Business

Computers in Human Behavior 27 (2011) 1152–1161

Contents lists available at ScienceDirect

Computers in Human Behavior

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c o m p h u m b e h

Why people use social networking sites: An empirical study integrating network externalities and motivation theory

Kuan-Yu Lin ⇑, Hsi-Peng Lu Department of Information Management, National Taiwan University of Science and Technology, No. 43 Keelung Road, Sec. 4, Taipei 106, Taiwan, ROC

a r t i c l e i n f o

Article history: Available online 22 January 2011

Keywords: Continued intention to use Motivation theory Network externalities Perceived benefit Social networking site

0747-5632/$ - see front matter � 2010 Elsevier Ltd. A doi:10.1016/j.chb.2010.12.009

⇑ Corresponding author. Tel.: +886 2 2737 6764; fa E-mail addresses: [email protected] (K.-Y. Li

P. Lu).

a b s t r a c t

Fast-developing social networking sites (SNS) have become the major media by which people develop their personal network online in recent years. To explore factors affecting user’s joining SNS, this study applies network externalities and motivation theory to explain why people continue to join SNS. This study used an online questionnaire to conduct empirical research, and collected and analyzed data of 402 samples by structural equation modeling (SEM) approach. The findings show that enjoyment is the most influential factor in people’s continued use of SNS, followed by number of peers, and usefulness. The number of peers and perceived complementarity have stronger influence than the number of mem- bers on perceived benefits (usefulness and enjoyment). This work also ran clustering analysis by gender, which found notable difference in both number of peers and number of members between men and women. The number of peers is an important factor affecting the continued intention to use for women but not for men; the number of members has no significant effect on enjoyment for men. The findings suggest that gender difference also produces different influences. The implication of research and discus- sions provides reference for SNS operators in marketing and operation.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction 2009). Users who propagate perceived benefit of use to their

Social networking sites (SNS) have infiltrated people’s daily life with amazing rapidity to become an important social platform for computer-mediated communication (Correa, Hinsley, & de Zuniga, 2010; Powell, 2009; Tapscott, 2008). Facebook, MySpace, and Friendster are successful examples (Kang & Lee, 2010; Lipsman, 2007; Pempek, Yermolayeva, & Yermolayeva, 2009). SNS, by defini- tion, provides a new method of communicating, employing com- puters as a collaborative tool to accelerate group formation and escalate group scope and influence (Kane, Fichman, Gallaugher, & Glaser, 2009; Pfeil, Arjan, & Zaphiris, 2009; Ross et al., 2009). The SNS innovative operation mode has not only successfully drawn the attention of industry and academia, but has also boosted user growth. SNS is currently the world’s fastest developing personal networking tool.

SNS is a cyber environment that allows the individual to con- struct his/her profile, sharing text, images, and photos, and to link other members of the site by applications and groups provided on the Internet (Boyd & Ellison, 2008; Pfeil et al., 2009; Powell, 2009; Tapscott, 2008). Hence, SNS enables users to present themselves, connect to a social network, and develop and maintain relation- ships with others (Ellison, Steinfield, & Lampe, 2007; Kane et al.,

ll rights reserved.

x: +886 2 2737 6777. n), [email protected] (H.-

friends and relatives achieve network externalities, and positive feedback gives rise to larger expansion, which increases platform members (Powell, 2009). Facebook is an obvious example. Face- book statistics indicate that its global members have rapidly in- creased from 150 million to about 350 million between January and December 2009 (Eldon, 2009). Hence, network externalities not only increase its economic benefits, but also have significant effect on expanding social network potential.

The above reveals that network externalities are an important factor affecting Internet users, a reason for people to use informa- tion technology (Gupta & Mela, 2008; Schmitz & Latzer, 2002). However, previous research has seldom studied how network externalities relate to the formation of user’s perception about SNS. In addition, as the SNS spirit emphasizes user’s interaction and involvement, users are the key to a successful website (Powell, 2009; Sledgianowski & Kulviwat, 2009). Thus, ‘‘what motives affect continued intention to use’’ becomes an important issue. Many researchers (Davis, Bagozzi, & Warshaw, 1992; Igbaria, Parasur- aman, & Baroudi, 1996; Lin & Bhattacherjee, 2008; Teo, Lim, & Lai, 1999; van der Heijden, 2004) have recently explicated individ- ual’s behavior of using information technology from the motiva- tion theory perspective. They discovered that both internal and external motivations influenced such behavioral intention, whereas, the benefits perceived by individuals was derived from the factor of motivation (Kim, Chan, & Gupta, 2007). In other words, the individual adopts information technology because

K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161 1153

he/she perceives the possibility of obtaining utility and enjoyment from it (Kim et al., 2007; Lin & Bhattacherjee, 2008; Lu & Su, 2009; Moon & Kim, 2001; Teo et al., 1999; van der Heijden, 2004).

SNS service providers need to investigate the correlation be- tween network externalities and individual motives to compre- hend the concerns of users to attract them. This study combines network externalities and motivation theory, to propose a rational research model to explain why people continue to join the SNS. Profitable online service performance depends on understanding users factors of use. Yet, few studies have investigated these fac- tors. The findings of this study could serve as a reference for SNS providers for the enhancement of the services they offer.

2. Theoretical background

2.1. Motivation theory

Previous research has widely used motivation theory to explain individual’s behavior of accepting information technology. Deci (1975) divided the motivations underlying individual’s behavior into extrinsic motivation and intrinsic motivation. Extrinsic motiva- tion refers to committing an action because of its perceived helpful- ness in achieving value (e.g., the performance of improvement), while intrinsic motivation refers to committing an action because of interest in the action itself, rather than external reinforcement (Da- vis et al., 1992).

Davis et al. (1992) found that both extrinsic (usefulness) and intrinsic (enjoyment) factors affect the motivation to use informa- tion technology systems. Later studies (Kim et al., 2007; Lin & Bhattacherjee, 2008; Lu & Su, 2009; Moon & Kim, 2001; Teo et al., 1999; van der Heijden, 2004) also found usefulness to be an extrinsic motivation, and perceived enjoyment an intrinsic motivation. These two motivations affect the individual’s intention to use information technology. Kim et al. (2007) pointed out that perceived benefit affects the individual’s use of information tech- nology, consisting of cognitive benefit and affective benefit, i.e., of extrinsic and intrinsic factors. Based on these reasons, this work proposes extrinsic benefit (usefulness) and intrinsic benefit (enjoy- ment) as the components of individual’s perceived benefit in SNS.

2.2. Network externalities

Katz and Shapiro (1985) defined network externalities as ‘‘the va- lue or effect that users obtain from a product or service will bring about more values to consumers with the increase of users, com- plementary product, or service.’’ Hence, once the scale of users reaches a critical number, external benefit emerges and attracts more users to join (Lin & Bhattacherjee, 2008). For instance, when the number of cell phone users reaches a critical mass, it generates relative benefit, providing subsequent users with more correspon- dents and a wider scope of use, as well as attracting third-party businesses (e.g., software developer) to join, which in turn bring in more users by making cell phone use easier and more conve- nient. As such, the number of users and availability of complemen- tary goods or services are factors that drive network externalities.

Many researchers (Gupta & Mela, 2008; Katz & Shapiro, 1985; Lin & Bhattacherjee, 2008) have pointed out the two types of net- work externalities: direct and indirect. Direct network externalities derive from the increase in users of a particular product or service, where user’s benefits increase. Taking online auction sites as an example, the more users that buy and sell, the more chances to choose from there are, and the higher the transaction value is (Gupta & Mela, 2008). Many researchers (Gupta & Mela, 2008; Iimi, 2005; Kim, Park, & Oh, 2008; Pae & Hyun, 2002; Wu, Chen, & Lin, 2007; Yang & Mai, 2010) have claimed that utility for users is

derived from market size, impacting the way that people use tele- com facilities, computer software, and websites. For example in the marketplace, different kinds of people can use these products, thereby increasing utility for users. This in turn encourages them to continue using these products and services. On the other hand, indirect network externalities display an increased sense of user va- lue from using a product or service, as the effect the user obtains from such product or service increases with the increase of related complementary products. The computer software spreadsheet is an example: consumers are willing to buy or use it to obtain net- work externalities rising from compatibility (Gandal, 1994). From the viewpoint of above-mentioned researchers, direct network externalities are due to the demand side of the network, while indirect network externalities are the supply side.

Findings from previous research show that researchers have distinct perspectives about sources of network externalities (as Ta- ble 1 shows). This investigation found that one single construct too often represented network externalities, a measurement incapable of reflecting sources of network externalities commonly consid- ered in the literature. A number of researchers (Gupta & Mela, 2008; Katz & Shapiro, 1985; Lin & Bhattacherjee, 2008) have indi- cated the two types of direct and indirect network externalities as sources of network externalities, thus the measurement in one sin- gle construct is insufficient. In addition, a number of researchers (Lin & Bhattacherjee, 2008, 2009) believed that the utility for users also comes from social effects. In the case of instant messaging (e.g., MSN Messenger), the more friends that join the network, the more users can maintain or develop their individual social cir- cles, thereby increasing the utility for users. Sledgianowski and Kulviwat (2009) argued that SNS is a pleasure-oriented informa- tion system that the individual becomes more willing to use as more friends or peers join (Baker & White, 2010; Li & Bernoff, 2008; Powell, 2009; Tapscott, 2008). Combining these perspec- tives, this study posits that in the context of a pleasure-oriented information system, peer network externality is one of the sources of network externalities. In summarizing the above-stated views of researchers (Baker & White, 2010; Gupta & Mela, 2008; Katz & Shapiro, 1985; Li & Bernoff, 2008; Lin & Bhattacherjee, 2008; Pow- ell, 2009; Sledgianowski & Kulviwat, 2009; van der Heijden, 2004), this study concluded that in the environment of SNS, the sources of direct, peer and indirect network externalities should all be consid- ered with regard to network externalities. Hence, these sources of network externalities were mentioned in the study to explore their effects on individual’s continued intention to use SNS.

3. Research model and hypotheses

Fig. 1 presents this study’s research model, developed based on network externalities and motivation theory. The model considers that perceived benefits and network externalities are key factors affecting individual’s continued intention to use, where the com- posing constructs of perceived benefits are extrinsic benefit (use- fulness) and intrinsic benefit (enjoyment), while for network externalities, the model considers three types of sources, namely, direct (number of members), peer (number of peers), and indirect (perceived complementarity) network externalities. The figure be- low presents the definition and hypothesis of each construct.

3.1. Perceived benefits

3.1.1. Extrinsic benefit: usefulness Davis (1989) defined usefulness as ‘‘the degree to which a person

believes that using a particular system would enhance his or her job performance,’’ when the individual feels a system is useful, he or she thinks positively about it. Many scholars (Lee, 2009; Lu,

Table 1 Previous research on network externalities.

Reference Context Source of network externalities

Gandal (1994) Spreadsheets Compatibility effects Gupta and Mela (2008) Online auction sites Market size effects, Compatibility effects Iimi (2005) Cellular phone services Market size effects Kim et al. (2008) Mobile communication service Market size effects Lin and Bhattacherjee (2009) Online social support Social effects Lin and Bhattacherjee (2008) Interactive information technologies Compatibility effects, Social effects Lou et al. (2000) Groupware Social effects Pae and Hyun (2002) Personal computer operating system Market size effects Wu et al. (2007) End user computing Market size effects Yang and Mai (2010) Online video game Market size effects

Fig. 1. The research model.

1154 K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161

Zhou, & Wang, 2009; Pontiggia & Virili, 2010; Sledgianowski & Kulviwat, 2009; Wu et al., 2007; Yen, Wu, Cheng, & Huang, 2010) have found that user’s thinking as to the usefulness of a system had great influence and positively related to adoption of informa- tion technology. An SNS user cares about whether the SNS allows him to effectively build and maintain relationships among the mechanisms that allow strangers to become acquainted and keep in touch, and that provides for the individual to form profiles and enable people to reach out toward one another (Li & Bernoff, 2008; Pfeil et al., 2009). Some scholars (Kang & Lee, 2010; Kwon & Wen, 2010; Sledgianowski & Kulviwat, 2009) have discovered that users’ perceived usefulness in SNS affects positive intention to use the SNS. Hence, this work proposes the following hypotheses:

H1. Usefulness will have a positive effect on continued intention to use of a social network service.

3.1.2. Intrinsic benefit: enjoyment Moon and Kim (2001) defined enjoyment as ‘‘the pleasure the

individual feels objectively when committing a particular behavior or carrying out a particular activity’’ and found in their study that enjoyment is a key factor for user’s acceptance of the Internet. Da- vis et al. (1992) incorporated intrinsic motivation in the discussion about Technology Acceptance Model (TAM) and believed the intrinsic enjoyment a user obtains from using computer technol- ogy to engage in work related behavior also promotes behavior intention. van der Heijden (2004) further pointed out that per- ceived enjoyment is an important factor in predicting the intention to use a pleasure-oriented information system. Many scholars

(Kang & Lee, 2010; Sledgianowski & Kulviwat, 2009) have consid- ered SNS as a pleasure-oriented information system, where users continue use with stronger motivation if they have more intense perceived enjoyment from it. Therefore, this study hypothesizes:

H2. Enjoyment will have a positive effect on continued intention to use of a social network service.

3.2. Network externalities

3.2.1. Using network externalities to explain continued intention to use Research has considered network externalities an important

factor directly affecting customer’s behavior of using information technology (Gupta & Mela, 2008; Kim & Lee, 2007; Pae & Hyun, 2002; Schmitz & Latzer, 2002; Yang & Mai, 2010). Sledgianowski and Kulviwat (2009) believed that a user intends to use an SNS once its participants reach a significant number. The number of members of SNS is similar to the install base referred to in prior net- work studies. The current study uses it to represent direct network externalities. Peer network externalities refers to the number of friends who are using, which is a major factor affecting people’s intention to join SNS, as the SNS design is for the purpose of letting the acquainted keep in touch, especially when they are allowed to share with their friends at any time (Baker & White, 2010; Li & Ber- noff, 2008; Powell, 2009; Tapscott, 2008). Hence, this work uses the number of peers to represent peer network externalities. A number of researchers (Gandal, 1994; Shurmer, 1993) have pointed out that the degree to which users perceive complemen- tary items or services (e.g., related soft- or hardware support or re- lated tutorial books) influences their intention to use such

K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161 1155

computer software. An SNS provides many complementary ser- vices for users to engage in various social applications on the web- site. One example is social games; examples of supporting tools are photo sharing, message sharing, and video sharing; an example of social activities is fan pages; and an example of friend searching tools is e-mail. These services help to increase the actual availabil- ity of complementary products perceived by users and further en- hance users’ continued intention to use (Powell, 2009; Tapscott, 2008). This study uses perceived complementarity to represent indi- rect network externalities. Summarizing the above arguments, this work proposes the following hypotheses:

H3a. Number of members will have a positive effect on continued intention to use of a social network service.

H4a. Number of peers will have a positive effect on continued intention to use of a social network service.

H5a. Perceived complementarity will have a positive effect on continued intention to use of a social network service.

3.2.2. Relationship between network externalities and perceived benefit

Some researchers (Katz & Shapiro, 1985; Lin & Bhattacherjee, 2008) have argued that network externalities affect user’s per- ceived benefits. When users use a product or service, the increase in user’s effect not only stems from the number of users, but also enhances the benefit with the increase of other compatible and complementary products or services. The best appeal of SNS is its capability of building trusted relationships outside traditional so- cial circles (Li & Bernoff, 2008; Powell, 2009; Sledgianowski & Kulviwat, 2009). Users who continue to contact their friends and those on extended SNS via SNS’s like Facebook, Myspace, and per- sonal blogs, affect more people affected, as the website gains mem- bers (Kane et al., 2009; Li & Bernoff, 2008; Sledgianowski & Kulviwat, 2009). Thus, when users perceive more members joining SNS, more people can help them become acquainted with those outside their individual network, further expanding their connec- tions (e.g., Fans page), and finding more enjoyment by interacting and sharing messages with more members (Li & Bernoff, 2008; Powell, 2009; Tapscott, 2008).

Thus, we hypothesize that:

H3b. Number of members will have a positive effect on usefulness of a social network service.

H3c. Number of members will have a positive effect on enjoyment of a social network service.

Users on most SNS normally do not aim to make new friends. Instead, they link their social networks in real life online to make further contacts (Boyd & Ellison, 2008); hence, greater numbers of peers in SNS helps to connect to more mutual friends (e.g., friend recommendation mechanism) and interaction and sharing be- tween more friends creates a greater sense of pleasure (Powell, 2009; Tapscott, 2008). Consequently, we hypothesize that:

H4b. Number of peers will have a positive effect on usefulness of a social network service.

H4c. Number of peers will have a positive effect on enjoyment of a social network service.

When indirect network externalities refer to a product or ser- vice with more complementary products or services, it creates higher benefit and more demand (Lin & Bhattacherjee, 2008). Thus, for SNS, more complementary products (e.g., supporting tools) help

users to show themselves and maintain interaction with others, thereby giving users more pleasure (Powell, 2009; Tapscott, 2008). For instance, users can take advantage of photos sharing, message sharing, and video sharing by using the supporting tools provided on the websites, to show themselves, share information, and interact with their friends in various ways. According, we hypothesize that:

H5b. Perceived complementarity will have a positive effect on usefulness of a social network service.

H5c. Perceived complementarity will have a positive effect on enjoyment of a social network service.

4. Research methods

4.1. Data collection and sampling

The current research targets subjects who are users of Taiwan Facebook (http://zh-tw.facebook.com/), currently the third most popular website in Taiwan, according to Alexa.com (Alexa, 2010) statistics. Facebook was created by Mark Zuckerberg in February 4, 2004, initially intended for use only by students at Harvard Uni- versity to share their status, photos, and other details with school- mates via a network. It gradually expanded to become an SNS across US universities until it assumed global proportions (Powell, 2009). Since June 2008, it has been providing web pages in Chinese for Taiwanese users. As such, the online platforms of Facebook in Taiwan and the US differ only in the language interface, while the services they provide are similar. Thanks to the availability of the Chinese version, the number of Facebook users in Taiwan has been growing rapidly. Statistics by CheckFacebook.com (2010) pointed out that the number of Taiwan users in September 2009 accounted for only 1.04% of the world’s users, a growth rate of 26.69%, ranking No. 1 in the world. Taiwan members exceeded 6.5 million in May 2010, making this site the largest SNS in Taiwan. Therefore, the current investigation chose it as the research subject.

Online questionnaires gathered data, distributed from January 15 to March 15, 2010 to randomly chosen Facebook users. Mean- while, this research posted messages about the questionnaire on websites of Facebook related activities and telnet://ptt.cc, the most popular bulletin board systems (BBS) in Taiwan. To encourage respondents to fill out the questionnaire, this project offered gifts, hoping to reinforce the sample return rate. Respondents’ identity was checked by their e-mail and IP address, when the question- naires were received, to avoid replications.

Four hundred and two usable responses were received, of which, male samples and female samples were nearly equal in number and the largest age group was 25–34 years, accounting for 40.5%. Regarding the educational and occupational levels, 51% of the responders were undergraduates and 53% were students. Of the three types of services that Facebook provides, communica- tion (i.e., comment and e-mail) and contents (i.e., games and news) emerged as the most frequently tried services, followed by com- merce (i.e., advertisement). Table 2 shows the detailed sample demographics.

4.2. Measurement development

The questionnaire adapted questionnaire items from previous literature; Appendix A lists the items. The scale for continued intention to use was adapted from Kim et al. (2008). The scale for usefulness was adapted from Davis (1989) and Kwon and Wen (2010). Enjoyment items were adapted from Agarwal and

Table 2 Sample demographics.

Measure Item Frequency Percentage (%)

Gender Male 204 50.7 Female 198 49.3

Age (years) Under 18 36 9 18–24 129 32.1 25–34 163 40.5 35–44 49 12.2 45–54 21 5.2 >54 4 1

Education High school or under 59 14.7 Undergraduate 205 51 Graduate degree 138 34.3

Occupation Student 213 53 Office worker 149 37.1 Self-employment 11 7.2 Home makers 29 2.7

Facebook servicesa Communications 184 45.8 Contents 197 49 Commerce 21 5.2

a Most frequently tried.

1156 K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161

Karahanna (2000) and Kim et al. (2007). As for the three constructs of network externalities, direct network externalities, which is the extent to which the number of SNS users increase, were modified from Pae and Hyun’s items (Pae & Hyun, 2002). Peer network externalities, which are the extent to which the number of friends using SNS the individual thinks will increase, were drawn up with reference to the measuring items of Lou, Luo, and Strong (2000), further modified according to the present topic. Indirect network externalities, which are the degree to which the individual thinks SNS complementary products or services will increase, were for- mulated with reference to the scale developed by Lin and Bhatt- acherjee (2008) and further modified as appropriate. All items were measured on a five-point Likert-type scale, ranging from ‘‘dis- agree strongly’’ (1) to ‘‘agree strongly’’ (5).

Both a pre-test and a pilot test were used to validate the instru- ment. The pre-test involved seven respondents, each with more than 2 years of experience using SNS. Respondents were asked to comment on the length of the instrument, the format, and the wording of the scales. The pilot test involved fifty respondents self-selected from the Facebook population. Based on respondents’ feedback at the pre-test and pilot test, several questionnaire items were modified to reflect more clearly the survey’s purpose. The reliability for all items was satisfactory (Cronbach’s alpha above 0.80) and items loaded in the correct factors in confirmatory factor analysis (with loadings of 0.60 or more). Therefore, the instrument has confirmed content validity and reliability. A list of the pilot test data is displayed in Appendix B.

5. Results

Data analysis followed the two-step approach by Anderson and Gerbing (1988), to test convergent validity and discriminant

Table 3 Fit indices for the Measurement models.

Fit Indices Recommended val

v2/df 53 Goodness of fit index (GFI) =0.9 Adjusted for degrees of freedom (AGFI) =0.8 Normed fit index (NFI) =0.9 Comparative fit index (CFI) =0.9 Root mean square error of approximation (RMSEA) 50.08

validity of the measurement model, followed by testing the re- search hypotheses and structural model framework.

5.1. Tests of the measurement model

Confirmatory factor analyses (CFA) used AMOS 7.0 for testing the measurement model. Hair, Anderson, Tatham, and Black (1998) argued that most model-fit indices should reach accepted standards before judging model fitness. Table 3 shows that every model-fit index exceeded the recommended value from previous studies, exhibiting an adequate fit to the collected data.

Reliability analysis used Cronbach’s alpha and composite reli- ability, CR, to assess the model’s internal consistency. Table 4 shows the results. The Cronbach’s alpha of each construct ranged from 0.82 to 0.91, i.e., greater than the accepted level of 0.7 recom- mended by Nunnally (1978). Every CR scored above 0.8, which ex- ceeded 0.7 suggested for CRs by Fornell and Larcker (1981), indicating good reliability and stability for the measurement items of each construct.

Convergent validity used the three standards recommended by Bagozzi and Yi (1988) to assess the measuring model: (1) all indicator factor loadings should exceed 0.5 (Hair et al., 1998); (2) CR should be above 0.7; and (3) the average variance ex- tracted, AVE, of every construct should exceed 0.5 (Fornell & Lar- cker, 1981). As Table 4 shows, the indicator factor loading of every item in the measuring model of this study exceeded 0.7. Composite reliability of constructs ranged from 0.83 to 0.91. AVE ranged from 0.59 to 0.77, therefore meeting all conditions for convergent validity.

In discriminant validity, as Fornell and Larcker (1981) sug- gested, the AVE of construct should exceed other correlation coef- ficients of the construct. Table 5 shows the matrix of correlation coefficients for all constructs in this research. Diagonal elements are the square roots of average variance extracted for the con- structs. The correlation coefficients between any two constructs are smaller than the square root of the average variance extracted for the constructs. Constructs in the measurement model of this re- search indeed are different from one another, indicating that all constructs in this research carry sufficient discriminant validity. Therefore, the measurement model in this research shows satisfac- tory reliability, convergent validity, and discriminant validity.

5.2. Tests of the structural model

The current research tested the structural model using AMOS 7.0. The model-fit indices for the structural model provided evi- dence of a good model fit (v2/df = 1.96, GFI = 0.94, AGFI = 0.91, NFI = 0.95, CFI = 0.97, RMSEA = 0.049). Fig. 3 displays the standard- ized path coefficients, path significances, and variance explained (R2) by each path, all supported by the path analysis results, except H3a, H3c, and H5a. As with variance explained (R2), R2 of continued intention to use reached 69% and that of usefulness 58%, and that of enjoyment 60%. R2 of every latent dependent variable was over 0.5, suggesting good explanatory power for the research model.

Apart from the test of total model fit and the evaluation of intrinsic model quality, when explaining the model, it is necessary

ue Suggested by authors Measurement model

Hayduck (1987) 1.86 Scott (1991) 0.94 Scott (1991) 0.92 Bentler and Bonett (1980) 0.95 Bagozzi and Yi (1988) 0.97 Bagozzi and Yi (1988) 0.046

Table 4 Statistics of construct items.

Construct Items Factor loadings t-Statistic Composite reliability (CR) Average variance extracted (AVE) Alpha

Number of members NM1 0.77 0.83 0.61 0.82 NM2 0.88 14.93 NM3 0.69 12.86

Number of peers NP1 0.85 0.86 0.68 0.86 NP2 0.81 19.13 NP3 0.81 18.29

Perceived complementarity PC1 0.79 0.85 0.59 0.85 PC2 0.83 17.23 PC3 0.74 14.80 PC4 0.70 13.81

Usefulness USE1 0.84 0.86 0.67 0.85 USE2 0.90 20.88 USE3 0.70 15.10

Enjoyment ENJ1 0.89 0.91 0.77 0.91 ENJ2 0.88 24.53 ENJ3 0.87 23.70

Continued intention to use CIU1 0.90 0.87 0.77 0.86 CIU2 0.85 20.35

Table 5 Discriminant validity.

Construct NM NP PC USE ENJ CIU

NM 0.78 NP 0.39 0.82 PC 0.31 0.53 0.77 USE 0.41 0.61 0.54 0.82 ENJ 0.39 0.64 0.57 0.60 0.88 CIU 0.40 0.63 0.54 0.58 0.70 0.88

Note: NM (number of members); NP (number of peers); PC (perceived comple- mentarity); USE (usefulness); ENJ (enjoyment); CIU (continued intention to use). Diagonal elements (bold) are the square root of average variance extracted (AVE) between the constructs and their measures. Off-diagonal elements are correlations between constructs. For discriminant validity, diagonal elements should be larger than off-diagonal elements. All correlations are significant at p < 0.01.

K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161 1157

to compare standardized direct, indirect, and total effects of the model before understanding the correlation between the variables. First, usefulness (b = 0.16, p < 0.05) and enjoyment (b = 0.44, p < 0.001) had positive direct effects on continued intention to use. Second, through usefulness (b = 0.14, p < 0.01), the number of members (direct network externalities) had positive indirect ef- fect on continued intention to use, where indirect effect was 0.02 (=0.14 � 0.16). Third, the number of peers (peer network external- ities) (b = 0.21, p < 0.01) had positive direct effect on continued intention to use and, both usefulness (b = 0.46, p < 0.001) and enjoyment (b = 0.50, p < 0.001), had positive indirect effect on con- tinued intention to use, resulting in the combined effect of 0.50 (=0.21 + 0.46 � 0.16 + 0.50 � 0.44). Finally, perceived complemen- tarity (indirect network externalities) through both usefulness (b = 0.30, p < 0.001) and enjoyment (b = 0.31, p < 0.001) had posi- tive indirect effect on continued intention to use, which indirect ef- fect was 0.18 (=0.30 � 0.16 + 0.31 � 0.44). Unfortunately, neither the direct effect of the number of members (b = 0.06, p > 0.05) nor that of perceived complementarity (b = 0.10, p > 0.05) met the significant level; this suggested that the number of members increased user’s continued intention to use only through useful- ness as a mediator. Further, we found that in the effect of the num- ber of peers on continued intention to use, the indirect effect through enjoyment was (0.22) > (0.21) of direct effect. The results of this study indicate that network externalities effectively in- crease user’s continued intention to use only through usefulness and enjoyment as a mediator.

5.3. Difference between men and women

Acceptance analysis for new information technology often uses gender difference, among all personality features (Sanchez-Franco, 2006; Venkatesh & Morris, 2000) mainly men and women have dif- ferent views for measuring value and benefit (Gefen & Straub, 1997; Venkatesh & Morris, 2000). This study thus uses the AMOS 7.0 Multi- ple-Group Analysis to understand whether male and female subjects (204 male and 198 female) have difference in the cause and effect of the model constructs in this study. The indices of fit for the two groups are consistent with the suggested values by other researchers (v2/df = 1.63, GFI = 0.91, AGFI = 0.87, NFI = 0.92, CFI = 0.97, RMSEA = 0.04), thus entitling us to verify a high degree of goodness of fit between the model and the sample data.

Fig. 3 (for men) and Fig. 4 (for women) show the estimates of path coefficients and the results of variance explained (R2) between the constructs. The results indicate that gender groups have signif- icant difference in the path ‘‘number of peers ? continued inten- tion to use’’ and the path ‘‘number of members ? enjoyment’’. On continued intention to use, both usefulness and enjoyment in men have direct influence, while enjoyment, usefulness, and num- ber of peers in women all have direct influence. On the benefit of network externalities, except for number of members in men, which does not affect enjoyment, all have significant effects. In wo- men, all three sources of network externalities significantly relate to perceived benefit. This study infers that women are more sus- ceptible to peer influence in using SNS, while men are not. Compar- isons show that men are more rational and less susceptible to empathy, so usefulness and enjoyment are more important to them.

6. Discussion

6.1. Correlations between users and network externalities, perceived benefit, and continued intention to use

This paper sheds light on why people continue to use SNS. The present research combines the network externalities theory and motivation theory to discover why people join SNS. The research results found that network externalities, usefulness, and enjoy- ment all play important roles in why people join SNS.

Fig. 2 shows the complete research results on all users. First, in the influence on user’s intention to join SNS, enjoyment has

Fig. 2. Path analysis result based on all valid samples (n = 402).

Fig. 3. Path analysis result for men samples (n = 204).

Fig. 4. Path analysis result for women samples (n = 198).

1158 K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161

K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161 1159

stronger significant effect on people’s continued use of SNS. The re- sult, consistent with studies by many researchers (Kang & Lee, 2010; Lin & Bhattacherjee, 2008; Sledgianowski & Kulviwat, 2009; van der Heijden, 2004), suggests that in the context of a pleasure-oriented information system, enjoyment plays an impor- tant role. Second, in the number of peers (peer network externali- ties), users essentially connect with old friends on Facebook (Ellison et al., 2007), and most users’ interactions via SNS are fre- quently with friends in off-line, real-world life (Pempek et al., 2009). User’s continued intention to use SNS intensifies when the user perceives many friends using SNS and anticipates more friends joining SNS in the future (Baker & White, 2010). Last, the positive influence of usefulness on continued intention to use SNS indicates that user’s continued intention to use SNS elevates when the user believes SNS upgrades the efficiency of his informa- tion sharing and connecting with others, or enables him to know more people (Kwon & Wen, 2010). The research result also ap- proves of van der Heijden (2004) perspective, that when predicting a pleasure-type information system, perceived enjoyment is an appropriate factor, whereas perceived usefulness is more for a task-oriented information system. Clearly, creating an enjoyable environment for interaction for pleasure-oriented SNS might be more effective than emphasizing utilitarian benefits.

The number of peers (peer network externalities) and perceived complementarity (indirect network externalities) are more influ- ential on extrinsic benefit (usefulness) than the number of mem- bers (direct network externalities). This finding suggests that the individual strongly believes that the breadth of his friends using SNS is great (Baker & White, 2010) or when complementary re- sources such as various supporting tools, applications, and groups of social connections are diverse (Lin & Bhattacherjee, 2008), the degree of SNS usefulness is naturally higher (e.g., broader circle of friends and more interactions). Similarly, the number of peers and perceived complementarity predict intrinsic benefit (enjoy- ment), suggesting that with increased peer connections and com- plementary tools, SNS interaction becomes more interesting. However, another index, the number of members, has no signifi- cant effect on enjoyment; this might be because SNS builds indi- vidual-centered networks and forms self-centered groups (Boyd & Ellison, 2008), where, despite a critical legion of users, it is diffi- cult to arouse an enjoyable mood in a user if he lacks development and connection with others. Hence, SNS service providers can en- hance interactions and exchanges between people with the same interests by sponsoring activities for jazzing up and further arous- ing user pleasure and fun, to intensify continued intention to use.

6.2. Correlations between gender groups and network externalities, perceived benefit, and continued intention to use

This study discovered that gender makes a notable difference in the effect of perceived benefit and network externalities on the continued intention to use SNS. Figs. 3 and 4 are the results of structural model analysis with men and women. First, in the paths of influence on continued intention to use, the number of peers has significant effect with women, but not with men. Previous research has found that women are more sensitive to other’s opinions (Venkatesh & Morris, 2000; Venkatesh, Morris, Davis, & Davis, 2003), and susceptible to the influence of their colleagues and friends to use a new technology. In contrast, men use new technol- ogies, as their task requires (Minton & Schneider, 1980). Second, in the paths of influence on enjoyment, the number of members does not have significance for men, indicating men do not feel pleasure with SNS’s with a large number of members; instead, it affects them in perceiving that expanding their own social circle is useful. This study found that of all three sources of network externalities, the number of peers has greatest influence on usefulness and

enjoyment for women, and the number of peers and perceived complementarity are most influential for men. Therefore, due to gender difference, different sources of network externalities have different influences on perceived benefit with SNS.

7. Conclusions

This study proposes an integrated theoretical framework for academic researchers by combining motivation theory and net- work externalities to investigate the attitudes and factors of user’s using SNS, and proposes possible factors of effects to understand why users continue to use. The study results suggest our research models exhibit good explanatory power to predict user’s continued intention to use SNS, providing a new direction for researchers to contemplate in subsequent research.

Social network service practitioners can draw several implica- tions from this study. First, the results suggest that enjoyment is the most important factor affecting the behavior of SNS users (Sledgianowski & Kulviwat, 2009). By enhancing users’ posting photos, films, and weblogs, and sharing links on their profiles, SNS service providers will be able to make users and their friends feel interested and have fun (Powell, 2009; Tapscott, 2008). In addition, SNS service providers should continue developing appli- cations and small games with novel, pleasurable experiences to reinforce pleasurable effects in using the site and further to strengthen its stickiness. Second, the results suggest that the num- ber of peers and perceived complementarity effectively reinforce SNS usefulness and enjoyment, providing SNS service providers with important information, that in the context of a pleasure-ori- ented information system, makes social effects an important con- struct. A user’s friends and relatives influence the level of user’s perceived enjoyment in SNS; also, through them, the user has the opportunity to meet new friends, whereby people can expand their social network (Li & Bernoff, 2008; Powell, 2009; Sledgianowski & Kulviwat, 2009; Tapscott, 2008). Practitioners should constantly incorporate and develop various activities or useful applications to allow people to reach out to each other, to reinforce user’s enjoy- ment, increase social connections, and further intensify user’s inten- tion to use, increasing SNS value. Third, the influences of different factors on continued use of information technology vary due to gen- der difference. Enjoyment is the most powerful factor affecting con- tinued intention to use SNS for both men and women. Among the reasons for attracting users to continue to use SNS, the ability to arouse inner pleasure is the crucial one. This research recommends that SNS operators develop specific applications for the demands of different genders, and promote users to have friends in their own social network joining the SNS to develop network externalities and encourage more people to use such a platform.

Despite its valuable findings and implications, this study con- tains some limitations. First, the implications are from a single study with samples in Taiwan. Therefore, research should use cau- tion when generalizing the findings to other SNS situations. Future studies should conduct research in cross-cultural and cross-mar- ketplace contexts to investigate and compare the differences in antecedents to continued intention to use. Second, this study em- ployed a quantitative statistics research model and collected data by means of an online questionnaire; it is thus difficult for repre- sented research sampling to avoid self-selection. Researchers should introduce qualitative interview methods for in-depth understanding of user’s use, to prevent online questionnaire respondents from casual answering to win prizes. Third, Facebook functions include providing various applications and socializing games. Future researchers could investigate whether the use inten- sity of social games helps expand user’s social network and analyze user’s use behavior.

1160 K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161

Appendix A

A.1. The questionnaire

A.1.1. Number of members (NM) NM1 I think a good number of people use Facebook. NM2 I think most people are using Facebook. NM3 I think there will still be many people joining Facebook.

A.1.2. Number of members Peers (NP) NP1 I think many friends around me use Facebook. NP2 I think most of my friends are using Facebook. NP3 I anticipate many friends will use Facebook in the future.

A.1.3. Perceived complementarity (PC) PC1 A wide range of applications is available on Facebook. PC2 A wide range of supporting tools is available on Facebook

(e.g., photo sharing, message sharing, video sharing). PC3 A wide range of social activities on Facebook can be joined

(e.g., fan pages). PC4 A wide range of friend-finding tools is available on

Facebook.

A.1.4. Usefulness (USE) USE1 Using Facebook enables me to acquire more information

or know more people. USE2 Using Facebook improves my efficiency in sharing infor-

mation and connecting with others. USE3 Facebook is a useful service for interaction between

members.

A.1.5. Enjoyment (ENJ) ENJ1 Using Facebook provides me with a lot of enjoyment. ENJ2 I have fun using Facebook. ENJ3 Using Facebook bores me (reversed).

A.1.6. Continued intention to use (CIU) CIU1 I intend to keep using Facebook in the future. CIU2 I intend to recommend my friends to use Facebook in the

future.

Appendix B

Table B1.

Table B1 Results of confirmatory factor analysis and reliability analysis.

Construct Items Factor loadings Cronbach‘s alpha

Number of members NM1 0.87 0.81 NM2 0.83 NM3 0.60

Number of peers NP1 0.83 0.83 NP2 0.87 NP3 0.70

Perceived complementarity PC1 0.65 0.83 PC2 0.68 PC3 0.73 PC4 0.85

Usefulness USE1 0.73 0.80 USE2 0.65 USE3 0.92

Enjoyment ENJ1 0.82 0.93 ENJ2 0.94 ENJ3 0.93

Continued intention to use CIU1 0.90 0.83 CIU2 0.78

References

Agarwal, R., & Karahanna, E. (2000). Time flies when you’re having fun: Cognitive absorption and beliefs about information technology usage. MIS Quarterly, 24, 665–694.

Alexa (2010). Top sties Taiwan. <http://www.alexa.com/topsites/countries/TW> (accessed May 2010).

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two step approach. Psychological Bulletin, 103, 411–423.

Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of Academy of Marking Science, 16, 74–94.

Baker, R. K., & White, K. M. (2010). Predicting adolescents’ use of social networking sites from an extended theory of planned behaviour perspective. Computers in Human Behavior, 26, 1591–1597.

Bentler, P. M., & Bonett, D. G. (1980). Significant tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588–606.

Boyd, D. M., & Ellison, N. B. (2008). Social network sites: definition, history, and scholarship. Journal of Computer-Mediated Communication, 13, 210–230.

CheckFacebook.com (2010). <http://www.checkfacebook.com/> (accessed June 2010).

Correa, T., Hinsley, A. W., & de Zuniga, H. G. (2010). Who interacts on the web? The intersection of users’ personality and social media use. Computers in Human Behavior, 26, 247–253.

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13, 319–340.

Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1992). Extrinsic and intrinsic motivation to use computers in the workplace. Journal of Applied Social Psychology, 22, 1111–1132.

Deci, E. L. (1975). Intrinsic motivation. New York: Plenum Press. Eldon, E. (2009). Facebook says it has reached 350 million monthly active users.

<http://www.insidefacebook.com/2009/12/01/facebook-says-it-has-reached- 350-million-monthly-active-users/> (accessed June 2010).

Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook ‘‘friends:’’ Social capital and college students’ use of online social network sites. Journal of Computer-Mediated Communication, 12, 1143–1168.

Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50.

Gandal, N. (1994). Hedonic price indexes for spreadsheets and an empirical-test for network externalities. Rand Journal of Economics, 25, 160–170.

Gefen, D., & Straub, D. W. (1997). Gender differences in the perception and use of e- mail: An extension to the technology acceptance model. MIS Quarterly, 1, 389–400.

Gupta, S., & Mela, C. F. (2008). What is a free customer worth? Harvard Business Review, 86, 102–109.

Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. London: Prentice Hall.

Hayduck, L. A. (1987). Structural equation modeling with LISREL. Baltimore, MD: Johns Hopkings University Press.

Igbaria, M., Parasuraman, S., & Baroudi, J. J. (1996). A motivational model of microcomputer usage. Journal of Management Information Systems, 13, 127–143.

Iimi, A. (2005). Estimating demand for cellular phone services in Japan. Telecommunications Policy, 29, 3–23.

Kane, G. C., Fichman, R. G., Gallaugher, J., & Glaser, J. (2009). Community relations 2.0. Harvard Business Review, 87, 45–50.

Kang, Y. S., & Lee, H. (2010). Understanding the role of an IT artifact in online service continuance: An extended perspective of user satisfaction. Computers in Human Behavior, 26, 353–364.

Katz, M. L., & Shapiro, C. (1985). Network externalities, competition and compatibility. American Economic Review, 75, 424–440.

Kim, H. W., Chan, H. C., & Gupta, S. (2007). Value-based adoption of mobile internet: An empirical investigation. Decision Support Systems, 43, 111–126.

Kim, E., & Lee, B. (2007). An economic analysis of customer selection and leveraging strategies in a market where network externalities exist. Decision Support Systems, 44, 124–134.

Kim, G. S., Park, S. B., & Oh, J. (2008). An examination of factors influencing consumer adoption of short message service (SMS). Psychology & Marketing, 25, 769–786.

Kwon, O., & Wen, Y. (2010). An empirical study of the factors affecting social network service use. Computers in Human Behavior, 26, 254–263.

Lee, M. C. (2009). Factor influencing the adoption of internet banking: An integration of TAM and TPB with perceived risk and perceived benefit. Electronic Commerce Research and Applications, 8, 130–141.

Li, C., & Bernoff, J. (2008). Groundswell: Winning in a world transformed by social technologies. Boston, MA: Harvard Business School Press.

Lin, C. P., & Bhattacherjee, A. (2008). Elucidating individual intention to use interactive information technologies: The role of network externalities. International Journal of Electronic Commerce, 13, 85–108.

Lin, C. P., & Bhattacherjee, A. (2009). Understanding online social support and its antecedents: A socio-cognitive model. The Social Science Journal, 46, 724–737.

Lipsman, A. (2007). Social networking goes global, Reston, VA. <http:// www.comscore.com/press/release.asp?press=1555> (accessed November 2009).

K.-Y. Lin, H.-P. Lu / Computers in Human Behavior 27 (2011) 1152–1161 1161

Lou, H., Luo, W., & Strong, D. (2000). Perceived critical mass effect on groupware acceptance. European Journal of Information System, 9, 91–103.

Lu, H. P., & Su, Y. J. P. (2009). Factors affecting purchase intention on mobile shopping web sites. Internet Research, 19, 442–458.

Lu, Y., Zhou, T., & Wang, B. (2009). Exploring Chinese users’ acceptance of instant messaging using the theory of planned behavior, the technology acceptance model, and the flow theory. Computers in Human Behavior, 25, 29–39.

Minton, H. L., & Schneider, F. W. (1980). Differential psychology. Prospect Heights, IL: Waveland Press.

Moon, J. W., & Kim, Y. G. (2001). Extending the TAM for a World-Wide-Web context. Information & Management, 38, 217–230.

Nunnally, J. C. (1978). Psychometric theory. New York: McGraw-Hill. Pae, J. H., & Hyun, J. S. (2002). The impact of technology advancement strategies on

consumers’ patronage decisions. Journal of Product Innovation Management, 19, 375–383.

Pempek, T. A., Yermolayeva, Y. A., & Yermolayeva, S. L. (2009). College students’ social networking experiences on Facebook. Journal of Applied Developmental Psychology, 30, 227–238.

Pfeil, U., Arjan, R., & Zaphiris, P. (2009). Age differences in online social networking: A study of user profiles and the social capital divide among teenagers and older users in MySpace. Computers in Human Behavior, 25, 643–654.

Pontiggia, A., & Virili, F. (2010). Network effects in technology acceptance: Laboratory experimental evidence. International Journal of Information Management, 30, 68–77.

Powell, J. (2009). 33 Million people in the room: How to create, influence, and run a successful business with social networking. NJ: FT Press.

Ross, C., Orr, E. S., Sisic, M., Arseneault, J. M., Simmering, M. G., & Orr, R. R. (2009). Personality and motivations associated with Facebook use. Computers in Human Behavior, 25, 578–586.

Sanchez-Franco, M. J. (2006). Exploring the influence of gender on the web usage via partial least squares. Behaviour & Information Technology, 25, 19–36.

Schmitz, S. W., & Latzer, M. (2002). Competition in B2C e-Commerce: Analytical issues and empirical evidence. Electronic Markets, 12, 163–174.

Scott, J. (1991). The measurement of information systems effectiveness: Evaluating a measuring instrument. In Proceedings of the Fifteenth International Conference on Information Systems, Vancouver, BC (pp. 111–128).

Shurmer, M. (1993). An investigation into sources of network externalities in the packaged PC software market. Information Economics & Policy, 5, 231–251.

Sledgianowski, D., & Kulviwat, S. (2009). Using social network sites: The effects of playfulness, critical mass and trust in a hedonic context. Journal of Computer Information Systems, 49, 74–83.

Tapscott, D. (2008). Grown up digital: How the next generation is changing your world. New York: McGraw-Hill.

Teo, T. S. H., Lim, V. K. G., & Lai, R. Y. C. (1999). Intrinsic and extrinsic motivation in Internet usage. Omega: International Journal of Management Science, 27, 25–37.

van der Heijden, H. (2004). User acceptance of hedonic information systems. MIS Quarterly, 28, 695–704.

Venkatesh, V., & Morris, M. G. (2000). Why don’t men ever stop to ask for directions? Gender, social influence, and their role in technology acceptance and usage behavior. MIS Quarterly, 24, 115–139.

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27, 425–478.

Wu, J. H., Chen, Y. C., & Lin, L. M. (2007). Empirical evaluation of the revised end user computing acceptance model. Computers in Human Behavior, 23, 162–174.

Yang, J., & Mai, E. (2010). Experiential goods with network externalities effects: An empirical study of online rating system. Journal of Business Research, 63, 1050–1057.

Yen, D. C., Wu, C. S., Cheng, F. F., & Huang, Y. W. (2010). Determinants of users’ intention to adopt wireless technology: An empirical study by integrating TTF with TAM. Computers in Human Behavior, 26, 906–915.