Online identity : constructing interpersonal trust and openness through participating in hospitality social networks

The present article describes the results of research on online identity construction during the participation in the hospitality social networks. Specifi cally the user references are analysed to understand patterns that form the image of a member. CouchSurfi ng service (couchsurfi ng.org) allows users to leave short texts where the experience of hosting/being hosted by a CS member is described, is an evaluation of the CS members of each other’s personal traits, skills and common experience. Therefore references can become a good instrument for portraying a CouchSurfi ng member and understanding his or her particular traits. References form an important part of a user’s virtual identity in the network. Using a sample of references of Spanish CouchSurfi ng users, the research established main characteristics of the references, which are the openness, readiness to share ideas and experiences and trustworthiness. These concepts illustrate the typical traits associated with a user of the network and also shed light on the activities common during offl ine CS meetings


Introduction
Hospitality social networks are the Internet social network services (SNS) used to arrange free accommodation during travel.These networks developed from the offl ine hospitality exchange directories and clubs.These early clubs allowed their members to access a special directory (essentially a yearly book) with contact details of other members, in order to arrange stay during travel.Early hospitality exchange networks did attract attention -US president Jimmy Carter was a founder of one such network: Friendship Force International, although they never became ubiquitous and popular enough to substantially affect the travel routines of many people.Everything changed after a rapid ascent of the Internet as a communication technology: communication had never been so effortless and everything that depended on the speed of human interactions naturally began to migrate to the web.Internet hospitality exchange networks were one of the fi rst to benefi t from social media boom.In 2004, a year before Facebook, couchsurfi ng.org became the fi rst hospitality service to exhibit all three defi ning properties of an SNS.It had user-created profi les, DOI: 10.15503/jecs20131-47-56 visible connections between users and a way to interact through these connections.
In the following years CouchSurfi ng steadily grew in popularity, attracting large amounts of media attention in the process.It has reached 4 000 000 users in 2012 and has become a truly global phenomenon (by comparison, the largest still existing offl ine hospitality exchange, the Friendship Force International has only 18 000 members)."Couch surfi ng" itself became a euphemism for the activity promoted by the website.So far online hospitality exchange services have not attracted much academic attention.Most notable works include study of online familiarity by Paula Bialski (2012), management of trust within hospitality SNS by Lada A. Adamic, Debra Lauterbach and Chun-Yuen Te (2011), and the research on interpersonal trust in CouchSurfi ng by Le Dieu Tran (2009).
While CouchSurfi ng is also an SNS like Facebook, MySpace or Twitter, it should be approached separately as it has several important differences from more conventional networks.Firstly and the most signifi cant is the nature of interactions between users: while in Facebook most of the connections are formed between people who had already known each other offl ine, the so-called weak ties, hospitality SNS in their essence are designed to encourage and facilitate the communication between strangers, people who had never met before.From this point of view hospitality SNS are closer to traditional dating sites than to the networks like Facebook.The second important difference also evokes a similar connotation: the fi nal goal of the interaction between users in the hospitality network is meeting in the real life.MySpace, Facebook, Google+ are all designed to provide immersive online experience for the Internet user, so the user can maintain its egocentric network, share content and experiences without any need for offl ine communication.Contrary to that, CouchSurfi ng is designed in a way to arrange, prepare and simplify the transaction from online to offl ine communication.In a sense, online experience in hospitality SNS is secondary to the offl ine interaction.This makes networks like CouchSurfi ng a particularly interesting and a tempting subject for research into how online and offl ine components relate and interplay in the lives of Internet users.

Identity construction online
One of the distinctly interesting discussions within the Internet research is centred on the question of how Internet users construct their identity.Online identity construction is one of the most interesting and research intensive focuses of Internet studies, although it is worth noting than most of the works dealing with identity construction on the Internet study the identities in the conditions of anonymity.Only with the rise of Facebook, studies of the identity constructions in the real-name communities began to appear (Zhao, Grasmuck, & Martin, 2008).Just like Facebook, CouchSurfi ng is not an anonymous community.Contrary to many other internet communities hospitality exchange networks demand the reality of identity information and addresses.This is done through verifi cation processes.On CouchSurfi ng verifi cation is a routine that demands the authenticity of the name and address, established through a "verifi cation payment" with a card with the name and billing address corresponding to the one indicated on the website.Because of its nature and inevitability of offl ine user interactions, CouchSurfi ng (and other hospitality SNS as well) practically require the authenticity not only of the members' names, but of also of addresses and photos.Such de-anonymisation is both directly encouraged by the system (through the system of verifi cation) and conditioned by the inner ethics of member interaction.As all the interaction in the hospitality networks is based on trust, photos are often required by hosts to form an image of a future visitor before letting her or him into the home.Related research dealing with real-name networks includes works on identity construction in Facebook by Shanyang Zhao, Sherri Grasmuck and Jason Martin (2008), Soraya Mehdizadeh (2010), Katie Ellis (2010), identity construction in Twitter was discussed by Dawn R. Giplin (2011).
As most of the user-to-user communication on CouchSurfi ng is meant to culminate in an offl ine meeting, identity construction in the hospitality SNS ideally should serve two functions.Firstly, identity should be trustworthy enough to serve its purpose, either being accepted as a guest by the hosting user, or be visited by the guest.Secondly, it should be the representation of a member.CouchSurfi ng users represent themselves primarily through their profi les.Profi le information is what forms a user's identity in the network.Couchsurfi ng profi les are the subject of the present research and the references which constitute a part of a profi le are what was analysed during the course of the study.Similar to the profi le pages on any other SNS, CouchSurfi ng profi les contain both user-and system-generated information.User-generated information is used for self-presentation and the direct construction of identity.System-generated data is the refl ection of the user's activity in the network; it includes lists of "vouch" and "friend" connections, references and statistical information.User-generated information is the interesting part for the purposes of the present paper -as it depends solely on a user and basically forms his or her identity in the network.In their research Paula Bialski and Dominic Batorski outlined three stages of the user experience in the CouchSurfing network, when the trust emerges.Pre-selection of users, website navigation and offl ine contact are the three dimensions of CouchSurfi ng usage that form a familiar environment and allow the trust between the community members.The second stage: website navigation is actually the creation and interaction through and with the help of the user profi les and it is arguably the most signifi cant one (Bialski, & Batorski, 2010, p. 202).Profi les are the sources of trust in users and they are what other members evaluate when they judge whether a profi le owner may be trusted.The profi le in Couch-Surfi ng is structured in such a way that when a user creates one, he or she answers a specifi c set of questions: "What is your personal philosophy?", "Why do you live your life the way you do?", "most amazing thing you've done?" and others.Whilst the way these questions are laid out creates a more trusting environment by making profi les more similar and thus familiar, it is the content that is evaluated by the other users to form an image about other hospitality network members.The apparent structure of the profi les also allows more refi ned and precise comparison of different users' profi les as this structure undoubtedly somewhat limits the member's activity of expression, making the profi les more similar to open-ended survey responds than to the free compositions on the vaguely identifi ed topic.
References are part of a user's profi le created as a result of a user-to-user interaction.Basically a reference is a usually a short comment left by the host and guest after the visit took place.There are no specifi c restrictions established by the system apart from the maximum number of words which is 1000.Typically though a reference is 40-150 words long and resembles a "thank you" note describing the time spent together and recommendation for other members concerning a particular host/guest.References may be neutral, negative or positive, but in reality the distribution is very uneven: roughly 90% of all references are positive (Adamic, Lauterbach, & Te, 2011, p. 7).The references are meant to be reciprocal, yet sometimes a reference is left only by one party.One possible reason for declining to leave a reference may be the unwillingness to leave a negative one.

Research on trust in hospitality networks
One important concept analysed in this paper is trust and its representation in the user profi les as a part of their Internet social identity.For the purposes of the present paper trust is defi ned according to the Lynne Zucker's defi nition as "an expectation or one's belief that the other party will behave in a dependable, ethical, and socially appropriate manner" (Zucker, 1986) and applied towards the unknown other social network users, essentially strangers.In this sense it is generalized social trust (Delhey, & Newton, 2004) that is being exercised by members of the social networks towards each other, when they have not engaged in any interaction, so that the trust might be affected by some other (personal) reasons.The existing research of trust on the online social networks followed two paths.The fi rst direction is aimed at computing and objectivizing trust for the purposes of the design improvement of existing and future online services.In this case trust is closely connected to privacy and safety on the Internet.The fi nal goal of this research is to provide intending users with reliable and trustworthy information, making the Internet experience safer and more reliable (see Smith, 2009 andFurht, 2010).Still, at present, managing trust online remains the area controlled primarily by the users, who may grant or restrict access to their information, using the tools provided by the services they use.The second direction of the trust research online deals exactly with the way people control their privacy and manage trust during the interactions on the Internet (Bhulyan, Josang, & Xu, 2010).This research of practical aspects of managing trust in the SNS concentrate on sociological and behavioural aspects of participation in social networks, privacy issues and the way the egocentric social networks are constructed by SNS members (see (Dwyer, Hiltz, & Passerini, 2007) (Trepte, & Reinecke, 2011)).The existing research is however limited to a number of most popular SNS.Facebook as the most widespread social network is the most common subject of analysis (see Johnson, Lavesson, Zhao, & Wu, 2011;Dwyer, Hiltz, & Passerini, 2007).
Trust is connected with familiarity.In the context of the Internet this suggests that more familiar and understandable services appear more trustful.And in the hospitality exchange SNS more familiar users are more likely to interact or connect with each other (Bialski, & Batorski, 2010, p. 190).Indeed research confi rmed that more similar network members are signifi cantly more likely to interact than very different ones.It was also established that trust is formed through the interaction with user profi les, and information contained in the profi les was the main factor that affected the person's (dis-)trust towards other network member (Bialski, & Batorski, 2010, p. 180).Profi les are the main focal point of the present research for the same reason: they are the main source of user information in the network and simultaneously the main tool used for identity construction.Another research on familiarity of trust conducted by Jie Zhang (2004, p. 23) suggested the possibility of measuring online familiarity through calculating several quantitative factors.The factors described by the author suggest that measuring familiarity (and thus trust related to familiarity) is advisable while analysing controlling for the amount of previous experience with the service.For that reason the sample used for the present work's analysis consists of users with similar backgrounds in using the CouchSurfi ng service.

Methodology of the study
Investigating the ethnography of Web, we use a plural paradigm under the general umbrella of Web Content Analysis.This complex method was explained and described in the research by Susan Herring in 2010 and is defi ned as a broader construal of web content analysis, one that goes beyond traditional content analysis, drawing on methods from other disciplines (in particular linguistics and sociology) (Herring, 2010).Web content analysis in this case should be understood as analysis of web content using various (traditional and non-traditional) techniques.Integrating different methods into the analysis of the users profi le enables us to deal properly with multimodal structure of a hospitality social network and cover a broad range of content.
Plural, broad or expanded Web Content Analysis paradigm involves methods of discourse analysis and social network analysis that are operationalized and implemented within a general content analytic framework, addressing characteristic features of the web.Besides, the methods of traditional content analysis adapted to the researching the blogs are applied.Blogs resemble the profi les on couchsurfi ng.org because of their structure.Profi les the same as blogs don't consist of texts only.Typically they are structured in various parts and the location and formatting of the text can be signifi cant, not only the direct meanings of words.In case of analysis of profi les we can take into account the links, announcements, participation in events, membership of belonging to special groups etc.These factors can infl uence the interpretation of the results.Thus analysis units, similarly to other researches of blogs 7 , include information blocks about the user, images, description of the host's apartment and references.The expanded methodological approach to content analysis considers content not just as thematic meanings present in text but interprets content as various types of information "contained" in the user profi les on CouchSurfi ng website, including features, links, and exchanges, all of which can communicate meaning.However in this particular research we analyse only one structural unit of a profi le: references.A broader analysis of multimodal and multifunctional web content of hospitality social networks (e.g.images, connections, references, links etc.) is still to be done.References are the short texts (usually up to 150 words) where the users describe and evaluate the experience of hosting or being 7 Such features of blogs were analyzed by Herring, Scheidt, et al. (2004, 2005, 2006) using a classical CA approach.
hosted by the particular CS user.We consider content framed within the references left on users' profi les on CouchSurfi ng as socio-cultural text.Therefore the elements of discourse analysis are applied to the profi les.Describing and exploring the texts, we are trying to uncover the "underlying meaning and signifi cance" (Herring, 2010) of the texts.References normally don't include any other content except for the text.We also don't take into account whether the reference is positive, negative or neutral (the ranking system is integrated in the web).Thus analysing the references we apply traditional content analysis and discourse analysis.288 references/ 144 pairs (usually both members of CS leave the references after the meeting).The criteria for choosing these particular 288 references is the geographical location of the member: references analysed were left by/for the members from the Basque country in Spain (primarily Vitoria and Donostia regions).When analysing the texts the fact that most of them were written by not-native-speakers was also taken into account.The rationale for choosing the language-focused approach combined with traditional content analysis is that in an online environment the human behaviour is carried out mostly through linguistic means (Herring, 2010).And in this particular case the texts become the instrument of an online identity construction.
The aim of the research is to analyse how the categories of trust and openness are expressed though the references left by/for the members of couchsurfi ng.org.The data for analysis is 288 references / 144 pairs by/for the experienced members from Basque country (primarily Vitoria and Donostia regions), collected in December 2012.We defi ned experienced member as the one having received at least ten references and active at least once in the last 6 months.For the present sample only positive references were collected as only 2 negative and one neutral reference were found while analysing intended profi les and no signifi cant conclusions could be made judging such a small number of references.Carefully reading the references we took into account not only the precise phrases and words used by the users, but also analysed the word choice and the general style of the text according to the categories chosen.

Results: interpersonal trust and openness
Interpersonal trust.The unit of reference itself is a very important element in the structure of a CoushSurfi ng profi le.Normally the CS members tend to ignore the requests without references or at least handle them cautiously.However reference is not a necessary requirement to participate in the community, it is a basic element that helps to create the atmosphere of mutual trust within the social network.Most of the references left on CS are positive, though negative ones also exist.If in other social networks like Facebook or MySpace the user constructs his or her identity as he wants others to see him/her, CouchSurfers are allowed to construct their online identity only partially.The rest of the work is done by other members of a community through the references they leave.Reference-based identity and profi le identity constructed by the CS user him/herself can differ.For instance, one of the references states: "No matter how scary his profi le seemed at the very beginning (to me at least ;-)), R. is a great and entertaining person and I really enjoyed his company and jokes" [sic].User R. wanted others to see him as a gloomy person, but references left by other users, who stayed at his place construct another identity of R., showing up his sense of humour, etc. Reference is a fundamental brick in building the trustworthy profi le, because it is the only element of a profi le (apart from statistics) that the member him/herself cannot infl uence.The simple tool that allows the CS members to evaluate and describe their hosting or surfi ng experience helps to create a social network with an "atmosphere of trust", described in one of the references.The CS members however rarely refer directly to the categories of trust in their references, preferring to describe the shared activities and personal traits of their host or a guest.It may be also suggested that writing that "you can trust him/her" is likely to be interpreted as if the member had doubts in trusting the particular user before meeting him or her personally.From 288 references only nine people write directly that a person is trustworthy, one of them thanks for trusting.The other way to express the trustworthiness of a member is to use language constructions addressing the potential hosts/ surfers of a member.The word "you" contain 140 references and 30 of them contain the structures "you can/ you have to/ you should".For instance: "If you host them", "with whom you can have a great time, you can talk about everything with him", "I highly recommend that you host her, you should defi nitely get in contact with him", "defi nitely a person that you have to meet", etc. 13 references contained direct recommendations to host or surf a couch of the referenced member, using the verb "recommend".References, being an instrument for making people trust each other in a virtual community contain direct messages of trust.Describing the shared experiences is also aimed at making other people interested in a particular user.In other words references are the disclaimer of trust within couchsurfi ng.org.
Openness.The single important mantra and the network-defi ning notion on Co-uchSurfi ng is openness.It is both arguably the most popular self-defi ning characteristic in the member profi les and at the same time the most wide-spread attribute of a good couchsurfer as described in the references.Partially because the openness to new experience and other people is the cornerstone of CouchSurfi ng sharing philosophy and is seen as a defi nite virtue for a member, and partially because "open" is a trait neutral enough for one to call him/herself so and not appear boastful.Almost a quarter of all references characterized a person on the receiving end as open or open-minded either directly or implying that by referring to a particular activity or attitude.Such references included mention of the willingness to discuss one's culture ("She shared her culture with us", "I would greatly recommend the experience if you are truly interested in learning more about the local culture", "Thank you for your kind explanations about different cultures"), member's knowledge of other culture ("he was speaking Euskera like a native", "he explained local traditions to us"), or indirectly, referring to the ability to listen and respect others.The willingness to share or exchange something is also expressed fairly often.14 references contain the keywords "share/ exchange", where stories, experiences, cultures, conversations, love for life, trip are shared.Out of 288 analysed references in total 62 references contained a mention of the openness of the referenced party, making "openness" the single most frequent positive characteristic in the sample.
Ability and desire to learn as well as the corresponding activity of tutoring is another recurring theme.As a result 31 references relate to learning or teaching.The members are both positively described as people eager to learn and people able to teach something (ready to learn new things, learns amazingly fast, would like to continue learning, teach, he has taught me more about Norway than I ever knew before).CouchSurfi ng is very often a transnational activity that's why it often becomes an instrument for learning new languages (especially in Europe where the concentration of diverse nations and languages is much higher than on the other continents).Besides that, considerable numbers of couchsurfers (28% of hosts in the sample) who practice couchsurfi ng in the Basque country are foreigners coming from a variety of different states.All this frequently makes couchsurfi ng a multilingual experience, and although in most cases users have English as a common language, the actual level of the language differs and more than that, the relation between different languages is a popular topic for CS conversations.Signifi cant numbers of keywords relate to the positive mental and social characteristics of the referenced members.Such attributes are usually listed in succession, praising a successful couchsurfi ng experience ("He is just so intelligent, interesting, easygoing and generous", "They are sensitive, cultivated, respectful, humble, funny, laid--back, easygoing, curious").Judging from how often they can be seen in references, such lists are more the signs of politeness bound by etiquette than objective descriptions of the person.It is worth noting though, that these attributes tend to levitate towards social aptness (the most common attributes being interesting (42), friendly (36), generous (21), positive (20) and easy-going (15), helpful (39) all relating mostly to the extroverted type of a person.Such clear prioritization depicts the SNS's focus on sharing and participative behaviour.In total more than 92% of all sample references contain positive descriptive attributes, of which 76% are those related to communicated skills.
Another relatively common theme in the references is CouchSurfi ng itself.5.2% of all references contained mention of CouchSurfi ng values, mission or spirit.Usually such mentions were coupled with positive valuation of someone's hospitality, respectfulness or willingness to help.Respect for CouchSurfi ng values is also connected to trust ("I think she fi ts perfectly to CS mission.I would trust her with my life").

Conclusions and future research
The analysis of the references sample of Spanish CouchSurfi ng members revealed the tendency of users to concentrate on social traits when writing a reference for the other couchsurfi ng.org member.Among these traits openness and desire to share (experiences, own culture, ideas) are the most wide-spread.Trust, respectfulness and the respect of the implicit CS mission and values are also a popular topics included in the references.Common activities are also an attribute of successful couchsurfi ng experiences as they were mentioned in a fi fth of all references.As a result of this research an outline of a typical positive reference can be created, mentioning the most valuable and important traits for a couchsurfer as seen by other CouchSurfi ng members: interesting, open-minded, generous, positive, trustful, easy-going, helpful and willing to share.
The content-analysis of the references may also help to reveal the most common shared activities of the members during CouchSurfi ng.Even cursory examination reveals plenty of information related to the activities undertaken by the members.20.2% of examined references contain mention of a specifi c shared activity: 5.9% (17) speak of "town tour", "excursion" or "guide through town", 10% (29) mention shared cooking and 7.2% (21) have mention of visiting bars or restaurants.Another frequent activity is language tutoring.User references may prove to be a valuable and objective view on the way CouchSurfi ng members communicate offl ine.Such indirect observation may be useful when studying this aspect of hospitality exchange, though undoubtedly further sophistication is needed to usefully apply the analysis to the references data sample and a larger sample is also advisable.
Further development of the analysis may lead to an increase of the analysed sample size, application of the approach to a wider geographical zone and inclusion into the analysis not only the references but also "friends comments" -short descriptions that accompany the friend-connection on CouchSurfi ng.Besides, it would be interesting to research the question of identity construction comprehensively, comparing and examining profi le information generated both by other users and by the profi le owner.Analysing a bigger sample may allow the establishing of patterns and concepts that were missed due to the limitation of the present work.

Fig. 1 .
Fig. 1.The most popular words found in the references.This tag cloud pretty well sums up the content of a typical reference left on CouchSurfi ng