A knowledge structures exploration on social network sites

Kybe ne es A knowledge s uc u es explo a ion on social ne wo k si es Jou nal: Kybe ne es Manusc ip ID K-01-2016-0013.R2 Manusc ip Type: Resea ch Pape Keywo ds: Pe cei ed Communi y Suppo , Rela ionship Quali y, Technology Accep ance Model, Social Ne wo k Si es, Pa h inde Ne wo k Scaling Kybe ne es Kybe ne es A knowledge s uc u es explo a ion on social ne wo k si es Key Wo ds: Pe cei ed Communi y Suppo , Rela ionship Quali y, Technology Accep ance Model, Enjoymen , Anxie y, Sel -E icacy, Pa h inde Ne wo k Scaling. JEL Classi ica ion: M15, M39. Page 1 o 15 Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es 1. In oduc ion and s udy pu pose As he g ow h and usage le el o Social Ne wo k Si es (he eina e , SNS) become a global phenomenon and o ganisa ions claim he ele ance o endu ing social bonds wi h espec o indi iduals’ decision-making p ocesses (A enas-Gai an e al., 2013; Wang & Chang, 2013), he ocus o ma ke ing esea ch shi s om en icing indi iduals o adop SNS o ensu ing high- alue ela ionships wi h o he s. “In he las ew yea s, social media has become an impo an sou ce o knowledge and enables he c ea ion o he con en alue chain” (Roblek e al., 2013, p.558). In his ein, he in angible cues o online social se ices imply he de elopmen o a ue ela ionship-based app oach and, consequen ly, g owing calls o esea ch which helps us o unde s and how use s’ ue-loyal y is o med in SNS. Ou s udy i s ly designs a esea ch p oposal o aid in explo ing pe cei ed communi y suppo (he eina e , PCS) based on (1) he de elopmen o bene i s de i ed om he p e e en ial ea men and coope a ion be ween indi iduals and g oups, and (2) high le els o ela ionship quali y (he eina e , RQ) e lec ed in sa is ac ion and us (e.g., C osby e al., 1990; Hennig-Thu au & Klee, 1997; Hennig-Thu au e al., 2002). On he one hand, PCS is made up o eelings o being suppo ed by o he s in he online communi y, ac i e in e ac ions wi h i (and i s membe ship g oups), and s onge bonds o close membe s. The exchange o mu ual suppo , in eg a ion and ul ilmen o needs a e essen ial mo i es o he long-las ing exis ence o SNS (Wellman & Guilia, 1999). Indi iduals employ online social se ices because hey a e connec ed in a pe son- o-pe son manne which is mo e di ec and in e pe sonal, pe cei e he possibili y o ob aining esou ces based on u ili y and enjoymen om hem, and expec o g a i y social-emo ional needs (Rau e al., 2008). On he o he hand, he success o SNS is also based on RQ ha is conside ed as a o emos ma ke ing concep o p edic he enhancing o elec onic loyal y, and de ined as “ he deg ee o app op ia eness o a ela ionship o ul ill he needs o he cus ome associa ed wi h ha ela ionship” (Hennig-Thu au & Klee, 1997, p.751). The p esen s udy ex ends p e ious esea ch on SNS by using a ma ke ing amewo k - he RQ model (e.g. C osby e al., 1990; Dwye e al., 1987, among o he s) - ocusing on he ue commi men phase o he beha iou chain model (c . Fogg & Eckles, 2007). Secondly, o e he yea s schola s ha e become inc easingly awa e o he ele ance o he non-cogni i e aspec s (e.g., how enjoyable he SNS would be) in unde s anding a i udes owa ds use o se ices (e.g., Bagozzi e al., 1999). A main pu pose o his s udy is hus o e eal how he in insic d i e s a e associa ed wi h a ue ela ionship-based app oach. On he one hand, hedonic mo i a ion is concep ualised as he pleasu e he indi idual eels objec i ely when commi ing a pa icula beha iou o ca ying ou a pa icula ac i i y (Moon and Kim, 2001). When his occu s, an indi idual de i es in insic enjoymen om he usage o he sys em and ends o con inue wi h i (Igba ia e al., 1996). On he o he hand, in o de o expe ience in insic enjoymen while engaged in online social in e ac ions, membe s mus pe cei e a balance be ween hei compe ences and he challenges o he ac i i y. Challenges mus be abo e a c i ical h eshold, and sel - e icacy is assumed as a belie o one’s capabili y o use he SNS (c . Wood & Bandu a, 1989). O he wise, use s may become anxious (c . compu e anxie y as discom o o ea when using he SNS o when hinking abou using he SNS). In his si ua ion o de elopmen , a p oposal based on Technology Accep ance Model (he eina e , TAM; c . Da is, 1989; Da is e al., 1992) and low (essen ially de ined as an in insically enjoyable expe ience; Sánchez-F anco & Roldán, 2005), is designed o desc ibe he main mo i es ha , i s , a ec SNS accep ance and usage and second, make using he SNS a compelling cus ome -expe ience. Flow is a posi i e, highly-enjoyable s a e o consciousness ha occu s when ou pe cei ed skills ma ch he pe cei ed challenges we a e unde aking. When his occu s, an indi idual de i es in insic enjoymen om he ac i i y and ends o con inue wi h i . Thi dly, assuming ha he s udy o hedonic-mo i a ion sys ems such as SNS is s ill a he ini ial s age, ou esea ch also p oposes ha he success o an SNS is associa ed pa ially wi h i s unc ionali y and u ili y (e.g., seeking in o ma ion). In his sense, he u ili a ian na u e o SNS is s ill unclea (E ns e al., 2013). As Cons an inides e al. (2013) no e, he e is no much esea ch abou he use o TAM o explain he accep ance o social echnologies (c . also Shin 2008). TAM is a p ominen heo e ical model in he ield o in o ma ion sys ems, “bu i has ecei ed much less a en ion and app ecia ion in communica ion esea ch” (Wang e al., 2012). TAM based on he Theo y o Reasoned Ac ion (Fishbein and Ajzen, 1975) and he Theo y o Planned Beha iou (Ajzen, 1991; Ajzen & Fishbein, 1980) is he e u ilised o explain which d i e s a e essen ial o os e communi y pa icipa ion in an SNS. TAM has, as he heo y mos employed in adop ion and accep ance esea ch (Venka esh & Bala, 2008), p o en o be a use ul heo e ical model in helping o unde s and and explain use beha iou in IS domain, and i has p o en o be pa simonious and obus (Pa lou, 2003; Sánchez-F anco, 2006). Fou hly, esea ch shows ha indi idual di e ences in luence ue ela ionship-based app oaches. Di e ences a ising om pe sonal ai s a e highly ele an o alidly p edic ing beha iou al choices o se ices because his a ia ion is a he hea o he cus ome decision-making p ocess (Holb ook & Hi schman, 1982; Roeh ich, 2002). Al hough he s udy o online social en i onmen s became an ea ly ocal Page 2 o 15Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es poin o esea che s, “ hey know li le abou cha ac e is ics o indi idual use s who spa k communica ion o in luence i s con en ” (Hu ake , 2010, p.593). Ou esea ch in pa icula examines he in e ac ing ole o accumula ion o knowledge (e.g., belie s and opinions) de i ed om se ice encoun e s (e.g., Ge en, 2000; Lee & Kwon, 2011; Ng, 2013). Following Casaló e al. (2008), he in ensi y o he in e ela ions o ou esea ch p oposal could change o e ime; i.e., loyal y an eceden s migh a y acco ding o he membe s’ expe ience. The e o e, as he benefi s o SNS migh be e alua ed di e en ly by new and long- e m membe s, ou esea ch sugges s examining he ela ional app oach by le els o use amilia i y. In eg a ing amilia i y – desc ibed as “an unde s anding based on p e ious in e ac ions, expe iences, and lea ning o wha , why, whe e and when o he s do wha hey do” (Ge en, 2000, p. 727) – b ings abou a iche unde s anding o knowledge s uc u es loca ed in he memo y o young use s o SNS. Familia i y (as a dynamic, si ua ion-speci ic indi idual di e ence) is indeed conside ed o be one o he main d i e s o indi iduals’ beha iou , and i is he e de ined as “an indi idual’s knowledge ega ding o he communi y membe s and hei ac i i ies wi hin he communi y” (Shen e al., 2010, p. 52). To sum up, ou esea ch examines he mode a ing ole o accumula ion o knowledge de i ed om se ice encoun e s du ing p e ious success ul in e ac ions (e.g. Ge en 2000; Lee & Kwon 2011; Ng 2013, among o he s). Gi en ha cus ome amilia i y is a signi ican p econdi ion o alidly p edic ing beha iou al choices o echnologies and se ices, communi y manage s could e ec i ely enhance cus ome s’ PCS by a ge ing hei unde lying amilia i y. Finally, he complexi y o indi iduals’ knowledge s uc u es based on le els o amilia i y is measu ed using a Pa h inde ne wo k algo i hm. The Pa h inde echnique is e y e ec i e and e icien o he display o complex ela ionships be ween objec s such as sophis ica ed ne wo ks, and i is well es ablished in he cogni i e psychology li e a u e (Fesel, 2015). Ou s udy p ecisely in es iga es belie s-s uc u es o amilia ised and non- amilia ised use s - ocusing on how use s o ganise concep s and a ibu es in memo y. As a use becomes mo e knowledgeable in a domain, he/she begins o o ganise his/he belie s abou PCS, RQ and ex ended TAM by cons uc ing meaning ul ela ionships be ween belie s and opinions a he han me ely accumula ing decla a i e knowledge. Nex , ou s udy desc ibes he esea ch me hod ha is adop ed o analyse he esea ch p oposal. The esul s and analysis ollow he esea ch design. This pape also ou lines he limi a ions and implica ions o u u e esea ch. 2. Resea ch ques ions and Me hod The p esen s udy is designed o examine knowledge s uc u es loca ed in he memo y o young use s o SNS. In pa icula , ou p esen esea ch demons a es a me hod o combining opinions and belie s abou PCS, RQ and ex ended TAM in o one ne wo k ep esen a ion in he con ex o SNS, and also analyses indi idual di e ences based on use amilia i y. In doing so, i a emp s o add ess he ollowing esea ch ques ions: Wha o m do he ne wo k ep esen a ions be ween PCS, RQ and TAM ake in online social en i onmen s? In pa icula , would he associa ions be ween he abo e dimensions be mo e di ec amongs highly amilia ised indi iduals? When use s log in o an SNS, hey exhibi somewha less logical beha iou as hey a e p obably explo ing he sys em. The ini ial beha iou is qui e di e en om mo e amilia ised beha iou s and may mani es i sel in di e en ep esen a ions. 2.1. Pa icipan s In Spain, 79 pe cen o In e ne use s (unde age 55) ha e a p o ile on an SNS. Among young adul s, his pe cen is highe . Resea ch ca ied ou wi h young unde g adua e s uden s concluded ha 91.2% pa icipa e in SNS (c . Gómez e al., 2012). 53% connec se e al imes a day and basically ag ee ha he usage o SNS is one o hei daily asks. Ou pa icipan s a e ec ui ed du ing he au umn semes e 2014 by eques s dis ibu ed in a a ie y o social si es (i.e., Facebook, Wha sApp, Linkedin, e c). The su ey pe iod is 4 weeks. Indi iduals a e asked o indica e he SNS ha hey use. The esul s show ha 85% young esponden s use Facebook (i.e., 289 indi iduals). Facebook, as he mos popula SNS (1.2 billion people a ound he wo ld). Ou esea ch hus chooses Facebook as he SNS o he p esen s udy. These a e 54% male and 46% emale. The age ange is om 18 o 34 yea s. The e a e se e al speci ic ad an ages o su eying homogeneous samples. On he one hand, he esea ch educes he mode a ing in luences o hi d-cons uc s. Rep esen a i e sampling o subjec s is no necessa y when conduc ing heo e ical use esea ch. Sampling a ba ely homogeneous g oup o subjec s also in la es he e o in s a is ical es s and educes he likelihood o de ec ing sys ema ic iola ions o a heo y when i is alse. In his ein, ou esea ch conside s a homogeneous sample o ensu e inc easing he in e nal alidi y o ou indings (Calde e al., 1981). On he o he hand, young people a e ele an use s o SNS in compa ison o he gene al popula ion. They use i o communica e wi h hei pee s (e.g., Bone a e al., 2006). Eme ging adul hood is de ined as “a s a e o explo a ion and ins abili y” (A ne , 2004, p. 21) which de ines hei iden i y and in imacy h ough connec ions wi h iends and amily membe s (Sub ahmanyam e Page 3 o 15 Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es al., 2008). Desc ip i e and e hnog aphic esea ch on he beha iou o young people in In e ne highligh s ha SNS a e essen ial o he de elopmen o pe sonal iden i y, media e in he ela ionships o young people wi h hei iends, pa ne s and b oad g oups (e.g., Agos o & Abbas, 2010; I o e al., 2009), boos us and loyal y (Radmache & Azmi ia, 2006), and os e he ob aining o social suppo (La ose e al., 2001). SNS p omo e social in e ac ion and i s ecip oci y o a ain pu poses o communica ion, enjoymen and socialisa ion (Leung & Lee, 2012; Oso io e al., 2014). Ou esea ch pe o ms a igo ous da a p e-p ocessing o handle anomalies, insince e da a, and missing alues. We i s ly exclude ques ionnai es wi h missing alues o inaccu a e alues o ele an i ems (i.e., 26 ques ionnai es). We emo e he o eign indi iduals om he da ase (which was less han 1% o he o al da ase ) because o possible in luences due o na ional-cul u al sc ip s. Secondly, K-means clus e ing allows segmen ing he da a based on membe s’ amilia i y in such a way ha he wi hin-clus e a ia ion is minimised. The e ec i e sample is analysed using NbClus (c . R p og amming language) o iden i y he numbe o po en ial clus e s in he da a. Nine indexes p opose wo clus e s as he bes numbe o clus e s (c . Cha ad e al., 2014, o a e iew o used indexes). We also emo e possible ou lie s ia K-means clus e ing (i.e., 6 ou lie s). We ge a o al numbe o 253 usable esponses. The unc ion Clusplo (c . R p og amming language) is used o iden i y he e ec i eness o segmen ing. In Table 1 he high ela i e g oup is ecognised wi h highe amilia i y (152 indi iduals), whe eas he low ela i e a i ude g oup was iden i ied wi h a lowe le el o amilia i y (101 indi iduals). The s udy is conduc ed using R p og amming language and so wa e en i onmen o s a is ical compu ing. Tab. 1: K-means clus e ing. Desc ip i e s a is ics. Clus e : High n mean sd FAM1 152 6.090 0.770 FAM 2 152 5.760 0.840 FAM 3 152 5.330 0.940 Clus e : Low FAM 1 101 4.340 1.200 FAM 2 101 3.930 1.130 FAM 3 101 2.780 1.040 2.2. Measu es A sel -adminis e ed ques ionnai e is used o collec explici belie s and opinions abou ex ended PCS, RQ and ex ended TAM, use amilia i y, and socio-demog aphic da a. Ins umen de elopmen consis s in e iewing he li e a u e so as o iden i y measu es o each cons uc (Appendix 1). A p e- es assesses he sui abili y o he wo ding and o ma , and he ex en o which measu es ep esen all ace s o he cons uc s. In pa icula , he ques ionnai e is p e es ed wi h ou e-business p o esso s. They assess he sui abili y o he wo ding and o ma and he ex en o which measu es ep esen ed all he ace s o he cons uc s. Sugges ions a e inco po a ed in o a second e sion ha is hen es ed by wo o he e-business p o esso s. No addi ional sugges ions a e made. All i ems use se en-poin Like - ype scales anging om 1, “s ongly disag ee”, o 7, “s ongly ag ee”. On he one hand, o in es iga e he mode a ing e ec s, ou esea ch conduc s sample segmen a ion based on sex and age (18 o 24, and 25 o 34 yea s old). I compa es he means o he same cons uc be ween g oups by sex and age by independen -sample - es s. O e all, he esul s indica e no significan di e ences be ween he g oups; hus we pool he da a om hese g oups oge he . On he o he hand, we conduc a confi ma o y ac o analysis (CFA) o es he alidi y o he cons uc s, including i em loading, cons uc eliabili y, and a e age a iance ex ac ed (AVE). The s udy is conduc ed using R language. CFA p o ides e idence o an accep able model i . Composi e eliabili y o all ac o s in ou measu emen model was abo e 0.7. The AVEs a e all abo e he ecommended 0.5 le el. Following Hai e al.’s (1992) ecommenda ions ac o loadings g ea e han 0.5 a e conside ed signi ican . The i ems in he esea ch model ha e ac o loadings g ea e han 0.5, excep ing TRUST7 and PEOU5. Bo h a e elimina ed. The cons uc alidi y is accep able. To examine disc iminan alidi y, ou esea ch compa es he sha ed a iances be ween ac o s wi h he AVE o he indi idual ac o s. The sha ed a iance be ween ac o s is lowe han he AVE o he indi idual ac o s, excep ing be ween communi y pa icipa ion and communi y in eg a ion –i.e., social in eg a ion o ‘sense o i ual communi y’ de ined he e as he “membe s’ eelings o membe ship, iden i y, belonging, and a achmen o a g oup ha in e ac s p ima ily h ough elec onic communica ion” (see Blancha d, 2007, p. 827). communi y pa icipa ion and communi y in eg a ion a e hus conside ed as a unidimensional cons uc named social in eg a ion ep esen ing bo h he sense o belongingness o and he iden i ica ion wi h he SNS -e.g., eeling like my online communi y is my own- and he social communi y’s in e ac i i y le el -e.g., pa icipa ing in o de o s imula e my online communi y - as an e ec i e ou e o well-being. See Table 2. Page 4 o 15Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es Tab. 2: AVE, and disc iminan alidi y FAM C.ORG S.INT TRUST SAT PU PEOU ANX ENJ SE FAM 0.728 C.ORG 0.354 0.779 S.INT 0.596 0.759 0.719 TRUST 0.217 0.505 0.450 0.728 SAT 0.538 0.536 0.654 0.552 0.701 PU 0.569 0.610 0.701 0.389 0.676 0.744 PEOU 0.363 0.182 0.305 0.153 0.267 0.392 0.745 ANX -0.318 -0.002 -0.183 -0.106 -0.240 -0.202 -0.096 0.696 ENJ 0.511 0.480 0.611 0.250 0.545 0.614 0.289 -0.309 0.795 SE 0.452 0.038 0.188 0.047 0.264 0.453 0.464 -0.117 0.266 0.910 Diagonal elemen s (bold) a e he squa e oo o he a iance sha ed be ween he cons uc s and hei measu es (AVE). O -diagonal elemen s a e he co ela ions among cons uc s. Fo disc iminan alidi y, diagonal elemen s should be la ge han o -diagonal elemen s. No e: FAM: amilia i y; C.ORG: communi y o ganisa ion; S.INT: social in eg a ion; TRUST: us ; SAT: sa is ac ion; PU: pe cei ed use ulness; PEOU: pe cei ed ease o use; ANX: anxie y; ENJ: enjoymen ; SE: sel -e icacy. 2.3. P ocedu e: Pa h inde ne wo k scaling The Pa h inde ne wo k scaling de eloped by Sch ane eld e al. (1985a) is a alid app oach o de i ing ne wo k s uc u es om ela edness da a. Pa h inde ’s ou pu is a di ec ionless ne wo k, no causal, and he p oposed model is no necessa ily assumed o e lec unobse ed, unde lying concep s. The li e a u e also sugges s ha Pa h inde ne wo k scaling e lec s he s uc u e o concep ual domains be e han mul idimensional scaling. The Pa h inde ne wo k scaling shows local ela ions among concep s (local s uc u es), “which a e psychologically meaning ul and possesses highe p edic i e powe o ee ecall pe o mance compa ed wi h o he mul idimensional scaling (MDS) ep esen a ions” (Lau & Yuen, 2009, p.233). MDS’s lack o explici links makes i di icul o in e p e he seman ic dimensions. The non-di ec ed leas -pa h ne wo ks Pa h inde p oduces allow o analysis and obse a ion o in e es ing ela ionships be ween concep s in a pa icula domain (K eie e al., 2005). Pa h inde does no o ce a hie a chical s uc u e on he da a bu iden i ies meaning ul links be ween concep s (Jonassen e al., 1993). The Pa h inde ne wo k scaling in ol es he con e sion o a se o ela edness judgmen s in o psychological p oximi ies, esul ing in a ully connec ed, weigh ed ne wo k o concep s and ela ionships. A key assump ion is he iangle inequali y condi ion; only hose links ha sa is y his condi ion appea in he inal ne wo k (Chen & Paul, 2001; Sch ane eld , 1990). Pa h inde ne wo k is a o m o s uc u al assessmen ha measu es he non-hie a chical in e ela ionships be ween he concep s comp ising an indi idual's knowledge s uc u e. I wo ks by selec ing he pai s wi h he sho es dis ance (i.e., g ea es simila i y) be ween hem (ei he di ec ly o h ough a pa h) while main aining comple e in e connec i i y among all he nodes. Ne wo k s uc u es he e o e depic mo e complex in e ela ionships be ween he concep s comp ising he indi idual’s knowledge s uc u e. Simila i y measu es o i em pai s a e de i ed by calcula ing he cosine measu e be ween he a ings ec o s o each i em wi h each o he i ems which a e p e iously ypi ied. Ou simila i y ma ix is he e symme ic, being he dis ance i espec i e o he di ec ion o he ela ionship be ween pai s o i ems. Following pa ly B anaghan and Hildeb and (2011), he esul ing ma ix is submi ed o a Pa h inde algo i hm (In e link, 2009) using he pa ame e s o q = n -1 (n is he numbe o nodes in a ne wo k) and = ∞. On he one hand, he -pa ame e is he Minkowski me ic used o compu e he dis ance o a pa h. On he o he hand, he q-pa ame e (he e, 45) speci ies he maximum leng h o a pa h subjec o he iangle inequali y es . The q pa ame e is indeed chosen in o de o gene a e he spa ses ne wo k possible om he gi en da a. The model p oposed is conduc ed using Knowledge Ne wo k O ganising Tool (KNOT). 3. Resul s and discussion A la ge numbe o simila i y judgmen s a e ob ained om ou pa icipan s and used o gene a e a ep esen a ion o hei s uc u es based on belie s and opinions, assumed o ep esen he assessmen s ha indi iduals ha e abou he a ious concep s analysed. S uc u es a e based on he p emise ha indi iduals o ganise hei belie s in o pa e ns ha e lec ela ionships be ween hem and he ea u es ha de ine hem. In his sense, wo Pa h inde ne wo ks, one o he mos amilia i y and one o he less amilia ised use s, a e gene a ed. Pa h inde allows he e o he di ec compa ison o he ep esen a ions, and such a compa ison makes i possible o de e mine he e ec i eness o s a egies based on ue ela ionships in online social se ings. The Pa h inde algo i hm supplies wo ne wo ks o compa able complexi y. KNOT p o ides g aphical ep esen a ions o ne wo ks p esen ed in Figu e 1. This layou may be used o quali a i ely e alua e he esul s. Bo h ne wo ks ha e been g aphically enhanced by Adobe Illus a o CS6. The numbe o links is Page 5 o 15 Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es equal o 52 ( amilia ised use s) and 47 (non- amilia ised use s). The simila i ies be ween pai s o nodes in one ne wo k a e compa ed o he simila i ies be ween he same pai s o concep s in ano he ne wo k, which p o ides a measu e o he simila i y o da a se s (c . Goldsmi h & Da enpo , 1990). The co ela ion o he da a in all he pai s o da a iles is he e equal o 0.845. Ou analysis also includes he cohe ence o each da a se (0.847, non- amilia ised use s’ da a se , and 0.912, amilia ised use s’ da a se ). Cohe ence alues e lec he consis ency o he da a. A low cohe ence alue (less han 0.20) indica es ha a es did no (o could no ) ake he a ing ask se iously (o echnical di icul ies). In pa icula , he cohe ence sco es o non- amilia ised use s a e lowe han hose o he mo e ad anced use s. Since he knowledge s uc u es o expe s a e mo e o ganised han hose o no ices, expe s in a gi en domain should possess ela i ely highe cohe ence sco es (Goldsmi h & K aige , 1996). The simila i y be ween bo h ne wo ks (Goldsmi h & Da enpo 1990) is de e mined by he co espondence o bonds in he wo ne wo ks, i.e., he numbe o bonds in common di ided by he o al numbe o unique bonds in he wo ne wo ks. This examines he deg ee o which a concep in he ne wo ks analysed is su ounded by simila neighbou ing concep s. Fo ou esea ch, his p ocedu e assesses he deg ee o which he non- amilia ised use s’ elici ed s uc u e, de ined by he p oximi y da a, app oxima es he amilia ised use s’ s uc u e. Two iden ical ne wo ks yield a simila i y o one ( o iden ical ne wo ks -concep s a e simila ly connec ed in bo h ne wo ks), and wo ne wo ks ha sha e no links yield a simila i y o ze o ( o complemen a y ne wo ks) (Lim & Klein, 2006). In his sense, ou esea ch compa es he numbe o links wo ne wo ks ha e in common (23 links), he numbe o links in common less he numbe expec ed by chance (20.634), he simila i y measu e (0.303), he simila i y co ec ed o chance (0.278), and he p obabili y o ob aining a leas he numbe o common links obse ed by chance (0.000). The esul s he e o e ind a sligh simila i y be ween bo h ne wo ks. Al hough indi iduals a e subjec ed o he same 46 nodes o judgmen s and i is expec ed ha he e would be commonali y be ween he ne wo ks since hey all deal wi h use s’ beha iou in SNS, ou esea ch shows ha membe s di e in he way hey ep esen and bundle hei ne wo ks ela ed o PCS, RQ and TAM. In pa icula , amilia i y abou he unc ioning o an SNS is he e de ined as he numbe o SNS- ela ed di ec o indi ec expe iences wi h he SNS ha has been accumula ed by membe s (c . Ro hae mel & Sugiyama 2001), and he amilia ised use s would he e o e emphasise he pe cei ed use ulness, in insic enjoymen and communi y pa icipa ion. In his ein, he h ee concep s show minimum eccen ici ies (see Figu e 1, le side). An SNS should hus be designed o be s imula ing o use and o e oke compelling use expe iences, and enhance he deg ee o which amilia ised use s belie e ha using he SNS helps hem o a ain gains in ask pe o mance. Familia ised use s (in an in insic enjoymen s a e) ocus hei a en ion on a limi ed s imulus ield, il e ing ou i ele an hough s and pe cep ions. Use s a e less conce ned wi h he isks associa ed wi h asks, and a e likely o eel mo i a ed o ac in an inno a i e and e icien way. In pa icula , amilia i y would a ec use s’ in o ma ion p ocessing ac i i ies in se e al ways, leading o a ep esen a ion based on e iciency-based se ice encoun e s and in insic enjoymen - de ined as he ex en o which pa icipa ion in an SNS is pe cei ed o be pleasu able, exci ing, enjoyable, and un (Ghani & Deshpandé, 1994; Sánchez-F anco, 2006). Fi s , as use s use a se ice mo e equen ly, hey a e mo e amilia wi h i , educing unce ain y in u u e decision si ua ions, and enhancing ela ionships wi h o he s by elabo a ing ad anced knowledge s uc u es (Bha ahe jee, 2001; Fla ián e al., 2006a; Ge en, 2000). Second, amilia ised use s belie e ha hey know an SNS well, showing an indi idual judgmen o one’s capabili y o use an SNS and mo e posi i e and enjoyable belie s and assessmen s abou i . Once use s a ain a le el o pe cei ed abili y and con ol wi h he SNS, hey a e ee o explo e i s capabili ies. Tha is o say, si ua ions in which challenges and skills a e pe cei ed o be simila a e assumed o acili a e in insically enjoyable expe iences. Thi d, Zajonc and Ma kus (1982) p opose ha amilia i y leads o posi i e assessmen s o a se ice. Familia i y equi ed o he success ul execu ion o SNS causes a posi i e ecep ion o he idea o social in e ac ions conduc ed online. In sum, highly amilia ised indi iduals a e likely o be less conce ned wi h he usabili y (i.e., pe cei ed ease o use) o SNS, being loca ed in lowe cen ed posi ions. Likewise, Ou esea ch e idences a di ec link be ween ac i e pa icipa ion and communi y o ganisa ion s essing he pe o mance expec ancy o such in e ac ions wi h he needs o he communi y o ganisa ion (e.g., esponding o calls o suppo my communi y – as an e ec i e ou e o wellbeing). Mo eo e , pe cei ed use ulness is weighed s ongly and di ec ly linked o in insic enjoymen . Al hough highe pe cei ed con ol should lead o g ea e expe ien ial beha iou h ough enjoymen emo ional esponses, ou esea ch shows how indi iduals do no ully e idence he di ec in e ela edness (i.e., be ween in insic enjoymen and sel -e icacy) in he domain o online social se ices. Page 6 o 15Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es Fig. 1: Visualisa ion o Pa h inde ne wo k scaling. Highe amilia i y s lowe amilia i y No e: FAM: amilia i y; C.ORG: communi y o ganisa ion; C.INT: communi y in eg a ion; C.PART: communi y pa icipa ion; TRUST: us ; SAT: sa is ac ion; PU: pe cei ed use ulness; PEOU: pe cei ed ease o use; ANX: anxie y; ENJ: enjoymen ; SE: sel -e icacy. Meanwhile, non- amilia ised use s emphasise i ems ela ed o pe cei ed ease o use, communi y o ganisa ion and sa is ac ion. See Figu e 1 ( igh side). The key e m ela ed o pe cei ed ease o use is an essen ial node in his ne wo k as a ep esen a ion o indi iduals’ men al models. Non- amilia ised use s desc ibe hei s uc u es in e ms o he deg ee o which a pe son belie es ha using a SNS would be ee o e o (c . Da is 1989) and as a sou ce o suppo , wi h online exchange o suppo being as help ul as suppo p o ided by o line con ac s (Tu ne e al., 2001). The lowe sel -e icacy o pe cep ions among non- amilia ised use s can cause an inc ease in he salience o he e o o expec ancy in de e mining beha iou al in en ion owa ds SNS. Lowe sel -e alua ions migh emphasise he conside a ion o mo e cues ela ed o he deg ee o which a use belie es ha using an SNS is ee o e o . Whe eas amilia ised use s p ima ily expec an SNS o be mainly use ul and enjoyable, and may dis ega d issues a ising om se ice complexi y, non- amilia ised use s p ima ily emphasise he pe cei ed ease o use and a eeling o being assis ed by he online communi y in e ms o suppo needs and esou ces a ailable o he use (communi y o ganisa ion; c . He e o and G acia, 2007). The high pe cei ed isk associa ed wi h SNS usage signi ican ly educes non- amilia ised use s’ pe cep ion o he belie in one's e ec i eness in pe o ming a ge asks (c . he e applica ion-speci ic sel -e icacy, c . Aga wal e al., 2000) and, in u n, ease o use. These use s may no be willing o accep a complex in e ace. Communi y manage s should hen design a suppo i e mechanism and an a able en i onmen (e.g., enhancing online in e ac ion and communica ion among membe s, o pe iodic ace- o- ace mee ings) ha mo e e ec i ely os e he indi iduals’ sel -e ficacy. Following Sch ane eld e al. (1985b, p.700; c . also Meye & Scha ane eld , 1976), “ wo concep s ha a e di ec ly linked a e iewed as mo e simila han a e wo concep s ha a e no linked o a e indi ec ly linked”. Highly- ela ed concep s consequen ly end up di ec ly connec ed, and less- ela ed concep s equi e he a e sal o wo o mo e links. On he one hand, amilia ised use s ind he cumula i e online social in e ac ions ewa ding, ul illing and s imula ing. In e ac ions in ol e in ospec i e social expe iences and in eg a ion -which e e s o conc e e in ol emen in ac i i ies. Familia i y esul s in a (pa ial) cogni i e link be ween pa icipa ion and communi y in eg a ion ia in insic enjoymen . Following he p e ious li e a u e abou use amilia i y would hus p esen social in eg a ion as a second-o de cons uc o med by communi y pa icipa ion and communi y in eg a ion which exis s on he same le el as he dimensions, and is inhe en ly associa ed wi h in insic enjoymen . Bo h dimensions o he social in eg a ion cons uc would be hemsel es cons uc s concei ed as speci ic componen s o he gene al cons uc hey collec i ely cons i u e. On he o he hand, non- amilia ised membe s would no be awa e o concep ual di e ences be ween he wo communi y dimensions, social in eg a ion exis ing as a supe o dina e second-o de cons uc e lec ed in communi y pa icipa ion and communi y in eg a ion –based on he in e -co ela ions be ween i s -o de ac o s o ming a sys em o in e dependence ha is i sel impo an in measu ing he high-o de in eg a ion. Page 7 o 15 Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es Social in eg a ion would be a gene al concep ha is mani es ed by i s dimensions. In his sense, Sánchez- F anco e al. (2015) also p opose a posi i e co a iance be ween communi y pa icipa ion and communi y in eg a ion de ined as ‘‘membe s’ eelings o membe ship, iden i y, belongingness, and a achmen o a g oup ha in e ac s p ima ily h ough elec onic communica ion” (Blancha d, 2007, p. 827; c . also Blancha d, 2008; McMillan & Cha is, 1986). Fu he mo e, he impo ance o p i acy in de e mining online us should be e en g ea e o ca ego ies wi h pe sonal in o ma ion a isk - o example, online social se ices - han i is o o he online ca ego ies (Ba e al. 2005). T us de e mines he ex en o which use s engage in a ela ionship. The pe cei ed isk o a ela ionship is high due o he lack o ace- o- ace con ac (Ridings e al. 2002) when communi y membe s a e s ill un amilia wi h one ano he . Non- amilia ised use s should hus end o demons a e a g ea e conce n wi h he isk o online se ices and SNS anxie y. Con a iwise, when amilia ised cus ome s e alua e an SNS h ough ul illing, g a i ying in e ac ions, he need o us – by de eloping PCS – becomes less ele an (c . Sánchez-F anco 2009); people may become mo e willing o accep unce ain y and ulne abili y as hey a e able o d aw om di ec expe ience, a he han su ace cues, o in o m hei expec a ions ega ding he u u e beha iou o an SNS. To sum up, non- amilia ised use s a e slowe in becoming loyal use s and spend longe deciding o y new se ices. An expe use o he SNS is mo e likely o ha e a g ea e con idence in he SNS han a no ice use . Al hough us cons uc s and main ains he exchange o ideas in an in e connec ion (Blau, 1964) when amilia ised/expe indi iduals e alua e an SNS h ough ul illing, and easy in e ac ions he need o us disappea s. Howe e , anxie y and us a e clea ly loca ed in he pe iphe y o bo h knowledge ne wo ks. The le el o a e age eccen ici y o us and anxie y is also high in he non- amilia ised sample. SNS cus ome s o e es ima e hei unce ain ies pe cep ions, unde es ima e hei us be ween he pa ies, and could damage he c ea ion o PCS and ela ional capi al which exis s when indi iduals ha e a s ong iden ifica ion and us in a ne wo k (Wasko & Fa aj, 2000). A possible explana ion could lie in use s’ desi e o use s’ popula i y. Al hough lowe le els o amilia i y in SNS and us in o he membe s should be a signi ican p edic o o highe le els o in o ma ion con ol and lowe in o ma ion disclosu e on SNS (Wal a e e al., 2012), membe s may easily p io i ise hei desi e o popula i y a he expense o hei p i acy (Sánchez-F anco e al., 2015). Acco ding o Ch is o ides e al. (2009, p.342), “despi e p i acy conce ns, indi iduals disclose a g ea deal o pe sonal in o ma ion”. Assuming hese associa i e ne wo ks, SNS mus he e o e p omo e iend g oupings, i em-le el access con ol, block lis s and back a ange o o he echniques o p i acy managemen . 4. Conclusion O e all, ou esea ch ocuses on he associa ion be ween PCS, RQ and u ili a ian- and hedonic-bene i s among young adul s. On he one hand, i assesses he pe cep ions o young adul s who a e cu en ly using SNSs, ocusing on he ue commi men phase o he beha iou chain model, and hus assuming ha he ul ima e success o SNSs de i es om a long-las ing usage. On he o he hand, i p oposes ha g ea e le els o mo i a ions based on ecei ing com o , empa hy, and o he ela ionship benefi s lead young adul s o eel ha hey a e being suppo ed by a whole po ion o hei social communi y, and s essing a sense o emo ional in ol emen wi h o he s. In his ein, social capi al inc eases membe s’ eelings o a achmen o hei SNS and expec a ions o con inui y, so ha young adul s s eng hen ela ionships wi h o he s and ob ain a ec i e bene i s om mu ual ela ionships (c . Casaló e al., 2007). The amoun and dep h o suppo ha membe s ecei e om o he membe s a e he e posi i ely associa ed wi h communi y pa icipa ion, a lowe unce ain y o hei ela ionship quali y wi h i , and an enhancemen o communi y in eg a ion. Likewise, pe cei ed enjoymen and ex insic an eceden s based on TAM (i.e., pe cei ed ease o use, and pe cei ed use ulness) a e o e all iden i ied as impo an mo i a ional ac o s in SNS accep ance and usage. Finally, ou esea ch p oposes he need o ocusing on di e en indi idual cha ac e is ics. In pa icula , o gua an ee he communi y’s su i al in he long e m, communi y manage s should encou age young adul s o sha e knowledge and ideas ela ed o mu ual in e es , o de elop s onge eelings o being suppo ed by he SNS and o s eng hen hei iden i y, belonging, and a achmen o i by s imula ing (a) he deg ee o which a membe belie es ha he in o ma ion ob ained om he SNS will p o ide a ange o ela ional benefi s ha would o he wise be di ficul o ob ain wi hou pa icipa ing in i , (b) he pleasu e he indi idual eels objec i ely when p ac ising a pa icula beha iou o ca ying ou a pa icula ac i i y, (c) ex insic mo i a ion ha desc ibes an indi idual’s pe sonal gain associa ed wi h he use o a pa icula SNS, and he deg ee o which a membe belie es ha using he online social se ice would be ee o e o o e alua e he i ual communi y posi i ely. Mo eo e , ou s udy ecommends he analysing o isola ed subse s o indi iduals’ cha ac e is ics and subsequen beha iou s ela ed o PCS, RQ o ex ended TAM based on u ili a ian- and hedonic-mo i es. Fo ins ance, assuming he e ha amilia i y is a signi ican cue o alidly explaining choices o SNSs, communi y manage s migh enhance membe s’ loyal y by a ge ing hei le els o in e ac ions, expe iences, and lea ning o wha , why, whe e and when o he s do wha hey do (Ge en, 2000). Ou esea ch empi ically analyses a heo y-g ounded amewo k o in eg a ing indi idual cha ac e is ics and ela ional d i e s, and i ocuses on associa i e s uc u es e idenced as a ep esen a ion o he mos salien loyal y-based concep s by modelling he in e ac i e e ec s o amilia i y. Page 8 o 15Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Kybe ne es Appendix 1: Measu emen scales. Cons uc s/Indica o s ENJ. In insic Enjoymen (Saadé and Bhali, 2005; Low y e al. 2013) ENJ1. On he whole, in e ac ing ia he SNS is pleasu able ENJ2. In gene al, I enjoy in e ac ing ia he SNS ENJ3. O e all, in e ac ing ia he SNS a ouses my cu iosi y PU. Pe cei ed Use ulness (Da is,1989; Da is e al. 1992; Lo enzo-Rome o e al. 2011; Sánchez-F anco, 2006) PU1. I eel ha he unc ion and he u ili y o he SNS a e use ul o me PU2. Using he SNS makes i easie o in e ac wi h o he use s PU3. Using he SNS gi es me access o a la ge quan i y o in o ma ion PU4. In gene al, I ind he SNS ex emely use ul PEOU. Pe cei ed Ease o Use (Da is,1989; Da is e al. 1992; Lo enzo - Rome o e al. 2011; Sánchez - F anco, 2006) PEOU1. Lea ning o use he SNS is simple PEOU 2. I ’s easy o lea n how o use he SNS PEOU 3. I ’s easy o me o do wha I wan o do on he SNS PEOU 4. I ’s easy o emembe how o use he SNS PEOU 5. In e ac ion wi h he SNS is clea and unde s andable PEOU 6. In gene al, I ind he SNS easy o use ANX. Anxie y (Ha ison and Raine , 1992) ANX1.The SNS make me eel uneasy and con used ANX2. Wo king wi h he SNS would make me e y ne ous ANX3. I ge a sinking eeling when I hink o ying o use he SNS SE. Sel - E icacy (Venka esh e al. 2003; Sánchez-F anco and Rod íguez, 2004) SE1. When I pos new commen s, in o ma ion, expe iences, pho os, links, e c. I clea ly know he igh hings o do SE2. I eel con iden pos ing new commen s, in o ma ion, expe iences, pho os, links, e c. i he e is no one a ound o ell me wha o do as I go C.INT. Communi y In eg a ion (Geyskens e al. 1996; Loewen eld, 2006; Sánchez - F anco, 2009) C.INT1. My emo ional ies in he SNS I belong o a e a key eason why I am s ill a membe C.INT 2. I eel a ease in he SNS; he ela ionships I ha e in he SNS a e e y enjoyable C.INT 3. I eel a sense o owne ship o he SNS; I eel like an in eg al pa o i C.INT 4. I iden i y mysel wi h he SNS I belong o C.INT 5. My opinions a e gene ally alued by o he membe s o he SNS whe e I am a membe C.INT 6. I eel emo ional ies wi h he SNS I am a membe o C.INT 7. In e ac ing wi h and ela ing o o he membe s means a lo o me C.ORG. Communi y O ganisa ion (Geyskens e al. 1996; Loewen eld, 2006; Sánchez-F anco, 2009) C.ORG1. I could ind people ha helped me eel be e C.ORG2. I could ind someone o lis en o me when I el down C.ORG 3. I could ind a sou ce o sa is ac ion o mysel C.ORG 4. I could be able o chee up and ge in o a be e mood C.ORG 5. I could elax and easily o ge my p oblems C.PART. Communi y Pa icipa ion (Geyskens e al. 1996; Loewen eld, 2006; Sánchez-F anco, 2009) C.PART1. In gene al, I ge ac i ely in ol ed in he SNS C.PART 2. I am a membe o di e en g oups o ganised h ough he SNS, and I pa icipa e in hese g oups C.PART 3. I con ibu e o he SNS by pos ing new commen s, in o ma ion, expe iences, pho os, links, e c. C.PART 4. In gene al, I ake no ice o he commen s and con ibu ions o o he membe s C.PART 5. In gene al, I espond eage ly o commen s made by o he membe s on he SNS TRUST. T us (Lo enzo - Rome o e al. 2010) TRUST1. I eel ha he SNS keeps he p omises and commi men s i makes TRUST2. I belie e ha he in o ma ion o e ed on he SNS is since e and hones TRUST3. Openness and anspa ency when p o iding se ices o use s a e cha ac e is ics o he SNS TRUST4. The SNS is us wo hy TRUST5. The SNS ac s esponsibly TRUST6. In gene al, I us he SNS TRUST7. Pos ing pe sonal in o ma ion on he SNS doesn’ eally wo y me TRUST8. I belie e ha I can us he SNS TRUST9. I eel ha he SNS’s beha iou is e hical TRUST10. I belie e ha he SNS is capable o pe o ming i s in ended unc ion SAT. Sa is ac ion (Fla ián e al. 2006b; Janda e al. 2002; Smi h and Ba clay, 1997) SAT1.In gene al e ms, I am sa is ied wi h my expe ience in he SNS SAT2.I hink ha I made he co ec decision o use he SNS SAT3.I ha e ob ained se e al bene i s de i ed om my pa icipa ion in he SNS FAM. Familia i y (Dick e al. 1995; Fla ián e al. 2006a; Gu soy, 2001) FAM1.O e all, I am qui e amilia wi h he SNS FAM2.I am e y amilia wi h he se ices o e ed by he SNS FAM3.In compa ison wi h he ypical use o he SNS, I belie e I am qui e amilia wi h i SNS: Facebook. Page 15 o 15 Kybe ne es 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60