scale developed and refined b& ?all2+o7eac, +o7eac, and <rube .3@AG/, <rant, <utrie, and ?all2+o7eac .3@@3/, and ?all2+o7eac, <rant, and ;orvat .3@@:/ is used0 I)+ is tus operationali>ed as respondents= composite mean score on te 3A item !S) scale .See Appendi$/0 #o measure action orientation, solitar& pla& and social understanding, composite mean scores for tree items eac on te !S) scale were used0 ;ence, mean scores for eac of te tree goal dimensions used in tis stud& are subsets of te overall dependenc& mean score0 #e dependent variable online sopping was conceptuali>ed bot as a range of sopping2related activities tat consumers engaged in online, as well as te number of actual products te& bougt online0 *perationall&, a five2item interval scale developed b& Patwardan .2003/ was used to measure respondents= self2reported online sopping activities, wile actual purcase was measured at te ratio level as te number of products bougt online in te last si$ monts0 *nline news reading was conceptuali>ed as te e$tent to wic respondents accessed and read news on an& news2based web site, including web sites of newspapers, broadcast media, or oter news2based organi>ations0 It is operationali>ed as a ratio level measure of te amount of time spent reading news online ever& da&0 *nline catting was conceptuali>ed as respondents= use of cat and instant messaging services available on te Internet, and measured at te ratio level as te average amount of time spent dail& catting wit oters via te Internet0 )emograpic variables were measured as followsD data on age were collected b& as7ing for te &ear of birtB income was measured categoricall& .Please select te appropriate range to reflect &our annual income from all sources including salar&, and1or parental support if students0 %nemplo&ed students were advised to select teir parents= income range/B <ender was te onl& dicotomous .nominal/ variable0 Method #e stud& used a cross2sectional email surve&0 #e population of interest was students, facult&, and administrative staff at a large mid2western universit& in te %nited States0 #e sampling frame was te universit& email director&0 #is sample was particularl& desirable for tis teor& testing stud& because universit& communities are 7nown to ave a ig proportion of Internet users0 ,uture researc will build on te foundations of tis stud&, surve&ing national1international Internet user populations0 Samplin& +espondents were selected using multi2stage stratified random sampling0 Stratified random sampling is a superior metod to simple random sampling and ensures representativeness in terms of variables important to te stud&0 #e universit& population was first stratified into tree groups 22 students, facult&, and staff 22 to ensure age and income variabilit&0 6ac population group was furter stratified on te basis of gender, to ensure gender representation0 ,ive percent of students, .since students were a muc larger group tan te oters/, and 30K of facult& and 30K of staff were sampled0 #us te total sample si>e selected for te stud& was 3,G42 respondents .3,200 students 32C facult& and A2 staff/0 5ile e$ploring differences b& group was not te primar& purpose of tis stud&, tis sampling metod ensured variabilit& on demograpic factors .age, income, and gender/ important to tis stud&0 +eliabilit& anal&sis was conducted for I)+ and online sopping scales0 !ultiple +egression was used to answer +P3 troug +P30 ?ivariate regression was used to test ;&poteses 3 troug G0 Survey Administration After creating an email address list for respondents in te selected sample, te surve& Nuestionnaire was delivered via email0 "onforming to researc InetiNuette,F te surve& was accompanied b& a cover letter e$plaining purpose and nature, time reNuired to complete te surve&, and te researcers= academic affiliation0 #e option to opt out was offered, as were confidentialit& assurances0 #e letter and Nuestionnaire were sent out as inline te$t0 +espondents were reNuested to it te repl& button, respond to Nuestions, and IsendF te surve& bac7 to te researcers0 #wo mailings were done, wit some te$tual adHustments made for te second mailing to overcome problems in administration0 ,or e$ample, a respondent reported tat te message was truncated wen te repl& button was it, so te revised cover letter offered suggestions for oter return routes li7e cutting and pasting te surve& into te repl&, or sending via campus mail0 Relia#ility and -alidity Post2test reliabilities for te scales were tested using "ronbac=s alpa0 5ile scale reliabilit& for te overall 3A2item I)+ scale was fairl& ig .0AA/, it was somewat lower for eac of te si$ individual goal dimensions Q self understanding .0CG/, social understanding .04C/, action orientation .04G/, interaction orientation .0:@/, self pla& .0A:/, and social pla& .04:/0 #is was probabl& because onl& tree items were used to measure eac dimension .larger number of items generall& increases reliabilit&/0 ,or te dependent variable online sopping, reliabilit& for te five2item scale was 0@0, similar to pre2 and post2test reliabilities reported for tis scale .0@3/ in previous researc .Patwardan 2003/0 )espite an inabilit& to generali>e te results of tis stud& be&ond te population of interest .universit& facult&, students, and staff/, it ma& be argued tat universit& populations are li7el& to reflect man& of te caracteristics of Internet users in te %nited States0 #e stud& does ave ig e$ternal validit& in terms of generali>ing from te sample to te universit& population, since probabilit& sampling was used0 %indin&s A total of 3,G42 Nuestionnaires were emailed over a 302da& period0 ,our undred and eigt emails were returned as failed deliveriesB and tere were twelve refusals to participate0 R#e ig number of failed deliveries were mostl& from te student group in te sample, suggesting tat student email addresses wit te universit& are not necessaril& currentS0 ;ence for a total of 3,003 emails successfull& delivered, 3C4 responses were received, a response rate of 3C04K0 SubseNuentl&, ten incomplete .truncated/ replies were discarded, leaving 344 usable sample Nuestionnaires0 5e ac7nowledge tat low response rate is a maHor limitation, and offer two possible defenses0 ,irst, man& online and email surve&s .including te <V% Surve&s/ use non2random sampling metods since it is difficult to obtain a sampling frame of all Internet users in a particular population0 A review of te literature also suggests tat response rate for email and online surve&s is generall& muc lower tan mail or telepone surve&s0 *ur response rate falls well witin te range reported b& researcers using probabilit& sampling in email surve&s0 ;owever, li7e oter researcers conducting surve&s using tese metods, we would caution against generali>ed interpretations of our results0 As a precaution against sampling error, we conducted a cec7 for non2response bias, and found a good matc0 5e also compared te sample demograpic profile wit te population profile obtained from universit& sources0 #ere was a good matc on age for all tree groups .facult&, students, and staff/0 In terms of gender, a moderate s7ew toward female respondents was observed among student and staff respondents but no s7ew was detected in te facult& group0 ,or te facult& group, iger income was over2 represented and lower income was under2represented0 ;owever, te overall matc between te sample and te population increases confidence in te generali>abilit& of our results to te universit& population0 Demo&raphic Pro$ile In terms of demograpics, a little more tan alf te respondents .::K, n T @2/ were students, 2GK .n T G0/ were facult&, and 23K .nT 3G/ were administrative staff0 +epresentation of females was sligtl& iger .:4K/ tan males .GGK/0 *ver alf te respondents .:2K, n T CA/ were in te age group of 3AQ3G, G0K .n T :@/ were between 3:Q:G &ears old, and onl& eigt percent .n T 32/ were more tan :: &ears old0 In terms of income distribution, over 43K of respondents .n T A@/ ad an annual income below 9:0,000D of tese, alf .n T G:/ earned below 92:,000 and alf .n T GG/ above0 #went& eigt percent of te respondents .n T G3/ earned between 9:0,000 to 9300,000 annuall&, and a smaller number .30K, n T 3:/ earned more tan 9300,0000 Mean IDR and Online Behavioral Activities #e mean intensit& of overall Internet )ependenc& +elations among respondents suggested a positive but somewat restrained dependence on te Internet=s resources to satisf& individual goals .mean T 303 on a scale of 3 to :, wit iger score indicating greater dependenc&/0 Similarl& positive but moderatel& intense dependenc& for understanding .mean T 300/, orientation .mean T 302/ and pla& .mean T 303/ were also observed .#able 3/0 Interestingl&, te igest means were found among te tree specific goal sub2dimensions used in tis stud&D action orientation .mean T 304/, solitar& pla& .mean T 303/ and social understanding .mean T 304/, wen compared to oter social and self dimensions in te !S) goal scope0 #able 3!ean Internet )ependenc& +elations and *nline ?eavioral Activities In terms of Internet2based activities, most respondents in te surve& engaged in online sopping activities fairl& freNuentl& on te Internet .mean T 202 on a scale of 3 to : running from Iver& freNuentl&F to IneverF/, and ad bougt an average of eigt products online in te last si$ monts0 5ile tere were no significant differences b& group in te use of te Internet for online sopping activities in general, differences were observed in te number of products bougt online b& staff .3: products in si$ monts/ and students .4 products in si$ monts/ ., T G03, p T 002/ .#able 3/0 +espondents also spent about 30 minutes dail& catting online0 Students and facult& differed significantl& in te time te& spent catting online ., T 304, p T 003/0 *n average, students spent te most time catting online dail& .mean T 3: minutes/, followed b& staff .mean T 4 minutes/, and facult& .mean T 3 minutes/ .#able 3/0 *n average respondents spent about 23 minutes dail& reading news online, and no statisticall& significant differences b& group were observed in time spent reading news online, toug facult& spent te most time on tis dail& activit& .mean T 24 minutes/, followed b& students .mean T 20 minutes/ and staff .mean T 3A minutes/0 R)': IDR and Online Shoppin& +P3 investigated te e$tent to wic intensit& of Internet )ependenc& +elations, and demograpic factors, affected consumers= online sopping activities0 -inear multiple regression was used to cec7 te relationsips0 #e overall model found tat eigt percent of variance in online sopping was e$plained b& I)+ and demograpic variables .+ sNuare T A03, , T 20C:, p T 003/ .#able 2/0 ;owever, none of te demograpic factors .age, gender, income/ were significant predictors0 I)+ was te onl& factor tat significantl& e$plained almost all te variance in online sopping .AK, t T 3022, p T 000/0 Ta#le * Multiple Re&ression Analysis o$ A&e (ender Income and Internet Dependency Relations to Predict -ariance in Online Shoppin& R)*: IDR and !um#er o$ Products Bou&ht Online +P2 e$amined te e$tent to wic te intensit& of Internet )ependenc& +elations, and demograpic factors, affected number of products bougt online0 -inear multiple regression was used to test te relationsips0 It was found tat te regression model did not significantl& predict te number of products bougt online0 +': Action Orientation (oal Dimension and Online Shoppin& ;3 e$pressed a relationsip between action orientation goal dimension and te online sopping e$perience0 ?ivariate linear regression was used to test te &potesi>ed relationsip0 Action orientation was found to be a strong, significant predictor .+ sNuare T 024, , T :G00:, p T 000/ .#able 3/0 ;3 was, terefore, supported0 Ta#le , .inear Re&ression /sin& Speci$ic IDR Dimensions to Predict -ariance in Online !e"s Readin& Online Chattin& Online Shoppin& and !um#er o$ Products Bou&ht Online +*: Action Orientation (oal Dimension and !um#er o$ Products Purchased Online ;2 as7ed weter action orientation goal dimension also predicted te number of products bougt online0 Surprisingl&, despite moderate significant correlation between online sopping activities as measured on te scale, and te number of products bougt online .r T 0G@3, p T 002, one tailed/, te predictive relationsip between action orientation and te actual number of products bougt online was not significant .+ sNuare T 002, , T 303A, p T 00C/ .#able 3/0 ;2 was, terefore, not supported0 R),: IDR and Time Spent Chattin& Online +P3 investigated te e$tent to wic I)+, and demograpic factors, influenced time spent catting online0 #e overall multiple regression model e$plained 3AK of variance in online catting .+ SNuare T 03AB , T 40:C, p T 000/ .#able G/0 ;owever, te variance in time spent catting online was not e$plained b& I)+, but b& demograpic variables .age and income/0 *f te total amount of variance in online catting, age ad a uniNue contribution of si$ percent, income e$plained 33K, and te rest was sared0 Ta#le 0 Multiple Re&ression Analysis o$ A&e (ender Income and Internet Dependency Relations to Predict -ariance in Time Spent Chattin& Online +,: Solitary Play (oal Dimension and Time Spent Chattin& Online ;3 investigated te predictive relationsip between solitar& pla& and time spent catting online0 ?ivariate regression anal&sis found tat solitar& pla& significantl& predicted te time spent catting online .+ sNuare T 00:, , T C04, p T 000/ .#able 3/0 ;&potesis 3 was, terefore, supported0 ;owever, onl& five percent of variance in te dependent variable was e$plained b& solitar& pla&, suggesting tat te predictive relationsip between te two was not ver& strong0 R)0: Demo&raphics IDR and Time Spent Readin& !e"s Online +PG e$amined te e$tent to wic I)+ intensit&, and demograpic variables, affected te time spent reading news online0 #e linear multiple regression model accounted for 3@K of variance in te dependent variable .+ sNuare T 03@, , T 404@, p T 000/ .#able :/0 Interestingl&, te two significant predictors were gender .# T 2G03:, p T 000/ wic e$plained 33K of variance in te dependent variable, and Internet )ependenc& +elations .# T 20A:, p T 000/ wic e$plained five percent of variance0 #e negative relationsip in te case of gender suggested tat male Internet users were more li7el& to read news online tan female Internet users0 Ta#le 1 Multiple Re&ression Analysis o$ A&e (ender Income and Internet Dependency Relations to Predict -ariance in Time Spent on Online !e"s Readin& +0: Social /nderstandin& (oal Dimension and Time Spent Readin& !e"s Online ;G &potesi>ed a predictive relationsip between social understanding and time spent reading news online0 ?ivariate regression anal&sis indicated tat social understanding was significant in predicting time spent reading news online .+ sNuare T 00A, , T 3303@, p T 000/ .#able 3/0 ;ence, ;G was supported0 Discussion As te fastest growing communication medium of all times, te Internet is not onl& canging people=s personal lifest&les but also resaping te interdependence between individuals, media, and societ&0 )ependenc& is te flip side of control0 As we argue for greater consumer empowerment and control over wat media content we consume in Internet environments, we are also more li7el& to grow increasingl& dependent on its resources to meet our goals0 In terms of individual2media relationsips tat develop over time, our stud& suggests tenable connections between individual goals and dependenc& on Internet resources0 *n average, Internet users did displa& moderatel& intense Internet )ependenc& +elations, indicating tat te medium as become an integral part of individuals= media environments0 I)+ intensit& appears to be strongest among &ounger people0 In te case of different I)+ goal dimensions, students, more tan facult& or staff, appear to be more strongl& motivated to see7 out Internet resources to meet teir overall pla& Q and solitar& pla& Q goals, empasi>ing te entertainment value of media to te &ounger generation0 #is stud& also finds support for previous researc attesting to te growing popularit& of online sopping, catting, and news reading activities among Internet users0 "onsumers in our stud& ad bougt an average of eigt products online in te last si$ monts, spent at least tirt& minutes per da& reading news online, and catted ten minutes dail& on te Internet0 )ifferences b& group were, owever, evident in te fact tat facult& spent te most time reading news online, students spent te most time catting online, and staff did te most sopping online0 *ur stud& also focused on te e$tent to wic Internet )ependenc& +elations influenced online sopping, catting, and news reading0 At tis stage of te Internet=s development, I)+ appears to be a moderate determinant of beavioral responses0 In te case of online sopping, te stud& is consistent wit previous researc findings tat suggest demograpic variables are not significant in e$plaining online sopping variance0 ;owever, statistical significance alone is not sufficient to draw conclusions about te predictive strengt of I)+, considering te low + sNuare, and future replications are necessar& to investigate te impact of I)+ intensit& on online beaviors, considering te criticalit& of te Internet2user interface in te commercial world0 5e also &potesi>ed a predictive lin7 between specific goals and online activities0 6$amining te connection between action orientation and online sopping, we found tat individuals wo depended on te Internet to meet teir action orientation goals were also more li7el& to engage in sopping2related activities online0 #is suggests tat greater consumer dependence on Internet resources to elp ma7e personal beavioral decisions .action orientation/ does indeed influence te online sopping e$perience0 ;owever, action orientation did not predict te number of products actuall& bougt online0 A possible e$planation migt be provided b& te differences between groups in te number of products bougt online0 Since staff bougt te most products online as compared to students and facult&, it suggests to us tat man& of te purcases were wor72related0 If te above conHecture is correct, it is possible tat Internet use in wor72related sopping conte$ts ma& differ from use in personal sopping conte$ts, and we ma& argue tat media dependenc& relations, based on te satisfaction of individual personal goals, ma& not influence purcase beavior in te wor7place0 In future researc, clearer distinction sould be made between wor72 related and personal online sopping0 Puestions related to te 7inds of products bougt online ma& also be included0 5ile demograpics did not affect online sopping, supporting previous researc findings, tis stud& indicated tat te& still ave potential to predict oter t&pes of Internet use0 Age and income are important predictors in online catting at presentB and it appears tat online catting is an activit& tat &ounger people wit associated lower incomes engage in for longer periods of time tan oters0 #e significant correlation of age and pla& goals also ma7es sense in te ligt of te greater intensit& of te overall pla& 22 as well as te more specific solitar& pla& 22 dimension among &ounger people0 ;ence &ounger people, wo are more dependent on te Internet to meet teir pla& goals, were also te ones more li7el& to cat online for longer periods of time0 I)+ was a significant predictor of te amount of time spent reading news online0 A significant gender difference was also observed, wit males spending more time tan females on tis activit&0 #e strong predictive correlation between social understanding and online news reading indicated tat people do depend on te Internet=s information resources to understand te world around tem0 Previous anal&sis of newspaper readersip and dependenc& relations teori>ed tat social understanding was lin7ed to newspaper reading because a reader=s goals of increasing integration in te communit& were addressed b& newspaper content .-oges and ?all2+o7eac 3@@3/0 #e same appears to old true in te Internet2based news environment as well0 Contri#utions and .imitations ?& introducing I)+ and its goal dimensions as a possible source of variance in online consumer beavior, we ope discussion of its importance and relevance will be furter stimulated0 ?ecause of its relational aspect, I)+ is potentiall& a better measure tan a simple Nuantification of te e$tent of Internet use0 #e Internet itself is inerentl& more consumer2involving, increasing te li7eliood of developing a relationsip wit it, wic in turn is li7el& to influence te nature and e$tent of online activities0 At present, tis relationsip appears to be of moderate intensit&, but we believe it will strengten over time0 ;owever, te findings in tis stud& are e$plorator&, and need to be furter validated troug future researc wit more general populations0 *ur researc as some implications for industr& as well0 62commerce companies, for e$ample, are strongl& motivated to discover reasons tat drive soppers online0 Internet portal companies are an$ious to increase web site traffic b& uncovering motivations tat lead people to use cat and instant messaging features0 And online newspapers and news web sites are 7een to understand ow news readers1viewers can be attracted to content on teir web sites0 #oug variance in online sopping and news reading e$plained b& I)+ was small, and I)+ did not e$plain variance in online catting, significant lin7ages between specific I)+ goal dimensions and online activities were observed0 #erefore, it is suggested tat online purcase action could be made easier, convenient, and action2oriented to serve online soppers betterB facilitating understanding goals could be te strategic focus to serve online news readersB and cat sites can increase traffic b& focusing on meeting pla& goals b& ma7ing sites fun and entertaining to use .for e$ample, te use of emoticons, view cams, and oter devices to ma7e te online catting e$perience multi2dimensional/0 #is stud& as some limitations0 6mail surve&s generall& result in lower response rate tan tose of telepone or mail surve&s0 6$ploring Saeffer and )illman=s .3@@A/ suggestion of using a multi2metod approac .combining email wit oter surve&ing metods li7e mail surve&s, for e$ample/ and initiation of multiple contacts .tis stud& used Hust two mailings/ to improve response rate, ma& provide some solutions in te future0 Second, te speed of tecnological advances constantl& alters te nature and scope of Internet activitiesB tis ma& in turn alter te nature and scope of dependenc& relations as well0 ;ence, trac7ing relations troug longitudinal anal&sis ma& provide a more consistent understanding of te development of individual2Internet relations over time tan te cross2sectional approac adopted in tis stud&0 #e use of te !S) teoretical perspective in tis researc ma& also invite some criticism, due to its limited use in media effects researc0 In our considered opinion, despite its comple$ conceptuali>ation, !S) provides a compreensive and organi>ed conceptual framewor7 to e$plore individual2media relations0 In terms of operationali>ation, !S) measurement allows cross2media as well as cross2genre comparisons, ma7ing it a strong and stable measuring instrument in media anal&sis0 ,uture researc could e$amine not Hust overall Internet dependenc&, but also dependenc& on specific t&pes of Internet content, for e$ample online advertising, political information, commercial information, or ealt1medical information0 +eplication wit national and international Internet user populations could provide anoter perspective on te development of Internet )ependenc& +elations and its effects on online consumer beavior0 "omparative studies of dependenc& on Internet and oter media, and te e$tent to wic te Internet is1is not affecting dependenc& on oter media or information sources also offer e$citing possibilities for future investigation0 Re$erences Astor, )avid .2000, !a& 3:/, ISurve& ,inds !ore Net %se and #rust, 6ditor ( Publiser, 3:3, 3:0 ?all2+o7eac, Sandra J0 .3@A:/, I#e *rigins of Individual !edia S&stem )ependenc&D A Sociological ,ramewor7, "ommunication +esearc, 32 .G/, GA:2:300 UUU .3@@A/, IA #eor& of !edia Power and a #eor& of !edia %seD )ifferent Stories, Puestions, and 5a&s of #in7ing, !ass "ommunication ( Societ&, 3 .312/, :2 G00 UUU and !elvin A0 )e,leur.3@C4/, IA )ependenc& !odel of !ass !edia 6ffects,F "ommunication +esearc, 3, 32230 ?all2+o7eac, Sandra J0, August <rant, and A0 ;orvat .3@@:/, IA Scale for !easuring !edia )ependenc&,F #&pescript, -os Angeles, "AD Annenberg Scool of "ommunication0 ?all2+o7eac, Sandra J0, !ilton +o7eac, and Joel 50 <rube .3@AG/, #e <reat American Values #estD Influencing ?eavior and ?elief #roug #elevision, N'D ,ree Press0 ?ellman, Steven, <erald -0 -ose, and 6ric J0 Jonson .3@@@/, IPredictors of *nline ?u&ing ?eavior, "ommunications of te A"!, G2 .32/, 3223A0 "&beratlas .2003, June 2C/, I'aoo, A" Nielsen )eclare 62"ommerce Strong and ;ealt& VttpD11c&beratlas0internet0com1mar7ets1retailing1article10,,4043WC@2423,000tmlX )e,leur, !elvin -0 and Sandra J0 ?all2+o7eac .3@A2/, #eories of !ass "ommunication0 .Gt ed0/, N'D -ongman0 UUU and UUU .3@A@/, #eories of !ass "ommunication0 .:t ed0/, N'D -ongman0 )ontu, Navin .3@@@/, I#e Internet Sopper,F Journal of Advertising +esearc, 3@ .3/, :22:A0 ,lanagin, Andrew J0 and !iriam J0 !et>ger .2003/, IInternet %se in te "ontemporar& !edia 6nvironment, ;uman "ommunication +esearc, 23 .3/, 3:32 3A30 ,orrester +esearc +eport .2000, June/, I#e !anufacturer <rowt Spiral,F Press ?rief issued 4124120000 Available on reNuest from pressYforrester0com0 UUU .3@@@, November/, ?rand )oesn=t )rive 'oung "onsumers *nline VttpD11www0forrester0com16+1Press1+elease10,3C4@,3A4,,,0tmlX0 <rant, August 60, J0 Jendall <utrie, and Sandra J0 ?all2+o7eac .3@@3/, I#elevision SoppingD !edia S&stem )ependenc& Perspective, I"ommunication +esearc, 3A .4/, CC32C@A0 <V% 30t 555 %ser Surve& .3@@A/, +eport VttpD11www0cc0gatec0edu1gvu1userWsurve&s1surve&23@@A2301tentreport0tmlZe$X ;oward, Pilip 60 N0, -ee +ainie, and Steve Jones .2003/, I)a&s and Nigts on te Internet, American ?eavioral Scientist, G:, 32330 Jat>, James and Pilip Aspden .3@@C/, I!otives, ;urdles and )ropouts,F "ommunications of te A"!, G0 .G/, @C23030 Jorgaon7ar, Pradeep J0 and -ori )0 5olin .3@@@/, IA !ultivariate Anal&sis of 5eb %sage,F Journal of Advertising +esearc, .!arc2April/, :324A0 -evins, ;oag .3@@A/, I<rowing %0S0 Audience +eads News on Net,F 6ditor ( Publiser, ,eb0 23, 3@@A, 3G0 -i, ;airong, "eng Juo, and !arta <0 +ussell .3@@@/, I#e Impact of Perceived "annel %tilities, Sopping *rientations, and )emograpics on te "onsumer=s *nline ?u&ing ?eavior, Journal of "omputer !ediated "ommunication, : .2/ VttpD11www0ascusc0org1Hcmc1vol:1issue21airong0tmlX0 -oges, 5illiam 60 .3@@G/, I"anaries in te "oal !ineD Perceptions of #reat and !edia S&stem )ependenc& +elations, "ommunication +esearc, 23 .3/, :2230 UUU and Sandra J0 ?all2+o7eac .3@@3/, I)ependenc& +elations and Newspaper +eadersip,F Journalism Puarterl&, C0 .3/, 402243G0 !orton, #omas A0 and Julie !0 )uc7 .2000/, ISocial Identit& and !edia )ependenc& in te <a& "ommunit&D #e Prediction of Safe Se$ Attitudes,F "ommunication +esearc, 2C .G/, G3A2G400 Nielsen Netratings1;arris Interactive .2003, April 2G/, INearl& ;alf of all Americans ?u& *nline, according to Nielsen1Net+atings and ;arris Interactive,F VttpD11www0arrisinteractive0com1news1allnewsb&date0aspONewsI)T2C3X0 Noac7, )avid .3@@@, Januar& 34/, IPoll sa&s 5eb News %se in !ainstream,F 6ditor ( Publiser, 332 .3/, 240 Palm "oast1,lagler Internet 2000, I"atting on te Internet,F VttpD11www0pcfl0net1cannels1cat1default0tmX0 Pastore, !icael .3@@@, Jul& 3:/, IInternet %sers #a7ing to "at,F "&berAtlas VttpD11c&beratlas0internet0com1bigWpicture1trafficWpatterns1print10,,:@33W342:X UUU .2003, ,ebruar& C/, I*nline "atters among ;eaviest 5eb %sers,F "&berAtlas VttpD11c&beratlas0internet0com1bigWpicture1trafficWpatterns1print10,,:@33W:A2G@3,000t mlX Patwardan, Padmini .2003/, I)o Purcasing Involvement, #ecnolog& +elationsip, and Visual1verbal *rientation Predict "onsumer *pinion about and actual use of te Internet for Product Information Searces and *nline SoppingOF, Presented to te American Academ& of Advertising Annual "onference, !arc 2@2April 3, Salt -a7e "it&, %ta0 Pew Internet ( American -ife0 .2000/, I)ail& Internet Activities,F VttpD11www0pewinternet0org1reports1cart0aspOimgT4Wdail&Wactivities0HpgX0 Pew +esearc "enter for People and Press .3@@A/, IPew +esearc "enter 3@@A ?iennial News "onsumption Surve&D 6vent )riven News Audience,F VttpD11www0people2press0org1med@Arpt0tmX0 Poinde$ter, Paula !0 .3@@@/, IMers and ?oomersD Are te& tat different in teir +elationsip to 5eb AdvertisingO In Proceedings of te 3@@@ "onference of te American Academ& of Advertising, !0S0 +oberts, ed0, <ainsville, ,loridaD %niversit& of ,lorida0 Scarboroug +esearc .2003, !a& @/, I,irst Scarboroug National Internet Stud& +eveals "anges in ow *nline "onsumers use #raditional and Internet !edia,F VttpD11www0scarboroug0com1scarb20001press1prWinternetstud&30tmX0 Scaeffer, )avid +0 and )on A0 )illman .3@@A/, I)evelopment of a Standard 6mail !etodolog&,F Public *pinion Puarterl&, 42, 3CA23@C0 S7umanic, Stepanie A0 and Jintsfater, )avid P0 .3@@A/, IIndividual !edia )ependenc& +elations witin #elevision Sopping ProgrammingD A "ausal !odel +evisited and +evised,F "ommunication +esearc, 2: .2/, 200223@0 5eir, #om .3@@@/, IInnovators or News ;oundsO A Stud& of 6arl& Adopters of te 6lectronic Newspapers,F Newspaper +esearc Journal, 20 .G/, 422A30 Appendi2 Scales Individual Media Dependency Relations Scale .,ive point scale from not at all elpful to ver& elpful/ In &our dail& life, ow useful is te Internet to[\ Sel$ understandin& <ain insigt into w& &ou do some of te tings &ou do Imagine wat &ou8ll be li7e wen &ou grow older *bserve ow oters cope wit problems or situations li7e &ours Social understandin& Sta& on top of wat is appening in te communit& ,ind out ow te countr& is doing Jeep up wit world events Action orientation )ecide were to go for services suc as ealt, financial, or ouseold ,igure out wat to bu& Plan were to go for evening and wee7end activities Interaction orientation )iscover better wa&s to communicate wit oters #in7 about ow to act wit friends, relatives, or people &ou wor7 wit <et ideas about ow to approac oters in important or difficult situations Solitary play %nwind after a ard da& or wee7 +ela$ wen &ou are b& &ourself ;ave someting to do wen nobod& else is around Social play <ive &ou someting to do wit &our friends ;ave fun wit famil& or friends ?e a part of events &ou enHo& witout aving to be tere Online Shoppin& Scale .,ive point scale from Ver& ,reNuentl& to Never/ I sop on te Internet I bu& man& different products on te Internet I ma7e use of online discounts on goods and services I follow up on good deals on te Internet I bu& a product online even if oter bu&ing options are available A#out the Authors Padmini Pat"ardhan is an Assistant Professor in Scool of !ass "ommunications at #e$as #ec %niversit&0 3in 4an& is a )octoral Student in "ollege of !ass "ommunication ( !edia Arts at Soutern Illinois %niversit& "arbondale0 ]#is is an invited article %+-D Hiad0org1vol31no21patwardan "op&rigt ^ 2003 Journal of Interactive Advertising