I need this assignment finish ASAP and I will provide you with the two scholarly sources

Are You Feeling Lonely? The Impact of Relationship Characteristics and Online Social Network Features on Loneliness

SABINE MATOOK, JEFF CUMMINGS, AND HILLOL BALA

SABINE MATOOK is a senior lecturer of information systems at the UQ Business School, University of Queensland, Australia. She received her Ph.D. from the Technische Universität Dresden, Germany. Her research interests include the IT artifact, social media, and agile IS development. Her work has appeared in the European Journal of Information Systems, Information and Management, MISQ Executive, Decision Support Systems, Journal of Strategic Information Systems, and other journals. She has served or is currently serving as an associate editor and track chair for major information systems conferences, including the International Conference on Information Systems and European Conference on Information Systems.

JEFF CUMMINGS is an assistant professor of information systems and operations management in the Cameron School of Business at University of North Carolina, Wilmington. He received his Ph.D. from Indiana University. His research interests include the impacts of social media on the organization, health-care IT, and virtual team collaboration. His work has been published or is forthcoming in Business Horizons and Journal of the American Society for Information Science and Technology.

HILLOL BALA is an assistant professor of information systems and Whirlpool Corporation Faculty Fellow in the Kelley School of Business at Indiana University, Bloomington. He received his Ph.D. from the University of Arkansas. His research interests include IT-enabled business process change and management, IT use, adaptation and impacts, and use of IT in health care. His work has been published or is forthcoming in Information Systems Research, Journal of Management Information Systems, MIS Quarterly, Management Science, Production and Operations Management, Decision Sciences, Information Society, Communications of the ACM, MISQ Executive, and other journals. He has served or currently serves on the editorial boards of Information Systems Research and Decision Sciences, and as a track chair, an associate editor, or a program committee member of major information systems conferences, such as the International Conference on Information Systems, the Pacific Asia Conference on Information Systems, and others.

ABSTRACT: In contemporary society, many people move away from their personal networks for extended periods to reach professional and/or educational goals. This separation can often lead to feelings of loneliness, which can be stressful and sometimes debilitating for the individual. We seek to understand how a person’s

Journal of Management Information Systems / Spring 2015, Vol. 31, No. 4, pp. 278–310.

Copyright © Taylor & Francis Group, LLC

ISSN 0742–1222 (print) / ISSN 1557–928X (online)

DOI: 10.1080/07421222.2014.1001282

use of online social networks (OSNs)—technology-enabled tools that assist users with creating and maintaining their relationships—might affect their perceptions of loneliness. Prior research has offered mixed results about how OSNs affect lone- liness—reporting both positive and negative effects. We argue in this study that a clearer perspective can be gained by taking a closer look at how individuals approach their relationship management in OSNs. Building on theoretical works on loneliness, we develop a model to explain the effects of relationship character- istics (i.e., relationship orientation, self-disclosure, and networking ability) and OSN features (i.e., active or passive) on perceived loneliness. Our findings show that OSN can be linked to both more and less perceived loneliness, that is, individuals’ relationship orientation significantly affects their feelings of loneliness, which are further moderated by their degree of self-disclosure within the OSN. Furthermore, how users engage in the OSN (either actively or passively) influences their percep- tions of loneliness. Practical implications regarding perceived loneliness include recommendations for firms to encourage mobile workers to utilize OSNs when separated from others, for education providers to connect with their new students before they arrive, and for users to utilize OSNs as a social bridge to others they feel close with.

KEY WORDS AND PHRASES: social media, online social networks, loneliness, relationship management, communal orientation, social exchange theory, self-disclosure, networking ability.

In recent years, individuals have become increasingly mobile, and in doing so travel significant distances from their homes to reside at the target destination for an extended period of time [30]. Consequently, individuals are separated from the social environments that are both familiar and comforting to them [58]. Prior research in sociology has suggested that this separation leads to negative social outcomes, including feelings of loneliness [66, 82]. It is assumed that lonely people desire human attachment that can be achieved through creating new or nurturing existing relationships [7]. Prior research has produced mixed findings regarding the impact of technology,

especially online social networks (OSNs), on feelings of loneliness [78]. Whereas some studies point to a negative association between OSN use and loneliness [31], others show the opposite [19]. When loneliness is reduced because OSNs engage users in relationships with their social network contacts [31], the technology serves as an outlet that captivates and helps users to buffer their social separation [60]. In contrast, other studies have shown that OSN use increases loneliness. Some users who only consume information that others share have developed feelings of envy, emotional withdrawal, and loneliness [50]. Indeed, studies showed that individuals’ life satisfaction and well-being decreased when using Facebook intensively because they compared their lives with their impressions of others in the OSN [52]. Because of these mixed findings, it is not clear whether OSNs alleviate or exacerbate perceived loneliness, and a more comprehensive and theoretically grounded account is needed.

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 279

The current study seeks to address this gap by focusing on the primary purpose of OSNs as relationship tools. OSNs are uniquely suited for managing relationships because users can virtually replicate their social network of relationships [34]. However, not everybody uses the OSN in the same way nor do they have the same approach to how they manage their relationships. We thus propose that (1) a user’s OSN feature use, (2) the user’s relationship orientation moderated by their degree of self-disclosure, and (3) a user’s networking ability have an impact on their loneliness. As the overarching theory for this study, we draw on the loneliness literature and insights from theories of social exchange, communication, and poli- tical skills to develop our model, and we test the model in the context of students’ use of OSNs. While all three theories center around human relationships and are core parts of human interactions, social exchange theory from the relationship literature enables us to theorize about relationship norms and how benefits between relationship partners (e.g., OSN users) are given and received. Communication theories are used to explain different forms of communication—either direct or via broadcasting—whereas the management theory on personal influences and political skills illuminates the OSN user’s ability to create and maintain relationships. This paper contributes to information systems (IS) research by enhancing our

understanding of OSN use and the impact it has on users’ perceived loneliness. This research expands the relationship literature by examining a user’s approach to relationship management in OSNs and extends theory in social psychology, in particular the body of knowledge of loneliness. We further add to the theoretical understanding of on self-disclosure in virtual settings by proposing it as a moderator between relationship orientation and loneliness. Finally, we contribute to manage- ment research by demonstrating the importance of networking for an OSN user to ease feelings of loneliness. This research also provides practical insights about how organizations, such as firms and education providers, can help alleviate loneliness experienced by those they are responsible for.

Background

This section presents prior research on OSNs and perceived social loneliness that is relevant to our study.

Online Social Networks

An OSN, such as Facebook or Google+, is a web-based technology that allows users to exchange social information with others who may be near or far, including friends, family, colleagues, and teammates [56]. Kane et al. [46, p. 279] summarized the unique characteristics of OSNs by describing how they enable users to “(1) have a unique user profile that is constructed by them, members of their network, and the platform; (2) access digital content through and protect it from various search mechanisms provided by the platform; (3) articulate a list of other users with

280 MATOOK, CUMMINGS, AND BALA

whom they share a connection; and (4) view and traverse their connections and those made by others within the system.” An OSN’s relationship management capability is enabled through features avail-

able in most OSNs, including the ability to list contacts, share social information (including photos, videos, text via microblogging [79]), send/receive private mes- sages, social search engines, and express support to others (e.g., “Likes” on Facebook) [11]. Through these features, user-generated content is created that facil- itates both active engagement with one’s network and passive consumption of content [64]. Social search engines help in finding contacts as well as filtering user-generated content [28]. In OSNs, self-disclosure is a common behavior [69]. Self-disclosure is defined as

“any message about the self that a person communicates to another” [87, p. 338]. In OSNs, disclosed information includes personal details such as interests, preferences, relationship status, and habits [85]. People self-disclose to others they like and trust, allowing for a relationship to become more intimate [59]. Reciprocal self-disclosure leads to a “you tell me, I tell you” behavior whereby disclosed information increases in depth and breadth [54, p. 170]. Indeed, research on bloggers’ online behavior has demonstrated the importance of reciprocity for knowledge sharing [17]. Research has shown that self-disclosure plays an important role in relationship growth, leading to a closer, high-quality relationship, especially for friendships [15]. Consequently, OSNs are apt for relationship creation (i.e., to form new and to revive neglected relationships) and relationship maintenance (i.e., to nurture and foster existing relationships) to impact perceived loneliness [70].

Perceived Social Loneliness

Loneliness results from a perceived absence of satisfying relationships and a deficit in an individual’s social network [39, 66]. According to the belongingness hypoth- esis, human beings have “a pervasive drive to form and maintain at least a minimum quantity of lasting, positive, and significant interpersonal relationships” [7, p. 497]. The loneliness literature differentiates between two types of loneliness: emotional

and social [86]. Emotional loneliness is associated with a lack of intimate ties and a deficit in intimate attachments, particularly in romantic relationships. Conversely, social loneliness “results from the lack of a network of social relationships in which the person is part of a group of friends who share common interests and activities” [74, p. 1314]. Given that we are interested in the perceived loneliness of people who are geographically separated from their familiar social network, we focus on social loneliness. Social loneliness is thought to be the result of social isolation because an indivi-

dual is not able to have repeated interactions with the same contacts [7]. The network of lonely individuals tends to be smaller, which gives them the impression that they do not belong to a group [39]. Research reports that social loneliness comes from infrequent interactions with friends [27], as well as less supportive behaviors

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 281

and unhelpfulness from one’s networks in times of need [80]. To compensate for feelings of social loneliness, prior research stipulates active relationship manage- ment [86].

Hypothesis Development

This section presents hypotheses on the impacts of relationship characteristics (relationship orientation, self-disclosure and networking ability) and OSN features (active and passive features) on social loneliness. Our theoretical foundation is based on human relationships and its core focuses on human interactions. We therefore build on the relationship literature using social exchange theory to theorize about self-disclosure in relationships as well as norms and benefits given and received between relationship partners (e.g., OSN users). We use communication theories to explain different forms of communicating via OSN networks. Furthermore, we use a management theory on personal influences and political skills to explain an OSN user’s ability to proactively create and maintain interpersonal relationships. The research model is presented in Figure 1 at the end of the section.

Relationship Orientation

Two social exchange theorists, Clark and Mills [21, 23], have shown that different norms govern a person’s behavior when creating and maintaining a relationship. These norms affect individuals’ orientations toward relationships and their under- standing of how benefits are given and received. In general, individuals maintain relationships only when the comparison between given and received benefits is perceived to be satisfactory, but individuals differ in how they judge the extent of

Exchange Orientation

Relationship Orientation

Communal Orientation

Self-Disclosure

Social Loneliness

Networking Ability

Passive Features

OSN Features

Active Features

H 1a

H 4a

H 3

H 2bH 2a

H 1b

H 4b

Broadcasting Direct

Communication H 4c

Figure 1. Research Model

282 MATOOK, CUMMINGS, AND BALA

reciprocity required [26]. This leads to two different orientations toward relation- ships: an exchange orientation and a communal orientation [21]. According to Clark and Mills, these two relationship types represent two distinct concepts and not a continuum that can vary in strength. An individual with an exchange orientation is concerned with equal reciprocity

and maintains a relationship with others only for instrumental reasons [22]. An exchange relationship orientation is characterized by giving benefits “with the expectation of receiving a comparable benefit in return or as repayment for a benefit received previously” [21, p. 684]. These individuals carefully record obligations and keep score of “give and take.” In contrast, a communal relationship orientation is characterized by giving benefits “in response to needs or to demonstrate a general concern for the other person” [21, p. 684]. A communal-oriented individual gen- erally has no expectations of immediate repayment of a supplied benefit, but shows a concern for the other’s welfare [22]. While these two types of relationship orienta- tion are general in nature, they can be translated into an OSN environment. In such a context, they manifest as different social-emotional benefits that stem from the user- generated content. The benefits one can give and receive in OSNs include an initial posting on a contact’s profile page, responding to a posting (textual or via a “like it” function), gift giving, sending private messages, and initiating chats. Communal relationship orientation: Within OSNs, we argue that users with a

communal relationship orientation behave in such a way that results in lower degrees of perceived social loneliness. These users undertake OSN activities to please others and without expectation of immediate repay. For example, a communal-oriented user would post a comment on a contact’s OSN profile (i.e., give a benefit) because he/ she sees the contact is in need or he/she cares about that contact (e.g., when the contact posted about a lost wallet or a canceled flight). As these OSN postings are undertaken without the expectation of receiving anything in return, the communal- oriented person might post on a contact’s profile multiple times without receiving a comment back. Posting continuously when a need is observed (i.e., providing the benefit), however, stimulates reciprocity because the receiver may eventually return the comment. In doing so, the interaction frequency between the two OSN contacts increases and the user feels more integrated in the social network. Simply engaging in the act of giving may also lead the communal-oriented user to feel more connected. For both reasons, perceptions of loneliness should decrease, and we thus expect that users who are more communal-oriented will feel less lonely.

Hypothesis 1a: A user’s communal relationship orientation is negatively asso- ciated with perceived social loneliness.

Exchange relationship orientation: Within OSNs, we argue that users with an exchange-relationship orientation exhibit behaviors that result in increased perceived loneliness. Exchange oriented individuals expect their OSN activity to be recipro- cated equally and on a timely basis. For example, if the user posts on a contact’s profile, a returning post is soon expected. If the user gives a virtual gift, then the

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 283

expectation is that the other returns a gift. If the gift is not returned, the exchange- oriented user refrains from giving a gift to this contact again [77]. Contacts who fail to reciprocate may even be removed from the user’s OSN network [9]. Because of an exchange-oriented user’s “scorekeeping,” the user knows of the failed reciprocity, and would stop giving benefits to contacts who do not reciprocate. In this case, the number of contacts with whom the user interacts becomes smaller, resulting in the user being more disconnected from the social network. Instead of feeling integrated, an exchange-oriented individual may feel isolated [16] and thus experience social loneliness [83]. Consequently, we would expect an OSN user who only gives benefits in expectation of equal and timely reciprocity to experience a higher degree of perceived social loneliness.

Hypothesis 1b: A user’s exchange-relationship orientation is positively asso- ciated with perceived social loneliness.

Self-Disclosure as a Moderating Factor

Prior research has studied the relationship between self-disclosure and loneliness, with ambiguous results. Some studies report that self-disclosure affects loneliness [53, 73], whereas other studies suggest the reverse [3, 45]. Despite mixed findings, these studies agree that self-disclosure is important because of its role in facilitating the deepening of relationships. We draw on this insight to argue that self-disclosure could be an important moderator in our study, in that it could alter the way in which a user’s relationship orientation affects loneliness. For a communal-oriented OSN user, we expect that higher self-disclosing behavior

will reduce perceived loneliness by strengthening the relationship between commu- nal orientation and loneliness. When a communal-oriented OSN user increases disclosure of social information, more opportunities for the user’s contacts to reciprocate emerge. As a result, the extent of the returning self-disclosure also increases [34], albeit not necessarily proportional to the giving. Because the com- munal-oriented user is not concerned with equal reciprocity, a lack of it would therefore not stop the user’s future self-disclosure, that is, the communal-oriented user would continue disclosing social information. Indeed, as communal-oriented individuals are more focused on giving than on receiving, and thus, an increase in giving would make the user feel close to his/her network, especially as no “scores” are kept on how frequently the network reciprocates. Moreover, continuing disclo- sure will most likely result in more reciprocated self-disclosure that will in turn diminish the communal-oriented user’s loneliness. Consequently, in the presence of increased self-disclosure, the negative relationship between communal orientation and perceived social loneliness is strengthened.

Hypothesis 2a: Self-disclosure moderates the relationship between communal- relationship orientation and perceived social loneliness such that the relation- ship becomes stronger when there are higher levels of self-disclosure.

284 MATOOK, CUMMINGS, AND BALA

For an exchange-oriented OSN user, we expect that a higher degree of self- disclosure will strengthen the relationship between exchange orientation and lone- liness, that is, more self-disclosing behavior for this type of user will result in more feelings of loneliness. An exchange-oriented OSN user expects higher self-disclo- sure to receive, in return, disclosed information of the same extent (e.g., based on novelty or interestingness, media richness or message length). However, recipients of disclosed information may only be a small number of OSN contacts—those who have proved to reciprocate. Because these OSN contacts are already having to reciprocate the “normal” information (e.g., postings made or photos sent), sending them additional information (through self-disclosure) could result in information overload [35, 48]. To ease this overload, these contacts may well ignore or hide information, further reducing reciprocity [9]. The exchange-oriented user is likely to feel that his/her calls are falling on increasingly deaf ears. This effect is likely to increase feelings of disconnectedness (i.e., reduced belongingness) that in turn stimulate perceptions of social loneliness [83]. Consequently, in the presence of higher self-disclosure, the positive relationship between exchange orientation and perceived social loneliness is strengthened.

Hypothesis 2b: Self-disclosure moderates the relationship between exchange- relationship orientation and perceived social loneliness such that the relation- ship becomes stronger when there are higher levels of self-disclosure.

Networking Ability

Networking is the proactive creation and maintenance of interpersonal relationships with the objective of leveraging these relationships at some point [29]. Networking ability is defined as an individual’s “capacity to identify and develop a diverse group of contacts” [55, p. 691]. Building on the theory of political skill [36], networking is a human skill leveraged for understanding and influencing others in professional settings to achieve personal and organizational objectives. Individuals with a strong networking ability find it easy to develop friendships, alliances, and coalitions [36]. Furthermore, these individuals master, effortlessly, the creation and maintenance of large and diverse networks to take advantage of the opportunities that emerge from these relationships. Yet, networking ability is just a personal trait; rather, it represents a social resource and an informational asset that stems from having access to the social network [10]. Networking ability is also an important factor in OSNs because it allows users to

create a large OSN network [34]. Users skilled in networking can also employ the OSN’s communication features to interact with their contacts and thereby deepen existing relationships [42]. These contacts in the OSN can provide social support (e.g., expressing concern and sharing news), and contribute to feelings of belonging and lower feelings of social loneliness. Individuals who feel integrated within a social network show reduced feelings of social loneliness because their desire for

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 285

attachment is satisfied [32]. When lacking networking ability, the OSN user may wait for others to initiate an OSN interaction instead of proactively approaching them. Missing out on interaction opportunities, a user with lower networking ability may feel disconnected, which is associated with feelings of loneliness [72]. Consequently, OSN users who possess the ability to proactively create and maintain interpersonal relationships experience lower feelings of social loneliness.

Hypothesis 3: Networking ability of an individual is negatively associated with perceived social loneliness.

Use of Active and Passive OSN Features

OSNs are used to create and consume user-generated content [47]. Content is produced through active engagement with a user’s contacts, for example, in the form of status updates or by sharing photos, videos, or links. Communication theorists differentiate active engagement as direct communication (one-to-one) and broadcasting (one-to-many) [12]. In addition, viewing user-generated content is referred to as “passive consumption,” also known as “social surveillance” [45]. The rich architecture of OSNs provides features for performing both active engage- ment and passive consumption [11]. OSN users experience social isolation when they passively consume user-gener-

ated content because of their lack of interaction [51]. Furthermore, passive con- sumption restricts individuals in creating and managing relationships [66]. The literature has linked passive consumption to feelings of disconnectedness and lone- liness [4]. For instance, passive content consumption on Facebook creates feelings of envy that reduce a user’s life satisfaction through social comparison between the user and their contacts [50]. These researchers showed that reading about the travel and leisure experiences of OSN contacts led users to feel envy, dissatisfaction, and loneliness. Consequently, we propose that the use of OSN features for observing others (i.e., passive consumption) creates perceptions of social loneliness because users do not engage with their OSN network. By only using passive OSN features, the user misses out on interactions that could create a sense of belonging.

Hypothesis 4a: The use of passive OSN features is positively associated with perceived social loneliness.

The creation of user-generated content and, as such, the active engagement with one’s OSN network, stimulates mutual content sharing among OSN contacts [69]. Active engagement is achieved through the use of active OSN features, where repeated interactions strengthen a user’s social integration and create a sense of belonging [7]. The enhanced sense of social belonging is related to feeling lower degrees of social loneliness [31]. Active OSN features that use either direct com- munication or broadcasting “may have dramatically different outcomes” on lone- liness perceptions [12, p. 572] because of the number of recipients, and thus the number of potential interaction encounters.

286 MATOOK, CUMMINGS, AND BALA

In direct communication, the user communicates with one or more recipients whereas in broadcasting no specific recipient exists and the content is shared with a larger audience. This greater number of recipients provides more opportunities for social interactions than direct communication. Nevertheless, actively engaging with one’s network contacts creates a sense of belonging and it is expected to result in lower degrees of social loneliness.

Hypothesis 4b: The use of active OSN features that facilitate broadcasting is negatively associated with perceived social loneliness.

Hypothesis 4c: The use of active OSN features that facilitate direct communica- tion is negatively associated with perceived social loneliness.

Methodology

Participants

For the study, OSN users were recruited from a master’s program in a business school at a major Australian university. We invited students from a large manage- ment information systems (MIS) course in which 205 students were enrolled. Prior studies on social media have repeatedly used student samples to test their hypothesis [see, e.g., 49, 54]. Students are particularly appropriate sample subjects for social media research because they represent the typical OSN user population based on age and gender [56]. Our decision to use a student sample is further supported by criteria and recommendations put forward by Compeau et al. [24] who posit that one can generalize from student samples when the intended population is clearly identified and a rationale for the use of students is provided. The sample of our study includes 61 percent females with an average age of 25 years.

Participants had substantial experiences with OSNs and had an OSN account for 5.5 years on average. Among the participants, 62 percent logged in at least once a day and 64 percent spent at least 30 minutes a day in their OSN. On average, participants spent about 55 minutes per day at their OSN. The OSN sites varied across participants, with Facebook being the primary site used (32 percent), followed by two Chinese OSNs, RenRen (27 percent) and QQ (20 percent), and various other OSNs (21 percent). The participants were first-semester students, mainly international students (92

percent Asian, 5 percent Australian, 1.5 percent European, and 1.5 percent South American) who had arrived in Australia two to three weeks prior to the data collection. Thus, the majority of students had recently been taken out of their familiar environment and, therefore, the country and the university were unfamiliar to them. Hence, we assume that these students were experiencing at least some level of social loneliness. Our assumption was supported by a research study reporting that social loneliness is a concerning but common phenomenon among university students [27]. We also believe that the students used their OSN for relationship management and to maintain relationships with their familiar network of people from their home country. It should be noted that the majority of students (66 percent)

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 287

came from China where access to Facebook is blocked. This number corresponds roughly to the percentage of participants who indicated that RenRen and QQ are their main OSNs. This is not a problem for our study, however, because we do not study Facebook specifically, but rather the features available in Facebook and other OSNs. We only refer to Facebook to introduce OSN features because it is a well- known OSN.

Data Collection

Participants were provided with an online questionnaire at the beginning of the study that included instructions to answer the questions regarding past OSN experience. Participants were informed that the questions concerned their OSN usage behavior. Specifically, they were asked to evaluate the questions in the context of how they currently participate in OSN. We undertook three rounds of data collection as illustrated in Figure 2. To facilitate the

tracking of participant responses over time, each participant was given a unique identification number. Participation in the study was voluntary. Responses varied cross the different data collection waves. Of the 205 students, 185 (20 missing responses), 178 (27 missing responses), and 169 (36 missing responses) students responded to the questionnaire at T1, T2, and T3, respectively. Only one student dropped the course and was removed from the sample. After deleting responses of those who did not participate all three times, we arrived at a final sample of 166 participants (final response rate 81 percent). Various measures were utilized to incentivize participation to achieve a high

response rate and to minimize dropouts, including prenotification of the data collec- tion, endorsement of the research project by the lecturer, participation reward through nonfinancial tokens, and the assurance of privacy and anonymity for all participants. We kept the survey length reasonably short by collecting different constructs at different time periods. For example, communal and exchange orienta- tions were collected at T3 because these two relationship orientations are relatively stable individual traits [21].

Figure 2. Data Collection Procedure

288 MATOOK, CUMMINGS, AND BALA

In survey research, common method bias has the potential to inflate the data collected [68]. As outlined in Appendix B, we followed prior research guidelines for procedural and statistical remedies to mitigate threats of common method biases [68]. Procedural remedies to address these issues included temporal, proximate, and methodological separation of the measurements. A temporal separation was achieved by collecting the independent and dependent variables at different time points. Proximate separation was achieved by distributing the questions on different pages of the online survey. A methodological separation was achieved by mixing different scales (Likert scale or binary) throughout the questionnaire [68]. Our statistical remedies relate to three different statistical analyses, such as

Harman’s single factor test, a partial correlation procedure (e.g., marker variable technique), and controlling for effects of an unmeasured latent method factor (i.e., single-common-factor method). All three tests, as detailed in Appendix B, produced results that suggest no major threat of common method bias. In particular, our Harman’s single factor test did not reveal a single factor with more than 50 percent variance explained. Further, the partial correlation approach did not indicate any significant correlations. Finally, the single-common-factor-method results did not significantly load on an unobserved method factor. Overall, the results of the procedural and statistical remedies suggested that common method bias was not a major concern.

Measures of Survey Constructs and Control Variables

For the survey, we adapted existing scales from the literature wherever possible. All constructs were multi-item measures with fixed answer categories. Except for the OSN feature-related constructs, which were operationalized as formative constructs, all the other constructs were operationalized as reflective constructs (see Appendix A). For the reflective constructs, shorter scales with fewer indicators were used. One of the important characteristics of reflective indicators is their substitutability [67]. In other words, an indicator may substitute another indicator of the same construct, and hence allow shorter scales. In addition, prior research has suggested that it is often practical and psychometrically viable to include a short version of a scale to manage the length of a survey [8]. In the IS literature, Venkatesh et al. [84] used short scales to operationalize the constructs. The important consideration is construct validity and reliability. If a reflective construct demonstrates psychometrically acceptable characteristics using a short scale, it is often useful to keep the short scale to reduce the length of the survey. In our study, we found strong psychometrical properties (e.g., reliability, factor loadings, convergent validity, and discriminant validity) for our constructs using a short scale during the pilot study. Hence, consistent with prior research, we kept the short scale for operationalizing the constructs in our study. Perceived social loneliness: This construct was measured as a reflective construct

via seven items of the Revised UCLA Loneliness Scale [75], which has become the

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 289

standard scale to determine social loneliness for younger populations [76]. The scale ranged from 1 (never) to 5 (very often). Relationship orientation: This was measured as two reflective, independent con-

structs which is consistent with the way that communal and exchange relationship orientation are measured in the literature [21]. First, we measured communal orien- tation via a three-item scale based on the original communal orientation scale [23]. Second, we measured exchange orientation via a three-item scale based on the revised versions of the exchange orientation scale [61]. The scale for both subcon- structs ranged from 1 (strongly disagree) to 5 (strongly agree). Networking ability: This construct was measured as a reflective construct via four

items of the political skill inventory [36]. The scale ranged from 1 (strongly disagree) to 7 (strongly agree). OSN features: This variable was measured as three separate and independent

constructs, which is consistent with the way that prior research distinguishes these different forms of communication. In particular, we differentiate the degree to which OSN users actively create and passively consume user-generated content that man- ifests as different forms of communication behavior. To determine an OSN user’s communication behavior, we listed a number of OSN activities. The choice of activities translates as either active or passive OSN feature use. We measured the use of active OSN in the form of broadcasting and direct communication via three items each based on [12, 13]. We measured the use of passive OSN features via five item that were also based on [12, 13]. Participants indicated their frequency of each activity; the scale ranged from 1 (never) to 5 (every time). These three subconstructs were measured as formative constructs. Self-disclosure: This was measured as a reflective construct with a five-item sub-

scale of the self-disclosure index [59]. On a scale from 1 (no information) to 5 (very detailed information), the degree of information disclosed on their OSN was rated. Control variables: Studies have shown that the degree to which users feel competent

to use computers in diverse situations impacts their usage behaviors [57]. It is reason- able to assume that how competent users feel using OSNs may impact their perceived social loneliness. Thus, computer self-efficacy was introduced as a control variable. It was measured as a reflective three-item subset of a construct developed by Compeau and Higgins [25]. The scale ranged from 1 (strongly disagree) to 7 (strongly agree). We also included age, gender, and home country as control variables. A preliminary survey was first pilot tested for comprehensiveness, clarity, lan-

guage usage, and face validity with a small sample of 20 OSN users and 3 experienced researchers in information systems, as recommended by Churchill [20]. The pilot test identified that some questions were difficult to understand and subsequently, wording was changed prior to launching the main study. For example, we had originally phrased all questions about Facebook only assuming that this is the major OSN for our target participants. However, this was not the case and we revised the questions to “on Facebook or the social networking site that you use the most.” Furthermore, we provided additional instructions in our surveys to ensure that participants knew to answer questions in the context of their OSN use. We also

290 MATOOK, CUMMINGS, AND BALA

revised the scales for several constructs from the traditional Likert scale (strongly disagree to strongly agree) to other scales that were more appropriate for those constructs. For example, we used a five-point extent scale (no information to very detailed information) to measure self-disclosure. Finally, the networking ability scale had the qualifier “at work.” We removed this qualifier and adapted the items to the context of OSN (see Appendix A).

Data Analysis and Results

Partial least squares (PLS), a component-based structural equation modeling (SEM) technique, was used to analyze the data. PLS provides reliable estimates for complex structural models when the sample size is not large, by placing less importance on model fit and more importance on prediction [18, 37]. Further, PLS is considered an appropriate data analytic approach for a research model such as ours that has formative construct(s) and a moderator [67]. We used SmartPLS Version 2 as our statistical software application to test the various PLS models [71].

Measurement Model

The measurement model for the reflective constructs was assessed for both reliability and validity [5]. In order to assess reliability and validity for the reflective constructs in our model, we followed the guidelines suggested by Fornell and Larcker [37], which include internal consistency reliability, convergent validity, and discriminant validity. Internal consistency reliabilities (ICRs) were evaluated using Cronbach’s alpha to ensure that model items reliably measured the proposed constructs. ICRs were greater than the recommended value of .70 for all constructs at all time periods [62] (see Table 1). From these results, we can assume acceptable internal consistency of our measures. Convergent validity was assessed using indicator loadings and average variance

explained (AVE) (see Table 1). As suggested by Hair et al. [40] and Bagozzi [6], indicator loadings were examined to confirm that factor loadings were greater than .70 on their intended constructs with minimal (less than .30) cross-loading on other constructs. Indicator loadings for our model were greater than .70 for all constructs at all time periods (except for LONL3 being slightly below .70), with cross-loadings being lower than .30, suggesting convergent validity of the structural model. In addition to indicator loadings, AVE was also examined to ensure that the variance explained by the construct is higher than variance from measurement error [37]. AVE for all constructs either met or exceeded .50, further suggesting convergent validity for the current model. Given these results, convergent validity can be assumed for the proposed model. We assessed discriminant validity by examining the square roots of the shared

variance between the constructs and their measures (see Table 2). The diagonal elements are the square root of the shared variance between the constructs and their

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 291

T ab le

1 . Q u al it y cr it er ia

o f th e co n st ru ct s

L at en t v ar ia b le

It em

s M ea n

S D

F ac to r lo ad in g s

an d w ei g h ts

C ro n b ac h ’s al p h a

C o m p o si te

re li ab il it y (I C R )

A V E

C o m m u n a l o ri e n ta tio

n C O M M 1

3 .3 2

.9 7

0 .7 7 **

.6 6

.8 0

.5 8

C O M M 2

3 .1 6

.9 6

0 .7 4 **

C O M M 3

3 .3 0

.8 7

0 .7 6 **

E xc

h a n g e o ri e n ta tio

n E X O R 1

3 .0 8

.9 5

0 .8 4 **

E X O R 2

2 .9 1

.8 7

0 .7 0 **

.6 4

.7 8

.5 4

E X O R 3

2 .5 6

.8 4

0 .6 6 **

N e tw o rk in g a b ili ty

N E T A 1

4 .2 4

1 .2 8

0 .7 6 **

N E T A 2

4 .3 8

1 .3 2

0 .9 0 **

.8 6

.8 8

.6 6

N E T A 3

4 .6 2

1 .3 4

0 .8 5 **

N E T A 4

4 .6 4

1 .3 5

0 .8 5 **

P a ss

iv e O S N

fe a tu re s

P A S F 1

3 .1 7

.9 0

0 .7 1 ** ^

P A S F 2

2 .5 8

.8 3

0 .7 4 ** ^

— —

—

P A S F 3

2 .4 0

.8 4

0 .4 3 ** ^

P A S F 4

2 .5 5

1 .0 1

0 .2 2 ** ^

P A S F 5

2 .8 2

.9 2

0 .2 5 ** ^

A C T F 1

2 .5 8

.8 3

0 .9 2 ** ^

A ct iv e O S N

fe a tu re s:

B ro a d c a st in g

A C T F 2

2 .5 5

.8 4

0 .8 1 ** ^

— —

—

A C T F 3

2 .3 8

.9 6

0 .6 3 ** ^

292 MATOOK, CUMMINGS, AND BALA

A ct iv e O S N

fe a tu re s:

D ir e c t c o m m u n ic a tio

n A C T F 4

2 .6 3

.8 6

0 .9 7 ** ^

A C T F 5

2 .9 0

.9 4

0 .6 1 ** ^

— —

—

A C T F 6

2 .8 8

.8 1

0 .5 9 ** ^

S e lf- d is cl o su

re S E L D 1

2 .1 8

1 .0 5

0 .7 3 **

S E L D 2

2 .8 8

1 .1 3

0 .7 4 **

.8 7

.8 8

.6 1

S E L D 3

2 .5 5

1 .0 9

0 .8 1 **

S E L D 4

1 .9 6

.9 7

0 .8 7 **

S E L D 5

3 .0 1

1 .1 1

0 .7 2 **

P e rc e iv e d so

ci a l lo n e lin e ss

L O N L 1

2 .5 8

.8 8

0 .7 9 **

L O N L 2

2 .2 9

.8 7

0 .8 6 **

.8 7

.9 0

.5 7

L O N L 3

2 .5 6

.7 7

0 .6 9 **

L O N L 4

2 .4 0

.9 2

0 .7 9 **

L O N L 5

2 .6 6

.8 0

0 .7 1 **

L O N L 6

2 .4 4

.8 2

0 .7 0 **

L O N L 7

2 .4 6

.7 9

0 .7 4 **

C o m p u te r se

lf- e ff ic a cy

C S E 1

5 .4 2

1 .1 4

0 .7 9 **

C S E 2

5 .8 1

1 .2 1

0 .8 6 **

.8 4

.8 9

.7 4

C S E 3

5 .4 1

1 .3 2

0 .9 3 **

C S E 3

5 .4 1

1 .3 2

0 .9 3 **

N o te s:

F ac to r lo ad in g s an d w ei g h ts ar e fr o m

th e P L S an al y si s (* * p < .0 1 ; ^ th es e v al u es

re p re se n t it em

w ei g h ts ); IC R : in te rn al

co n si st en cy

re li ab il it y (ρ

c = (Σ λ i )2 /

[( Σ λ i )2

+ Σ iv ar (ε i) ], w h er e λ i

is th e co m p o n en t lo ad in g to

an in d ic at o r an d v ar (ε i) = 1 -λ

i2 ).

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 293

T ab le

2 . D es cr ip ti v e S ta ti st ic s an d C o rr el at io n s

M ea n

S D

1 2

3 4

5 6

7 8

9

1 . C o m m u n a l o ri e n ta tio

n 3 .2 6

.7 0

.7 6

2 . E xc

h a n g e o ri e n ta tio

n 2 .8 5

.6 2

– .1 4 *

.7 4

3 . N e tw o rk in g a b ili ty

4 .4 7

1 .1 1

.0 7

– .1 6 *

.8 1

4 . S e lf- d is cl o su

re 2 .5 1

.8 6

– .1 2 *

– .1 4 *

.2 5 **

.7 8

5 . P e rc e iv e d so

ci a l lo n e lin e ss

2 .5 2

.5 5

– .4 3 ** *

.4 3 **

– .3 7 **

– .1 3 *

.7 5

6 . C o m p u te r se

lf- e ff ic a cy

5 .5 4

1 .0 3

.1 4 *

– .0 2

.0 7

.1 2 *

– .2 2 **

.8 5

7 . P a ss

iv e O S N

fe a tu re s

2 .7 1

.7 0

– .3 0 ** *

.0 9

.2 2 **

– .0 5

.4 2 ** *

– .2 2 **

– –

8 . B ro a d ca

st in g

2 .5 1

.7 6

– .1 4 *

.0 6

.2 4 **

.1 9 **

– .1 8 **

– .0 8

.2 9 ** *

– –

9 . D ir e ct

co m m u n ic a tio

n 2 .8 0

.7 3

.1 7 **

– .1 0

.1 7 **

.3 0 ** *

– .0 8

– .1 1

.1 7 **

.3 4 ** *

– –

1 0 . G e n d e r

.6 1

.4 9

.0 9

.0 3

– .2 0 **

– .1 4 *

.0 3

.0 9

– .0 4

.0 9

.0 9

1 1 . A g e

2 4 .9 1

3 .5 6

.1 7 **

– .1 1

.0 1

.0 3

– .0 5

– .0 7

– .0 4

– .0 9

– .1 5

1 2 . C o u n tr y

.7 1

.4 5

– .0 8

.1 4 *

– .0 8

.1 2 *

.1 7 **

.0 2

.0 4

.1 8 **

.1 0

* p < 0 .0 5 , * * p < 0 .0 1 , * * * p < 0 .0 0 1 .

N o te :S q u a re

ro o t o f th e av er ag e v ar ia n ce

ex tr ac te d (A V E ) is o n d ia g o n al

el em

en ts .

294 MATOOK, CUMMINGS, AND BALA

measures; off-diagonal elements are correlations between constructs. For discrimi- nant validity, diagonal elements should be larger than off-diagonal elements [37]. We found all diagonal elements to be higher than the correlations across constructs, hence supporting discriminant validity. To determine the quality of the formative construct of OSN features, we inspected

the item weights and their multicollinearity (see Table 1). For all items, the weights are significant (p < .01). We examined multicollinearity because it can destabilize the model [67]. The variance inflation factor (VIF) statistic was used to determine whether the formative measures were too highly correlated. We did not find any major multicollinearity issues with all VIFs below the strict threshold of 3.3.

Structural Model: Hypotheses Testing

Following the examination of the measurement model, we then tested the structural model to assess the significance of the proposed hypotheses using a bootstrap procedure of 1,000 resamples [38]. As presented in Table 3, we ran three structural models to test our hypotheses: Model 1 included only the control variable, Model 2

Table 3. Structural Model Results for Perceived Social Loneliness

Predictors Model 1 Model 2 Model 3

Control variables Gender (female = 1) –.12* –.03 –.06 Age .04 .06 .09 Country of origin (China = 1, other countries = 0) .19** –.11 .04 Computer self-efficacy [T1] –.17** –.09 –.09

Direct/indirect effects Communal orientation [T3] –.34*** –.21** Exchange orientation [T3] .22** .16** Networking Ability [T1] –.32*** –.28*** Passive OSN features [T2] .24** .23** Active OSN features: Broadcasting [T2] –.14* –.12* Active OSN features: Direct communication [T2] –.05 –.04

Moderator Self-disclosure [T1] –.09 –.03

Moderating effectsa

Communal orientation [T3] × Self-disclosure [T1] –.33*** Exchange orientation [T3] × Self-disclosure [T1] .29*** R2 .07 .44 .56 ΔR2 .37*** .12**

* p < 0.05, ** p < 0.01, *** p < 0.001. aWe included other second-order interaction terms in Model 3. Given that these were nonsignificant and did not change the overall model estimates substantially, we excluded these terms from this table for brevity and parsimony.

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 295

included the main effects (H1a/b, H3, H4a/b/c),and Model 3 (H2a/b) included both main and interaction effects. To assess the incremental variance explained by the interaction terms, a comparison of the R2 between these models was conducted using the guidelines suggested in the literature when testing for interaction effects [2, 18, 43]. Following these guidelines, variables at the indicator level were mean-centered prior to creating the interaction terms. We hypothesized that communal orientation would have a negative influence

(i.e., decrease) on perceived social loneliness (H1a) and exchange orientation would have a positive influence (i.e., increase) on perceived social loneliness (H1b). As per Table 3, communal orientation had a negative influence on per- ceived loneliness (Model 3: β = –.21, p < .01) and exchange orientation had a positive influence on perceived loneliness (Model 3: β = .16, p < .01), thus supporting H1a and H1b. Self-disclosure was hypothesized to moderate the relationships between communal

and exchange orientation on perceived social loneliness such that the relation would be stronger when self-disclosure is high (H2a and H2b). We found support for both hypotheses. Perusal of Figure 3 indicates that the relationship between communal orientation and loneliness was found to be moderated by self-disclosure such that for OSN users with high self-disclosure, communal orientation had a stronger negative effect on perceived social loneliness (Model 3: β = –.33, p < .01). In other words, self-disclosure and communal orientation will work in tandem such that when both are high, individuals will feel less lonely. Alternatively, as shown in Figure 4, exchange-oriented individuals displayed greater loneliness when they exercised high levels of self-disclosure (Model 3: β = .29, p < .01). In other words, if an individual with high self-disclosure is exchange-oriented, he or she will feel even lonelier. In the presence of both, loneliness will increase. Table 3 shows that the addition of the interaction terms increases R2 significantly. Overall, the interaction model explained 56 percent of the variance in perceived social loneliness compared to 44 percent without the interaction terms, which is significantly different. Networking ability was hypothesized to negatively influence (or decrease) per-

ceived social loneliness (H3). As per Table 3, network ability had a significant negative influence on perceived social loneliness in both models (Model 3: β = –.28,

2

2.5

3

3.5

Low Communal Orientation High Communal Orientation

P er

ce iv

ed S

oc ia

l L

on el

in es

s

Low Self- disclosure

High Self- disclosure

Figure 3. Moderating Effect of Self-Disclosure on Communal Orientation and Loneliness

296 MATOOK, CUMMINGS, AND BALA

p < .001), suggesting that individuals who had a high degree of proclivity to connect with others are less likely to be lonely. Thus, H3 was supported. OSN feature use was also hypothesized to impact perceived social loneliness. We

hypothesized that the use of passive features would positively influence (or increase) loneliness (H4a). Results of Table 3 indicate that passive features used had a significant positive influence on feelings of loneliness (Model 3: β = .23, p < .01), supporting H4a. The use of active features in the form of broadcasting was hypothe- sized to negatively influence (or decrease) loneliness (H4b). We found that broad- casting had a strong negative influence on loneliness (Model 3: β = –.12, p < .05). In addition, direct communication was hypothesized to negatively influence (or decrease) loneliness (H4c). However, we found no support for active direct com- munication on perceived social loneliness (Model 3: β = –.04, ns.), rejecting H4c. Table 4 summarizes the hypothesized relationships including the path coefficients for those relationships.

2

2.5

3

3.5

Low Exchange Orientation High Exchange Orientation

P er

ce iv

ed S

oc ia

l L

on el

in es

s

Low Self- disclosure

High Self- disclosure

Figure 4. Moderating Effect of Self-Disclosure on Exchange Orientation and Loneliness

Table 4. Summary of Results as per Model 3 from Table 3

Path Path

coefficient Supported/ rejected

H1a Communal orientation → Perceived social loneliness –.21** Supported H1b Exchange orientation → Perceived social loneliness .16** Supported H2a Self-disclosure × Communal orientation → Perceived

social loneliness –.33*** Supported

H2b Self-disclosure × Exchange orientation → Perceived social loneliness

.29*** Supported

H3 Network ability → Perceived social loneliness –.28*** Supported H4a Passive OSN features → Perceived social loneliness .23** Supported H4b Active OSN features: Broadcasting → Perceived social

loneliness –.12* Supported

H4c Active OSN features: Direct communication → Perceived social loneliness

–.04 Rejected

* p < 0.05, ** p < 0.01, *** p < 0.001.

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 297

Discussion

The research aimed to determine how using an OSN can impact feelings of lone- liness. To this end, we used the literature on loneliness as our overarching theory to examine the influence of relationship characteristics (i.e., relationship orientation moderated by self-disclosure and networking ability) and active versus passive OSN features on perceived loneliness. The study shows how OSNs can be associated with both more and less perceived loneliness. Specifically, loneliness increased for individuals who were involved in passive feature use and for those who had exchange-relationship orientation and high degrees of self-disclosure. However, loneliness was reduced when a user had increased networking ability and used active OSN features for broadcasting and when a user with a communal relationship orientation had high degrees of self-disclosure. Yet, active OSN features via direct communication was not associated with feelings of social loneliness. We elaborate in the following on the theoretical and practical implications of the study.

Theoretical Implications and Contributions

This research contributes to the literature in several ways. First, the study contributes to IS research on OSN use regarding an individual’s social loneliness when sepa- rated. Prior research has produced contradicting results on the relationship between OSN use and loneliness, specifically, there is disagreement whether OSN use reduces or increases loneliness feelings [31, 52, 78]. This study suggests that the relationship between OSN and loneliness depends how OSNs are used. Many prior works treated OSNs as monolithic without much consideration that these platforms offer diverse features and functionalities [63]. We extended prior research by examining OSNs at a feature level and differentiated between active and passive features of an OSN. Our findings demonstrate that loneliness is impacted by the use of both active and passive OSN features. Loneliness is reduced when the active features related to broadcasting are used, but the use of only passive features leads to an increased level of loneliness. Using OSNs for broadcasting to all contacts facilitates sharing information with

the entire network, which in turn attracts reciprocity. As such, our study shows that features to support mass communication within OSNs can lower social loneliness by creating feelings of belonging. Broadcasting is a time-efficient approach to distribute social information, especially when the user is busy, something not uncommon in today’s fast-paced society. In contrast, our findings also show that passive content consumption expressed

through the use of certain OSN features (e.g., reading postings) increases loneliness. The fact that such passive behavior is labeled “social surveillance” strongly suggests isolation because the user takes on an observer role, hence monitoring others from a distance but carefully avoiding interaction. Finally, for direct communications, we did not find a significant impact on lone-

liness. One explanation for the insignificant results could be the misalignment of the

298 MATOOK, CUMMINGS, AND BALA

OSN design goals compared to its use. When directly communicating, OSN users commit their full attention to one person; however, the approached OSN contact may not be willing or able to reply, leaving the sent messages unanswered. Given that OSNs are designed to support a network approach, with a broadcasting functionality in mind, in which messages spread through the entire social network, the private one-to-one communication does not align with this design purpose and may lead to ineffective use of the technology [14]. Other systems, such as e-mail and chat rooms, or even the old-fashioned phone, might be better suited for one-to-one communica- tion. Further, directly communicating requires more effort to reach the same number of people compared to broadcasting, and users might feel that the benefits of direct communication do not outweigh the costs. Based on this imbalance, Thibaut and Kelley’s [81] exchange theory suggests that a user would refrain from directly communicating via OSNs and, as such, may not be able to ease perceived social loneliness. In sum, this study contributes to IS by suggesting that theorizing about OSNs and how they can affect social outcomes needs to investigate this relationship at the feature level because the different technology features allow for different outcomes. Second, we contribute to the relationship literature and especially to research on

interpersonal relationships. Prior research has indicated that people have different understandings of the degree of reciprocity in creating and maintaining relationships [21]. Our study now shows that users benefit differently from OSNs, depending on their relationship orientation. More importantly, a communal orientation is a bene- ficial characteristic because these users will naturally perform OSN-related activities that are associated with lower degrees of perceived social loneliness, if any. Hence, users with a communal orientation (e.g., where benefits are given in response to needs or to show general concern for a user’s OSN contacts) are able to create a sense of belonging that makes them feel less lonely when compared to exchange- oriented users. Although an exchange orientation can be successful in professional environments, our findings illustrate that this tit-for-tat behavior in OSNs leads to increased feelings of loneliness. The third contribution relates to the identification of self-disclosure as a moderat-

ing factor. Extensive research has examined the direct influence of self-disclosure in social media [54, 69]. In contrast, this study demonstrates the interaction effects of self-disclosure on the relationship between a user’s relationship orientation and loneliness. The impact on social loneliness for both exchange-oriented and commu- nal-oriented users becomes stronger as their self-disclosure increases. Sharing social information increases liking and leads to closer relationships [49, 69], and our findings support this for communal-oriented users. When they increase self-disclo- sure, improved social outcomes (such as reduced loneliness) are the result. However, for exchange-oriented users, increased self-disclosure does not have such positive effects. Indeed, for these users, their disclosing behavior may be perceived as excessive, and recipients experience information overload. As hypothesized, our findings may be an example showing that increased levels of self-disclosure burden the relationship between exchange orientation and loneliness. Thus, exchange-

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 299

oriented users might fare better by not increasing their disclosure because their contacts may not be able to keep up with reciprocating. Yet, if they fail to return the information, the user may discard the contact for future interactions. This means that for exchange-oriented users, prior research findings that higher self-disclosure leads to reduced loneliness do not apply. We find a similar theoretical association of “less is more” in prior research on electronic word of mouth. A study on the optimal number of online product recommendations has shown that after three recommendations, wear-out effects manifest, causing future recommendations to be ignored, even if the recommendations are still valuable [1]. Consequently, OSNs provide various possibilities for self-disclosure, but increased levels of sharing user-generated content can impose, even indirectly, detrimental effects, namely, that users feel more lonely. Fourth, we contribute to the management literature on personal influences and

political skills via networking abilities. Our results demonstrate that within OSNs the ability to establish interpersonal relationships is a crucial factor as to why people feel and likely remain lonely. In the literature on workplace influences networking ability was deemed as a key aspect to improving social capital that resides in a relationship [36]. Our study illustrates that networking abilities are also valuable for an OSN user because these relationships can be leveraged for creating and maintaining relationships resulting in feelings of belonging thereby reducing social loneliness perceptions. Fifth, this research contributes to theory in social psychology, in particular to the

body of knowledge of loneliness and the factors impacting these feelings. Loneliness has been extensively researched in psychology in offline settings and exemplary in an online context, however, this study is the first we know of that shows which factors affect perceived social loneliness in the technology-mediated environment of OSNs. Most important, our research explains loneliness from a relationship point of view whereas prior research to date has examined individual characteristics (e.g., personality or self-esteem), which are only loosely related to relationships. However, the literature on loneliness stresses the qualitative and quantitative deficits of relationships as a key reason for a person to be in a lonely state [66]. Hence, this study highlights the theoretical importance of relationship management factors to understand social loneliness perceptions of OSN users.

Implications for Practice

A number of practical implications arise from this research for firms and OSN users. For firms, our findings help to highlight how they can utilize OSNs to support their workers. Corporate managers should be mindful of how heavy travel demands can disrupt workers’ personal relationships and how they can use OSNs to reduce feelings of loneliness. In modern workplaces, travel is often unavoidable and many individuals work in remote locations or in foreign countries for an extended period of time. Thus, firms should actively seek ways to address the negative

300 MATOOK, CUMMINGS, AND BALA

implications of perceived social loneliness. Allowing and endorsing the use of an OSN can be a promising strategy. Prior research has shown that firms often use technology internally for networking, collaboration, and knowledge sharing [33]. Our study points to the potential benefits a firm can gain through the “private” use of OSN by their employees during work times. This research suggests that firms focus on usage policies that encourage positive outcomes for the employee and the firm rather than prohibiting OSN use at work [41]. Our study participants were university students, which makes our findings directly

relevant to education providers, many of which have large international student cohorts. This research suggests that a university should actively approach its stu- dents through an OSN while the students are still in their home country. The objective of the university should be to establish personal ties with the students and to connect with them before their arrival at the university. In this way, the OSN can be actively used prior to any development of social loneliness in an unfamiliar environment, and can be used further to help education providers to develop relationships once the student has arrived in the foreign country. And finally, the findings are also valuable for individuals who are the focus of the

study—hose who are physically separated and away from home. Parents, family members, couples, and close friends need to understand OSNs as a social tool that allows them to comfort and communicate with meaningful people in their lives. If OSNs can ease feelings of loneliness, these technologies should be used in a way that allows for bridging temporal and spatial separation and making the contact person feel connected, for example, by uploading photos, posting related comments, and using the “like it” option more often.

Limitations and Future Research

A number of study limitations should be discussed. First, there are limitations in regard to sampling and data collection. Data collected from university students might raise issues of generalizability to other populations that are separated from their familiar social environments. The majority of the students in this study had just moved to Australia, and thus the participants were away from home in an unfamiliar environment and we assumed that they experienced social loneliness. We believe that our findings could be generalized to other student or organizational settings, but further empirical work would be necessary. We suggest two settings that are particularly interesting and valuable to study: industries that utilize “fly-in-fly-out” workers (such as the mining industry) and industries in which staff (and often their immediate families) regularly move from base to base (e.g., a country’s armed forces). It would be valuable in such studies to control for age and life stage, the use intensity of OSNs, and whether and how social loneliness may be experienced differently. A second limitation refers to the evolving nature of OSN. The active and passive

features OSN users can employ today will most likely change in the future as

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 301

technology develops. Yet we believe the nature of such features will not change because there will always be features that support more active use or more passive use. Furthermore, we did not restrict our research to one particular OSN (i.e., participants responded to questions in relation to their main OSN site), and hence, we focused on features available across popular OSNs (e.g., Facebook, RenRen). Another study might want to explore variations in features across different OSNs and what impact the variations have on OSN use and loneliness. In the same light, one might want to develop a taxonomy of OSN features to

classify different OSNs. In doing so, OSN features can be differentiated at a more fine-grained level than our active–passive distinction. In addition, future research can examine the extent to which active and passive features impact other factors, for example, interpersonal influence (word of mouth) and the benevolence of users. Furthermore, we have exclusively focused on perceived social loneliness and dif- ferent antecedents in the virtual world. A future research study might undertake a comparison between the virtual world and the physical world to determine whether people would react differently and to what extent relationship orientation, self- disclosure, and networking ability affect perceived social loneliness. Finally, prior research has highlighted the importance of privacy in relation to OSN users’ self- disclosure [54]. Thus, future research might want to examine the extent to which privacy concerns affect a user’s relationship management, because these concerns could restrict self-disclosure or active use of OSN features to an extent that it may increase perceived loneliness.

Conclusion

This research aimed at explaining the impact of relationship characteristics and OSN features on feelings of social loneliness within an OSN. We have drawn on the literature of loneliness and integrated theories of social exchange, communication, and political skills to study the phenomenon. We find that an OSN user’s communal- relationship orientation moderated by self-disclosure, use of active OSN features, and networking ability are negatively associated with perceived social loneliness, whereas an exchange orientation and use of passive OSN features are positively associated with these feelings. Our study highlights how and why creating and maintaining interpersonal relationships in the OSN influences perceived social lone- liness. The findings enhance our understanding of OSNs and their ability to bridge distances of time and place and thus, enhance users’ sense of well-being, connect- edness, and human attachment.

REFERENCES

1. Abendroth, L.J., and Heyman, J.E. Honesty is the best policy: The effects of disclosure in word-of-mouth marketing. Journal of Maketing Communication, 19, 4 (2012), 245–257.

2. Aiken, L.S., and West, S.G. Multiple Regression: Testing and Interpreting Interactions. Thousand Oaks, CA: Sage, 1991.

302 MATOOK, CUMMINGS, AND BALA

3. Al-Saggaf, Y., and Nielsen, S. Self-disclosure on Facebook among female users and its relationship to feelings of loneliness. Computers in Human Behavior, 36 (2014), 460–468.

4. Amichai-Hamburger, Y., and Ben-Artzi, E. Loneliness and Internet use. Computers in Human Behavior, 19, 1 (2003), 71–80.

5. Bagozzi, R.P. The role of measurement in theory construction and hypothesis testing: Toward a holistic model. In C. Fornell (ed.), A Second Generation of Multivariate Analysis. New York: Praeger, 1982, pp. 5–23.

6. Bagozzi, R.P. On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16, 1 (1988), 74–94.

7. Baumeister, R.F., and Leary, M.R. The need to belong: Desire for interpersonal attach- ments as a fundamental human motivation. Psychological Bulletin, 117, 3 (1995), 497–529.

8. Bergkvist, L., and Rossiter, J.R. The predictive validity of multiple-item versus single- item measures of the same constructs. Journal of Marketing Research, 44, 2 (2007), 175–184.

9. Bevan, J.L.; Pfyl, J.; and Barclay, B. Negative emotional and cognitive responses to being unfriended on Facebook: An exploratory study. Computers in Human Behavior, 28, 4 (2012), 1458–1464. 10. Blass, F.R., and Ferris, G.R. Leader reputation: The role of mentoring, political skill,

contextual learning, and adaptation. Human Resource Management, 46, 1 (2007), 5–19. 11. Boyd, D., and Ellison, N. Social network sites: Definition, history, and scholarship.

Journal of Computer-Mediated Communication, 13, 1 (2007), 210–230. 12. Burke, M.; Kraut, R.; and Marlow, C. Social capital on Facebook: Differentiating uses

and users. In D. Tab (ed.), CHI ‘11: Annual Conference on Human Factors in Computing Systems. Vancouver: ACM, 2011, pp. 571–580. 13. Burke, M.; Marlow, C.; and Lento, T. Social network activity and social well-being. In

E. Mynatt (ed.), CHI ‘10: Annual Conference on Human Factors in Computing Systems. Atlanta: ACM, 2010, pp. 1909–1912. 14. Burton-Jones, A., and Grange, C. From use to effective use: A representation theory

perspective. Information Systems Research, 24, 3 (2013), 632–658. 15. Butler, B.S., and Matook, S. Social media and relationships. In The International

Encyclopedia of Digital Communication and Society. London: Wiley-Blackwell, 2015, pp. 1–20. 16. Buunk, B.P., and Prins, K.S. Loneliness, exchange orientation, and reciprocity in

friendships. Personal Relationships, 5, 1 (1998), 1–14. 17. Chai, S.; Das, S.; and Rao, H.R. Factors affecting bloggers’ knowledge sharing: An

investigation across gender. Journal of Management Information Systems, 28, 3 (2011), 309–342. 18. Chin, W.W.; Marcolin, B.L.; and Newsted, P.R. A partial least squares latent variable

modeling approach for measuring interaction effects: Results from a Monte Carlo simulation study and an electronic-mail emotion/adoption study. Information Systems Research, 14, 2 (2003), 189–217. 19. Chou, H.-T.G., and Edge, N. “They are happier and having better lives than I am”: The

impact of using Facebook on perceptions of others’ lives. Cyberpsychology, Behavior, and Social Networking, 15, 2 (2012), 117–121. 20. Churchill, G.A. A paradigm for developing better measures of marketing constructs.

Journal of Marketing Research, 16, 1 (1979), 64–73. 21. Clark, M., and Mills, J. The difference between communal and exchange relationships:

What it is and is not. Personality and Social Psychology Bulletin, 19, 6 (1993), 684–691. 22. Clark, M.S. Record keeping in two types of relationships. Journal of Personality and

Social Psychology, 47, 3 (1984), 549–557. 23. Clark, M.S.; Ouellette, R.; Powell, M.C.; and Milberg, S. Recipient’s mood, relationship

type, and helping. Journal of Personality and Social Psychology, 53, 1 (1987), 94–103. 24. Compeau, D.; Marcolin, B.; Kelley, H.; and Higgins, C. Research Commentary:

Generalizability of information systems research using student subjects—A reflection on our practices and recommendations for future research. Information Systems Research, 23, 4 (2012), 1093–1109.

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 303

25. Compeau, D.R., and Higgins, C.A. Computer self-efficacy: Development of a measure and initial test. MIS Quarterly, 19, 2 (1995), 189–211. 26. Cropanzano, R., and Mitchell, M. Social exchange theory: An interdisciplinary review.

Journal of Management, 31, 6 (2005), 874–900. 27. Cutrona, C.E. Transition to college: Loneliness and the process of social adjustment. In

L.A. Peplau and D. Perlman (eds.), Loneliness: A Sourcebook of Current Theory, Research, and Therapy. New York: Wiley-Interscience, 1982, pp. 291–309. 28. Dang, Y.; Zhang, Y.; Chen, H.; Brown, S.A.; Hu, P.J.H.; and Nunamaker, J.F. Theory-

informed design and evaluation of an advanced search and knowledge mapping system in nanotechnology. Journal of Management Information Systems, 28, 4 (2012), 99–127. 29. de Janasz, S.C., and Forret, M.L. Learning the art of networking: A critical skill for

enhancing social capital and career success. Journal of Management Education, 32, 5 (2008), 629–650. 30. de Silva, H.; Johnson, L.; and Wade, K. Long distance commuters in Australia: A

socio-economic and demographic profile. Thirty-Fourth Australasian Transport Research Forum, 2011 Adelaide, Australia. 31. Deters, F.G., and Mehl, M.R. Does posting Facebook status updates increase or

decrease loneliness? An online social networking experiment. Social Psychological and Personality Science, 4, 5 (2013), 579–586. 32. DiTommaso, E., and Spinner, B. Social and emotional loneliness: A re-examination of

Weiss’ typology of loneliness. Personality and Individual Differences, 22, 3 (1997), 417–427. 33. Durcikova, A., and Gray, P. How knowledge validation processes affect knowledge

contribution. Journal of Management Information Systems, 25, 4 (2009), 81–107. 34. Ellison, N.B.; Steinfeld, C.; and Lampe, C. The benefits of Facebook “friends”: Social

capital and college students’ use of online social network sites. Journal of Computer-Mediated Communication, 12, 4 (2007), 1143–1168. 35. Eppler, M.J., and Mengis, J. The concept of information overload: A review of literature

from organization science, accounting, marketing, MIS, and related disciplines. Information Society, 20, 5 (2004), 325–344. 36. Ferris, G.R.; Treadway, D.C.; Kolodinsky, R.W.; Hochwarter, W.A.; Kacmar, C.J.;

Douglas, C.; and Frink, D.D. Development and validation of the political skill inventory. Journal of Management, 31, 1 (2005), 126–152. 37. Fornell, C., and Larcker, D.F. Evaluating structural equation models with unobservable

variables and measurement error. Journal of Marketing Research, 18, 1 (1981), 39–59. 38. Gil-Garcia, J.R. Using partial least squares in digital government research. In G.D.

Garson and M. Khosrov-Pour (eds.), Handbook of Research in Public Information Technology. Hershey, PA: IGI Global, 2008, pp. 239–253. 39. Green, L.R.; Richardson, D.S.; Lago, T.; and Schatten-Jones, E.C. Network correlates of

social and emotional loneliness in young and older adults. Personality and Social Psychology Bulletin, 27, 3 (2001), 281–288. 40. Hair, J.; Black, W.; Babin, B.; and Anderson, R. Multivariate Data Analysis. Upper

Saddle River, NJ: Prentice Hall, 2010. 41. Huang, Y.; Singh, P.; and Ghose, A. A structural model of employee behavioral

dynamics in enterprise social media. Management Science. Forthcoming (2015). 42. Huffaker, D.A., and Calvert, S.L. Gender, identity, and language use in teenage blogs.

Journal of Computer Mediated Communication, 10, 2 (2005), 1–12. 43. Jaccard, J., and Turrisi, R. Interaction Effects in Multiple Regression. Thousand Oaks,

CA: Sage, 2003. 44. Joinson, A.N. “Looking at,” “Looking up” or “Keeping up with” people? Motives and

uses of Facebook. In D. Tan (ed.), Twenty-Sixth Annual SIGCHI Conference on Human Factors in Computing Systems. Florence: ACM, 2008, pp. 1027–1036. 45. Jones, W.H.; Freemon, J.; and Goswick, R.A. The persistence of loneliness: Self and

other determinants. Journal of Personality, 49, 1 (1981), 27–48. 46. Kane, G.C.; Alavi, M.; Labianca, G.J.; and Borgatti, S. What’s different about social

media networks? A framework and research agenda. MIS Quarterly, 38, 1 (2014), 274–304.

304 MATOOK, CUMMINGS, AND BALA

47. Kaplan, A.M., and Haenlein, M. Users of the world, unite! The challenges and opportunities of social media. Business Horizons, 53, 1 (2010), 59–68. 48. Koroleva, K.; Krasnova, H.; and Günther, O. “STOP SPAMMING ME!” Exploring

information overload on Facebook. In D.E. Leidner and J.J. Elam (eds.), Americas Conference on Information Systems. Lima, Peru: Association for Information Systems, 2010. 49. Krasnova, H.; Spiekermann, S.; Koroleva, K.; and Hildebrand, T. Online social net-

works: Why we disclose. Journal of Information Technology, 25, 2 (2010), 109–125. 50. Krasnova, H.; Wenninger, H.; Widjaja, T.; and Buxmann, P. Envy on Facebook: A

hidden threat to users’ life satisfaction? In: Proceedings of 11th Annual Conference on Wirtschaftsinformatik, Alt, R., and Franczyk, B., (eds.), (2013), Feb 27 – Mar 01, pp. 1477–1491, Leipzig, Germany. 51. Kraut, R.; Patterson, M.; Lundmark, V.; Kiesler, S.; Mukophadhyay, T.; and Scherlis, W.

Internet paradox: A social technology that reduces social involvement and psychological well- being? American psychologist, 53, 9 (1998), 1017–1031. 52. Kross, E.; Verduyn, P.; Demiralp, E.; Park, J.; Lee, D.; Lin, N.; Shablack, H.; Jonides,

J.; and Ybarra, O. Facebook use predicts declines in subjective well-being in young adults. PLoS ONE, 8, 8 (2013). 53. Leung, L. Loneliness, self-disclosure, and ICQ (“I Seek You”) use. CyberPsychology

and Behavior, 5, 3 (2002), 241–251. 54. Lowry, P.B.; Cao, J.; and Everard, A. Privacy concerns versus desire for interpersonal

awareness in driving the use of self-disclosure technologies: The case of instant messaging in two cultures. Journal of Management Information Systems, 27, 4 (2011), 163–200. 55. Magnusen, M.J.; Mondello, M.; Kim, Y.K.; and Ferris, G.R. Roles of recruiter political

skill, influence strategy, and organization reputation in recruitment effectiveness in college sports. Thunderbird International Business Review, 53, 6 (2011), 687–700. 56. Manago, A.M.; Taylor, T.; and Greenfield, P.M. Me and my 400 friends: The anatomy

of college students’ Facebook networks, their communication patterns, and well-being. Developmental Psychology, 48, 2 (2012), 369–380. 57. Marakas, G.M.; Mun, Y.Y.; and Johnson, R.D. The multilevel and multifaceted char-

acter of computer self-efficacy: Toward clarification of the construct and an integrative framework for research. Information Systems Research, 9, 2 (1998), 126–163. 58. Marshall, G.W.; Michaels, C.E.; and Mulki, J.P. Workplace Isolation: Exploring the

Construct and Its Measurements. Psychology & Marketing, 24, 3 (2007), 195–223. 59. Miller, L.C.; Berg, J.H.; and Archer, R.L. Openers: Individuals who elicit intimate self-

disclosure. Journal of Personality and Social Psychology, 44, 6 (1983), 1234–1244. 60. Morahan-Martin, J., and Schumacher, P. Loneliness and social uses of the Internet.

Computers in Human Behavior, 19, 6 (2003), 659–671. 61. Murstein, B.I.; Wadlin, R.; and Bond, C.F. The revised exchange-orientation scale.

Small Group Research, 18, 2 (1987), 212–223. 62. Nunnally, J.C. Psychometric Theory. New York: McGraw-Hill, 1978. 63. Orlikowski, W.J., and Iacono, C.S. Research commentary: Desperately seeking the “IT”

in IT research. Information Systems Research, 12, 2 (2001), 121–134. 64. Pagani, M., Hofacker, C.F., and Goldsmith, R.E. The influence of personality on active

and passive use of social networking sites. Psychology and Marketing, 28, 5 (2011), 441–456. 65. Pavlou, P.; Liang, H.; and Xue, Y. Understanding and mitigating uncertainty in online

exchange relationships: A principal-agent perspective. MIS Quarterly, 31, 1 (2007), 105–136. 66. Peplau, L.A., and Perlman, D. Perspectives on loneliness. In L.A. Peplau and D.

Perlman (eds.), Loneliness: A Sourcebook of Current Theory, Research, and Therapy. New York: Wiley-Interscience, 1982, pp. 1–20. 67. Petter, S.; Straub, D.W.; and Rai, A. Specifying formative constructs in information

systems research. MIS Quarterly, 31, 4 (2007), 623–656. 68. Podsakoff, P.M.; MacKenzie, S.B.; Lee, J.Y.; and Podsakoff, N.P. Common method

biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 5 (2003), 879–903.

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 305

69. Posey, C.; Lowry, P.B.; Roberts, T.L.; and Ellis, T.S. Proposing the online community self-disclosure model: the case of working professionals in France and the UK who use online communities. European Journal of Information Systems, 19, 2 (2010), 181–195. 70. Reis, H.T.; Collins, W.A.; and Berscheid, E. The relationship context of human behavior

and development. Psychological Bulletin, 126, 6 (2000), 844–872. 71. Ringle, C.M.; Wende, S.; and Will, A. SmartPLS 2.0 (M3) beta. Hamburg, Germany:

Retrieved 04 June 2013, from http://www.smartpls.de, 2005. 72. Rook, K.S. Research on social support, loneliness, and social isolation: Toward an

integration. Review of Personality and Social Psychology, 1984, pp. 239–264. 73. Rook, K.S., and Peplau, L.A. Perspectives on helping the lonely. In L.A. Peplau and D.

Perlman (eds.), Loneliness: A Sourcebook of Current Theory, Research, and Therapy. New York: Wiley-Interscience, 1982, pp. 351–378. 74. Russell, D.; Cutrona, C.E.; Rose, J.; and Yurko, K. Social and emotional loneliness: An

examination of Weiss’s typology of loneliness. Journal of personality and social psychology, 46, 6 (1984), 1313–1321. 75. Russell, D.; Peplau, L.A.; and Cutrona, C.E. The revised UCLA Loneliness Scale:

Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 3 (1980), 472–480. 76. Russell, D.W. UCLA Loneliness Scale (Version 3): Reliability, validity, and factor

structure. Journal of Personality Assessment, 66, 1 (1996), 20–40. 77. Skågeby, J. Gift-giving as a conceptual framework: Framing social behavior in online

networks. Journal of Information Technology, 25, 2 (2010), 170–177. 78. Song, H.; Zmyslinski-Seelig, A.; Kim, J.; Drent, A.; Victor, A.; Omori, K.; and Allen,

M. Does Facebook make you lonely? A meta analysis. Computers in Human Behavior, 36, (2014), 446–452. 79. Stieglitz, S., and Dang-Xuan, L. Emotions and information diffusion in social media:

Sentiment of microblogs and sharing behavior. Journal of Management Information Systems, 29, 4 (2013), 217–248. 80. Suh, A.; Shin, K.-S.; Ahuja, M.; and Kim, M.S. The influence of virtuality on social

networks within and across work groups: A multilevel approach. Journal of Management Information Systems, 28, 1 (2011), 351–386. 81. Thibaut, J.W., and Kelley, H.H. The Social Psychology of Groups. New York: Wiley, 1959. 82. Torkington, A.M.; Larkins, S.; and Gupta, T.S. The psychosocial impacts of fly-in fly-

out and drive-in drive-out mining on mining employees: A qualitative study. Australian Journal of Rural Health, 19, 3 (2011), 135–141. 83. Townsend, K.C., and McWhirter, B.T. Connectedness: A review of the literature with

implications for counseling, assessment, and research. Journal of Counseling and Development, 83, 2 (2005), 191–201. 84. Venkatesh, V.; Morris, M.; Davis, G.; and Davis, F. User acceptance of information

technology: Toward a unified view. MIS Quarterly, 27, 3 (2003), 425–478. 85. Walther, J.B.; Heide, B.V.D.; Kim, S.-Y.; Westerman, D.; and Tong, S.T. The role of

friends’ appearance and behavior on evaluations of individuals on Facebook: Are we known by the company we keep? Human Communication Research, 34, 1 (2008), 28–49. 86. Weiss, R.S. Loneliness: The experience of emotional and social isolation. Cambridge,

MA: MIT Press, 1973. 87. Wheeless, L., and Grotz, J. Conceptualization and measurement of reported self-dis-

closure. Human Communication Research, 2, 4 (1976), 338–346.

APPENDIX A: Constructs and Measurement Items

Instructions were given to participants to answer the questionnaire in the context of their OSN experience, we stated especially on the title page that the goal of the survey is to find out more about participants’ online social networks and how they use it or have used it in the past.

306 MATOOK, CUMMINGS, AND BALA

C o n st ru ct

It em

w o rd in g

B as ed

o n an d d ev el o p ed

fr o m

C o m m u n a l o ri e n ta tio

n (s tr o n g ly

d is a g re e — st ro n g ly

a g re e )

C O M M 1

I’m n o t th e so

rt o f p e rs o n w h o o ft e n co

m e s to

th e a id

o f o th e rs . (R

) C la rk

e t a l. [2 3 ]

C O M M 2

W h e n p e o p le

g e t e m o tio

n a lly

u p se

t, I te n d to

a vo

id th e m . (R

) C O M M 3

P e o p le

sh o u ld

ke e p th e ir tr o u b le s to

th e m se

lv e s.

(R )

E xc

h a n g e o ri e n ta tio

n (s tr o n g ly

d is a g re e —

st ro n g ly

a g re e )

E X C H 1

If I te ll so

m e o n e a b o u t m y p ri va

te a ff a ir s (b u si n e ss

, fa m ily , lo ve

e xp

e ri e n ce

s) , I

e xp

e ct

th e m

to te ll m e so

m e th in g a b o u t th e ir s.

M u rs te in

e t a l. [6 1 ]

E X C H 2

If I ta ke

a fr ie n d o u t fo r d in n e r, I e xp

e ct

h im

o r h e r to

d o th e sa

m e fo r m e

so m e tim

e .

E X C H 3

It d o e s n o t m a tt e r if p e o p le

I lik e d o le ss

fo r m e th a n I d o fo r th e m . (R

)

N e tw o rk in g a b ili ty

(s tr o n g ly

d is a g re e —

st ro n g ly

a g re e )

N E T A 1

I sp

e n d a lo t o f tim

e d e ve

lo p in g co

n n e ct io n s w ith

o th e rs .

F e rr is

e t a l. [3 6 ]

N E T A 2

I kn

o w

a lo t o f p e o p le

a n d a m

w e ll co

n n e ct e d .

N E T A 3

I h a ve

d e ve

lo p e d re la tio

n sh

ip s w ith

a lo t o f p e o p le

w h o m

I ca

n a sk

fo r su

p p o rt .

N E T A 4

I a m

g o o d a t b u ild in g re la tio

n sh

ip s w ith

p e o p le .

P a ss

iv e O S N

fe a tu re s (n e ve

r— e ve

ry tim

e )

P A S F 1

R e a d o th e r p e o p le ’s

st a tu s u p d a te s

B u rk e e t a l. [1 2 , 1 3 ]

P A S F 2

L o o k fo r fr ie n d s

P A S F 3

A d d g ro u p s a n d fa n p a g e s

P A S F 4

U se

a p p lic a tio

n s

P A S F 5

C h e ck

e ve

n ts

A ct iv e O S N

fe a tu re s (n e ve

r— e ve

ry tim

e )

B ro a d c a st in g

A C T F 1

U p d a te

m y st a tu s

B u rk e e t a l. [1 2 , 1 3 ]

A C T F 2

U p lo a d p h o to s a n d vi d e o s

A C T F 3

T a g p e o p le

in p h o to s a n d vi d e o s

(c o n ti n u es )

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 307

C o n ti n u ed

C o n st ru ct

It em

w o rd in g

B as ed

o n an d d ev el o p ed

fr o m

D ir e c t c o m m u n ic a tio

n A C T F 4

W ri te

o n o th e r p e o p le ’s

w a lls

A C T F 5

C h a t w ith

fr ie n d s

A C T F 6

S e n d p e rs o n a l m e ss

a g e s to

fr ie n d s

S e lf- d is cl o su

re (n o in fo rm

a tio

n — ve

ry d e ta ile d in fo rm

a tio

n )

S E L D 1

W h a t I lik e a n d d is lik e a b o u t m ys

e lf

M ill e r e t a l. [5 9 ]

S E L D 2

W h a t is

im p o rt a n t to

m e in

lif e

S E L D 3

W h a t m a ke

s m e th e p e rs o n I a m

S E L D 4

M y w o rs t fe a rs

S E L D 5

T h in g s I h a ve

d o n e th a t I a m

p ro u d o f.

P e rc e iv e d so

ci a l lo n e lin e ss

(n e ve

r— ve

ry o ft e n )

L O N L 1

I la ck

co m p a n io n sh

ip .

R u ss

e ll e t a l. [7 5 ]

L O N L 2

T h e re

is n o o n e I ca

n tu rn

to .

L O N L 3

I d o n o t fe e l a lo n e . [R ].

L O N L 4

I a m

n o t cl o se

to a n yo

n e .

L O N L 5

M y in te re st s a n d id e a s a re

n o t sh

a re d b y th o se

a ro u n d m e .

L O N L 6

T h e re

a re

p e o p le

I fe e l cl o se

to . [R ]

L O N L 7

I fe e l le ft o u t.

C o m p u te r se

lf- e ff ic a cy

(s tr o n g ly

d is a g re e — st ro n g ly

a g re e )

C o m p e a u a n d H ig g in s [2 5 ]

I a m

co n fid

e n t th a t I co

u ld

fin is h a ta sk

u si n g a co

m p u te r …

C O S E 1

if I h a d se

e n so

m e o n e e ls e u si n g it b e fo re

tr yi n g it m ys

e lf.

C O S E 2

if I ju st

h a d th e b u ilt -i n h e lp

fa ci lit y fo r a ss

is ta n ce

. C O S E 3

if so

m e o n e sh

o w e d m e h o w

to d o it fir st .

308 MATOOK, CUMMINGS, AND BALA

A P P E N D IX

B : C o n tr o ll in g fo r C o m m o n M et h o d B ia se s

T ec h n iq u es

A ct io n s ta k en

P ro c e d u ra l re m e d ie s

T e m p o ra l, p ro xi m a te , p sy

ch o lo g ic a l, o r m e th o d o lo g ic a l

se p a ra tio

n o f m e a su

re m e n t

T e m p o ra ls e p a ra tio

n : W e m e a su

re d th e ke

y d e p e n d e n t va

ri a b le

(i .e ., p e rc e iv e d so

ci a l

lo n e lin e ss

) se

p a ra te ly

fr o m

th e in d e p e n d e n t va

ri a b le s a t d iff e re n t p o in ts

in tim

e . F o r

in st a n ce

, p e rc e iv e d so

ci a l lo n e lin e ss

w a s m e a su

re d a t T 3 ; n e tw o rk in g a b ili ty , se

lf- d is cl o su

re , a n d o th e r co

n tr o lv a ri a b le

w e re

m e a su

re d a t T 1 ; a n d O S N fe a tu re s w e re

m e a su

re d a t T 2 .

P ro xi m a te

se p a ra tio

n : W e d is tr ib u te d th e q u e st io n s in

th e q u e st io n n a ir e a cr o ss

d iff e re n t w e b p a g e s.

M e th o d o lo g ic a l se

p a ra tio

n : W e u se

d d iff e re n t sc

a le s (L ik e rt sc

a le , e xt e n t sc

a le ,

b in a ry

sc a le ) fo r o u r q u e st io n s to

d im

in is h a n y re ca

ll o f in fo rm

a tio

n fr o m

sh o rt -t e rm

m e m o ry .

P ro te ct in g re sp

o n d e n t a n o n ym

ity a n d re d u ci n g e va

lu a tio

n a p p re h e n si o n .

W e in fo rm

e d th e p a rt ic ip a n ts

th a t th e ir re sp

o n se

s w o u ld

b e a n o n ym

o u s,

a ss

u re d th e m

th a t th e re

is n o ri g h t o r w ro n g a n sw

e r, a n d re q u e st e d th a t th e y a n sw

e r q u e st io n s a s

h o n e st ly

a s p o ss

ib le .

C o u n te rb a la n ci n g q u e st io n o rd e r

W e co

u n te rb a la n ce

d th e ite

m s b y ra n d o m iz in g th e m

w ith

in e a ch

su rv e y b lo ck

a n d b y

ra n d o m iz in g th e su

rv e y b lo ck

s re p re se

n tin

g d iff e re n t co

n st ru ct s.

Im p ro vi n g sc

a le

ite m s

W e u se

d p re va

lid a te d re lia b le

ite m s (s e e d is cu

ss io n o f m e a su

re m e n t) a n d p ro vi d e d

d e fin

iti o n s a n d e xa

m p le s fo r p o te n tia

lly u n fa m ili a r te rm

s.

S ta tis tic

a l re m e d ie s

H a rm

a n ’s

si n g le

fa ct o r te st

T h e H a rm

a n ’s si n g le

fa ct o r te st

in d ic a te d th a t th e re

w a s n o si n g le

fa ct o r th a t e xp

la in e d

m o st

o f th e va

ri a n ce

. T h e fir st

fa ct o r e xp

la in e d o n ly

4 6 p e rc e n t o f va

ri a n ce

. P ri o r

re se

a rc h h a s su

g g e st e d se

ve ra l lim

ita tio

n s o f H a rm

a n ’s

si n g le

fa ct o r te st .

T h e re fo re , w e co

n d u ct e d tw o a d d iti o n a ls

ta tis tic a la

n a ly se

s th a t w e d e sc

ri b e b e lo w .

(c o n ti n u es )

IMPACT OF ONLINE SOCIAL NETWORK FEATURES ON LONELINESS 309

A P P E N D IX

B : C o n ti n u ed

T ec h n iq u es

A ct io n s ta k en

P a rt ia l co

rr e la tio

n p ro ce

d u re

(e .g ., m a rk e r va

ri a b le

te ch

n iq u e )

G iv e n th a t w e d id

n o t in cl u d e a n y co

n st ru ct s th a t w e re

co m p le te ly

th e o re tic a lly

u n re la te d to

o n e o r m o re

co n st ru ct s in

o u r re se

a rc h m o d e l to

re d u ce

th e su

rv e y

le n g th , w e , fo llo w in g P a vl o u e t a l. [6 5 ], u se

d a co

n st ru ct

th a t w a s n o t p a rt o f o u r

re se

a rc h m o d e l a n d w a s w e a kl y re la te d to

o th e r co

n st ru ct s in

th e re se

a rc h m o d e l.

W e co

m p a re d th e co

rr e la tio

n b e tw e e n le n g th

o f co

m p u te r u se

a n d o th e r co

n st ru ct s

in th e st u d y a n d d id

n o t fin

d a n y m a jo r si g n ifi ca

n t co

rr e la tio

n s (i .e ., th e a ve

ra g e

co rr e la tio

n w a s 0 .1 1 ), in d ic a tin

g th a t th e re

w a s n o e vi d e n ce

o f m a jo r co

m m o n

m e th o d b ia s.

C o n tr o lli n g fo r th e e ff e ct s o f a n u n m e a su

re d la te n t m e th o d s

fa ct o r (i .e ., si n g le -c o m m o n -m

e th o d -f a ct o r a p p ro a ch

) W e d id

n o t fin

d a g o o d fit

(i .e ., ite

m lo a d in g s)

fo r th e m o d e l w h e n w e u se

d th e si n g le -

co m m o n -m

e th o d -f a ct o r a p p ro a ch

. In

p a rt ic u la r, w e cr e a te d a co

m m o n fa ct o r in

P L S

a n d a d d e d a ll th e ite

m s a s in d ic a to rs

o f th is

u n m e a su

re d la te n t m e th o d fa ct o r. T h e

ite m

lo a d in g s ra n g e d fr o m

– .4 0 to

.6 2 . W e fo u n d th a t a ll th e ite

m s h a d su

b st a n tia

lly h ig h e r lo a d in g s o n th e tr a its

(i .e ., th e ir re sp

e ct iv e co

n st ru ct s)

th a n o n th e m e th o d

fa ct o r. F u rt h e r, e xc

e p t fo r tw o in d ic a to rs

(N E T A 3 a n d N E T A 4 ), a ll th e in d ic a to rs

h a d

n o n si g n ifi ca

n t ite

m lo a d in g s (p

> .0 5 ) o n th e m e th o d fa ct o r.

310 MATOOK, CUMMINGS, AND BALA

Copyright of Journal of Management Information Systems is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.