From frequency of use to social integration: The mediation of routinization and infusion in Tuenti community

Abstract

This study examines post-adoption behaviors (i.e., frequency of use, routinization and infusion) and their effects on the sense of community in the domain of social network sites. In particular, this contribution formulates mediation hypotheses, which posit how frequency of use affects social integration via routinization and infusion. The data was collected from 278 users of Tuenti, a highly-popular social network site among the Spanish college student population during the period 2006–2012. Results from partial least squares structural equation modeling (PLS-SEM) show these sophisticated types of usage are interrelated in such a way that routinization and infusion (a) fully mediate the effect of frequency of use on social integration; and (b) exert significant influences on social integration, as an active sense of belongingness to a social network site. In order to attain social integration, it is therefore essential for managers to devise strategies to foster advanced post-adoption behaviors.

JEL classification

D8;M15;M21;O33

Keywords

Frequency of use;Routinization;Infusion;Social integration;Social network sites;Partial least squares

1. Introduction

Social network sites (SNS) are here associated with online environments that allow individuals to connect with each other in order to seek or share content, professional advancement, friendship, and entertainment (Sánchez-Franco, Buitrago-Esquinas, & Yñiguez-Ovando, 2015). Their long-term viability will thus depend on (a) their users’ enduring participation, and (b) the formation and maintenance of social capital (Blanchard & Markus, 2004). Nevertheless, there is a paucity of research about the motives which lead people to socially integrate.

In particular, given the relatively recent introduction of SNSs, there is currently limited research related to sophisticated or deeper types of usage and their influences on community successor social integration as an individual-level outcome of the use of an SNS. Initial usage is not enough to fully bring about the desired benefits from an application (Cooper & Zmud, 1990). In this vein, Saga and Zmud (1994) highlighted the importance of expanded usages (i.e., from using more of the technologys features to use the technology to accomplish new tasks or activities) over initial acceptance.

Firstly, growing system research calls for expanding the scope of research from simple usage behavior to post hoc usage (i.e., expanded usages) evaluation (cf. Burton-Jones and Straub, 2006; Chin and Marcolin, 2001 ;  Po-An Hsieh and Wang, 2007, among others). Schwarz and Chin (2007) particularly encouraged a broadening of the understanding of Information System (IS) usage from extent or frequency of use to interrelated usage-based types such as infusion and routinization (cf. Saga & Zmud, 1994). It is necessary (a) “to demystify the complex relationships between different types of usage” (Po-An Hsieh & Zmud, 2006, p. 8) and (b) to analyze their influences on performance outcomes (cf. Sundaram, Schwarz, Jones, & Chin, 2007), such as involvement and participation in social activities, and, consequently, the individuals feelings of identity with and attachment to SNSs. Social integration will be here considered as an essential indicator of community success (or performance outcome).

Secondly, one well-known model that has been used extensively to account for a variety of adoption-acceptance-post adoption usage is proposed by Cooper and Zmud (1990). Our research will concentrate on acceptance, routinization and infusion – broadening our understanding of system acceptance toward post-adoption usage. When members use an SNS in a way that goes beyond simple, or standardized usage, they achieve a higher level of usage that may allow them to exploit the fullest potential of the system, fostering more positive individual benefits (Po-An Hsieh & Wang, 2007).

To achieve the objectives proposed, this paper is organized as follows. First, we detail the literature review and hypotheses, followed by the research method. Next, we present our results, and a discussion of them based on the analysis of data collected from 278 users of Tuenti, the most popular computer-mediated SNS among the Spanish youth (college student and teenager population) during the period 2006–2012 (Apaolaza, Hartmann, Medina, Barrutia, & Echebarria, 2013). Finally, we conclude with a discussion of theoretical and practical implications and limitations.

2. Literature review and research hypotheses

2.1. Social integration

The development and testing of models that may aid in predicting a general positive feeling toward an SNS and a motivation to affectively stay with the SNS is a primary concern in our research (Sánchez-Franco et al., 2015). Nevertheless, there is still a lack of research that analyzes the effective drivers associated with social integration, omitting essential research questions of why members perceive community support provided by one SNS over a competing SNS.

Whereas social capital distinguishably relates to receiving emotional support from people who would help one feel better, social integration measures both the sense of belongingness to and the identification with the SNS and the social communitys interactivity level. In particular, our research draws on two aspects of social integration: “social-psychological or emotional integration which involves introspective social experiences or perceived depth of connectedness, and structural integration which refers to concrete involvement in activities” (Gracia & Herrero, 2004, p. 7), aiming to capture the construct of social integration in its broader sense.

On the one hand, community integration is defined as “members’ feelings of membership, identity, belongingness, and attachment to a group that interacts primarily through electronic communication” (Blanchard, 2007, p. 827; see also Blanchard, 2008). On the other hand, community participation deals with the extent to which members actively engage in community activities, and stresses the social communitys interactivity level (Herrero & Gracia, 2007). Our research presents social integration as a superordinate second-order construct reflected in community participation and community integration, and it focuses on the intercorrelations between first-order factors which form a system of interdependence that is itself important in measuring social integration.

To sum up, the success of SNSs – as communities of relationships – should be the development of a sense of belongingness to an SNS and an active participation in an SNS. Next, we focus on issues such as what motivates users to join and regularly participate in SNSs.

2.2. The relationship between frequency of use and social integration

Frequency of use represents the traditional surrogate measure of the acceptance of an Information Technology (IT), being the behavioral result of attitude and intention to use (Saga & Zmud, 1994). Therefore, taking into account the complexity and flexibility of current SNSs, this represents the first stage in the progression of sophistication in the use of such technologies. We will, therefore, conceptualize frequency of use as the number of times that an individual uses an SNS, the periods being daily, weekly, etc. (cf. Burton-Jones & Straub, 2006).

The IT literature has traditionally supported the link between usage and individual performance as a sequential process (cf. Sundaram et al., 2007). In addition, DeLone & McLeans IS success model (2003) propose overall that the amount of use of a system (e.g., an SNS) is a direct antecedent of the individual impact or net benefits for the user. Accordingly, first, the duration, frequency and intensity of interaction are likely to help members develop a special bonding with others in the online community and thus strengthen their ties (cf. Haron & Razzaque, 2008). Second, a higher frequency implies a high level of exposure to information and communication about issues discussed among community members (cf. de Valck, Van Bruggen, & Wierenga, 2009). This fosters the members affective commitment; that is, the desire to maintain a valued relationship with others. This leads to the first hypothesis.

Hypothesis 1 (H1).

Frequency of use is positively related to social integration.

2.3. The mediating role of routinization

Frequency has been here qualified as initial usage. However, the initial usage may not always be sufficient to fully derive the desired benefits from an SNS. According to innovation diffusion research, users still need to institutionalize (or routinize) the innovation (i.e., an SNS) as part of regular behaviors (Cooper & Zmud, 1990).

Routinization describes the state in which SNS use is no longer perceived as out-of-the-ordinary but actually becomes institutionalized (Saga & Zmud, 1994), and this is associated with habitual and standardized usage; that is, the integrating of the SNS into daily routines (Schwarz & Chin, 2007). Routinization involves an efficient use of the system (Sundaram et al., 2007). In this regard, there is an interrelationship between frequency of use and routinization. Frequent members become accustomed to an SNS and increasingly recognize its use as a normal activity (Saga & Zmud, 1994). An increased (or frequent) usage will, therefore, create an opportunity for the SNS to be routinized (Sundaram et al., 2007).

On the one hand, according to Cooper and Zmud (1990) the institutionalization of behaviors is more important than their initial manifestation (or initial use). Users initially develop a simple usage. After obtaining more experience, users progress to the routinization stage. Routinization will be a consequence of a learning process (Po-An Hsieh & Zmud, 2006) known as exploitation, which can drive individual performance outcomes. Exploitation refers to the routine execution of knowledge (cf. March, 1991). That is, the improvement of existing competencies by using what has already been learnt, namely, by adaptation; i.e., usage that implements and executes ones knowledge of ones SNS and tasks (Burton-Jones & Straub, 2006). Its results are predictable, achieved quickly and positive, leading to foreseeable and immediate benefits for social users in the short run.

On the other hand, higher routinization will provide sufficient time for members to leverage on their learning acquired from daily usage of the SNS (cf. Jasperson, Carter, & Zmud, 2005). We therefore expect a positive relationship between routinized members and the formation and maintenance of social integration. The routinized virtual settings create a level of comfort such that members feel free to discuss and share their ideas, fostering mutual aid and empathy. Routinized members will tend to be more informal and are more willing to share personal information, thereby achieving greater social integration.

In accordance with our presentation so far, we postulate that an increase of frequency of use can intensify a routinized behavior of an SNS, raising the likelihood of intensifying the members’ social integration. Thus, this leads to H2.

Hypothesis 2 (H2).

The relationship between frequency of use and social integration is mediated by routinization.

2.4. The mediating role of infusion

The achievement of the routinization stage does not mean that a person uses an SNS’ full potential (Sundaram et al., 2007). In this regard, we conceptualize infusion as the extent to which a person uses an SNS to its highest level to maximize its potential (Jones, Sundaram, & Chin, 2002), implying the notion of a deeper use. In particular, the infusion stage of an SNS can be operationalized in three alternative ways (Saga & Zmud, 1994): extended use, where individuals utilize more of an SNS’ features to support their life activities; integrative use, reflecting the use of an SNS to enhance linkages between diverse social tasks or life activities; and emergent use, or using an SNS to perform social or life activities that were not previously acknowledged as being amenable to SNS support.

Once members have begun using an SNS, it can be expected that their attitude toward the SNS will be likely to influence the members’ infusion (Jones et al., 2002). As Hall and Loucks (1977) noted, continued interaction with the technology (or frequency of use) can lead to achieving a higher level of use (or infusion) (cf., Afonso, Schwarz, Roldán, & Sánchez-Franco, 2015). Moreover, the IT implementation model can be thought of as activities, some of which may occur in parallel (Cooper & Zmud, 1990). As hypothesized before, frequency of use may lead to routinization and, likewise, could simultaneously influence infusion.

Likewise, taking into account DeLone & McLeans (2003) research, infusion may lead to individual impacts (or net benefits) for the SNS user. Infusion is achieved through the direct experience of users with the SNS and associated learning processes that enable individuals to use an SNS to its full potential. This kind of learning could be identified as exploration (March, 1991). Particularly, exploration refers to the search for novel or innovative ways of doing things with an SNS (cf. Burton-Jones & Straub, 2006), it being postulated that long-run performance depends on sustaining a reasonable level of exploration (March, 1991). Indeed, when individuals use an SNS in a way that goes beyond routinized usage, they achieve a higher level of usage that may allow them to exploit the fullest potential of an SNS, resulting in more positive individual impacts (Cooper & Zmud, 1990) (cf., Afonso et al., 2015).

Infusion can be compared with high quality contents creation by users. Users who reach the infusion stage integrate more socially with the SNS rather than merely being routinized members. As Lim and Lee (2011) comment, “members in this stage are well known by others in the community due to the contribution of good quality information as well as well-built social ties with other core members such as ‘devotees’ and ‘insiders”’.

Therefore, we theoretically expect that the more users explore an SNS, the more likely they will use the full potential that it offers, enhance their own social integration and, consequently, foster ideas or social activities related to mutual interest that were not previously acknowledged as being amenable to SNS support. Based on this logic and previous research, we hypothesize the following:

Hypothesis 3 (H3).

The relationship between frequency of use and social integration is mediated by infusion.

According to the process model of IT implementation, routinization also precedes infusion. As users become familiar with an SNS, they might not be satisfied with the current use situation and may explore and find more useful functionalities to support their life activities (Saga & Zmud, 1994). Thus, through the direct experience and learning processes accumulated in prior stages, individuals who attain the routinization level have the potential to use an SNS in a more comprehensive and sophisticated manner, allowing them to attain the infusion stage (Po-An Hsieh & Wang, 2007) (cf., Afonso et al., 2015). At the same time, Hassandoust, Techatassnasoontorn, and Tan (2016) recognize routinization as the main individual user behavioral factor to explain infusion. Accordingly, we also propose that infusion is preceded by routinization.

Given the above-mentioned theory and empirical evidence, frequency of use will be related to social integration first through routinization and then via infusion. Thus:

Hypothesis 4 (H4).

The relationship between frequency of use and social integration is sequentially mediated by routinization and infusion.

3. Research method

3.1. Participants

Participants were recruited from social studies at a public University in Southern Spain. In January 2009, undergraduate students were asked if they were Tuenti members, and, consequently, were presented with a series of items related to their Tuenti usage. A total of 278 members responded to the offline survey. Table 1 summarizes the respondents’ demographic characteristics.

Table 1. Demographic characteristics of the sample.

Demographic characteristics		Number	Percentage
Gender	Male	119	42.8
	Female	159	57.2
Age (years)	Average	21.04
	SD	2.403
Tuenti friends	Total	110
	Active	63
Frequency of Tuenti usage	Never/almost never	1	0.4
	Less than once a month	2	0.7
	A few times a month	5	1.8
	A few times a week	45	16.2
	About once a day	56	20.1
	Several times a day	169	60.8

3.2. Measures

The instrument development first consisted of reviewing contributions analyzed in the Literature Review and Research Hypotheses Section so as to identify measures for each construct which had established their reliability and validity, thereby satisfying content validity. Likewise, selected scales were modified to make them relevant to the SNS-based context.

In particular, the frequency of Tuenti usage was operationalized by two self-reported measures (see Table 2). As the community behavior we are interested in is not particularly sensitive (with respect to data about income, ethnicity, financial practices, etc.), self-reports could be accurate (Ajzen, 1988). While the routinization measure was developed by Sundaram et al. (2007), the infusion scale has been operationalized by Jones et al. (2002).

Table 2. Measurement model evaluation.

Construct/dimension/indicator	Loadings	Composite reliability	Average variance extracted
Frequency of use		0.908	0.832
Frequency of Tuenti usage	0.892
Frequency of Tuenti usage (very infrequently – very frequently)	0.932
Routinization		0.961	0.890
My use of Tuenti has been incorporated into my regular life schedule	0.944
My use of Tuenti is pretty much integrated as part of my normal life routine	0.944
My use of Tuenti is a normal part of my life	0.942
Infusion		0.943	0.806
I am using Tuenti to its fullest potential for supporting my life	0.918
I am using all capabilities of Tuenti in the best fashion to help me in my life	0.905
I doubt that there are any better ways for me to use Tuenti to support my life	0.858
My use of Tuenti in my life has been integrated and incorporated at the highest level	0.911
Social integrationa		0.914	0.841
Community participation	0.907	0.920	0.659
I intend to actively participate in Tuenti community activities	0.789
In general, I participate in order to stimulate Tuenti community	0.869
I often provide useful information/contents for our Tuenti members	0.826
I eagerly reply to postings by the help-seeker of Tuenti community	0.759
I actively participate in Tuenti community activities to satisfy my social needs	0.813
In general, I’m very motivated to participate actively in Tuenti community activities	0.811
Community integration	0.927	0.923	0.750
I identify with my Tuenti community	0.895
My opinions are valued in my Tuenti community	0.923
Lots of people in my Tuenti community know who I am	0.821
I feel like my Tuenti community is my own	0.822

a. Superordinate multidimensional construct.

Following Herrero and Gracia (2004), our research has tried to capture the construct of social integration in its broader sense (i.e., that of belonging to an SNS and the degree of participation), being designed as a superordinate multidimensional construct (Polites, Roberts, & Thatcher, 2012) composed of its community participation and integration dimensions. Community participation and integration dimensions represent different manifestations of the underlying multidimensional construct (i.e., social integration). Ten items were used to assess community participation and integration – adapted from Algesheimer, Dholakia, and Herrmann (2005), Casaló, Flavián, and Guinalíu (2007), Herrero and Gracia (2007), Koh and Kim (2004), and von Loewenfeld (2006).

Both indicators and dimensions have been modeled following a reflective measurement model (Mode A). All items are seven-point Likert-type, ranging from “strongly disagree” (1) to “strongly agree” (7). A pilot test was conducted with four e-business professors to assess content validity. Table 2 includes a list of the 19 items used in this study.

3.3. Data analysis

Our research model has been tested using Partial Least Squares (PLS), a variance-based structural equation modeling (Roldán & Sánchez-Franco, 2012). We have chosen PLS because (1) the research model is complex according to the type of relationships hypothesized (direct and mediation) and the levels of dimensionality (first- order and second-order constructs); (2) this study uses latent variables scores in the subsequent analysis of predictive relevance, particularly in the implementation of the two-stage approach for modeling the multidimensional construct; and (3) this research defines the nature of the main dependent theoretical construct (i.e., social integration) as defined. This implies that this study relies on a composite measurement model with a reflective design approximation (Mode A), which means that, in the case of the social integration variable, the dimensions (community integration and community participation) represent different facets, although correlations exist between them. Therefore, the use of traditional PLS is advisable (Sarstedt, Hair, Ringle, Thiele, & Gudergan, 2016). This study uses SmartPLS 3.2.4 software (Ringle et al., 2015).

4. Results

4.1. Measurement model

The assessment of the measurement model for reflective indicators in PLS is based on individual item reliability, construct reliability, convergent validity, and discriminant validity (Roldán & Sánchez-Franco, 2012). Individual item reliability is considered adequate in this study because all the indicators and dimensions have loadings above 0.759 (Table 2). All the constructs, dimensions and the multidimensional construct meet the requisite of construct reliability as their composite reliabilities (CR) are greater than 0.7 (Table 2). To assess convergent validity, we examine the average variance extracted (AVE). All latent variables achieve convergent validity given that their AVEs surpass the 0.5 level (Table 2).

Finally, Table 3 shows that all the constructs attain discriminant validity following both the Fornell-Larcker and the strictest HTMT85 criterion (Hair, Hult, Ringle, & Sarstedt, 2017). This means that all the constructs are empirically distinct

Table 3. Measurement model. discriminant validity.

Fornell–Larcker criterion					Heterotrait–Monotrait ratio (HTMT)
	FU	R	I	SI		FU	R	I	SI
FU	0.912				FU
R	0.613	0.944			R	0.703
I	0.482	0.548	0.898		I	0.548	0.577
SI	0.425	0.559	0.657	0.917	SI	0.519	0.639	0.751

Notes: FU: frequency of use; R: routinization; I: infusion; SI: social integration. Fornell–Larcker Criterion: Diagonal elements (bold) are the square root of the variance shared between the constructs and their measures (average variance extracted). Off-diagonal elements are the correlations among constructs. For discriminant validity, diagonal elements should be larger than off-diagonal elements. AVE: average variance extracted.

4.2. Structural model

The sign, size and significance of the structural path coefficients, the R² values, and the Q² test for predictive relevance initially allow an evaluation of the structural model. Consistent with Hair et al. (2017), bootstrapping (5000 resamples) was used to generate t-statistics and confidence intervals. This enables us to assess the statistical significance of the path coefficients. Five of the six direct effects described in Fig. 1 are significant. On the other hand, this analysis shows that H1 is not supported. The direct effect of frequency of use on social integration (c′) is not significant. In addition, our research model seems to have an appropriate predictive power for all the dependent variables ( Table 4). Hence, social integration attains the largest explained variance (0.489). Furthermore, with the exception of the frequency of use variable as an explanatory construct of social integration, the rest of the variables represent key antecedent constructs of their respective dependent variables. This is supported after analyzing the Cohen f² values, which are well above the base level of 0.02 (Roldán & Sánchez-Franco, 2012). We also evaluate the model with the cross-validated redundancy index (Q²) for the endogenous reflective constructs. A Q² greater than 0 implies that the model has predictive relevance. The results summarized in Table 4 confirm that the structural model has satisfactory predictive relevance for all the endogenous constructs. Finally, we have additionally computed the standardized root mean square residual (SRMR) of our model as the root mean square discrepancy between the correlations observed and the model-implied correlations (Henseler, Hubona, & Ray, 2016). Our research model achieves a SRMR of 0.059 (Fig. 1), which means an appropriate fit taking the usual cut-off of 0.08 into account.

Fig. 1. Structural model results.

Table 4. Effects on endogenous variables.

Effects on endogenous variables	Direct effect	t-Value	Percentile bootstrap 90% CI	Explained variance	f2
Routinization (R2 = 0.376/Q2 = 0.331)
Frequency of use (a1)	0.613***	15.989	[0.546; 0.673] Sig.	37.55%	0.601
Infusion (R2 = 0.334/Q2 = 0.260)
Frequency of use (a2)	0.235***	3.666	[0.129; 0.340] Sig.	11.33%	0.052
Routinization (a3)	0.404***	5.622	[0.282; 0.517] Sig.	22.11%	0.153
Social integration (R2 = 0.489/Q2 = 0.407)
H1: Frequency of use (c′)	0.014ns	0.267	[−0.072; 0.104] N.Sig.	0.62%	0.000
Routinization (b1)	0.277***	4.158	[0.161; 0.386] Sig.	15.51%	0.082
Infusion (b2)	0.498***	9.32	[0.409; 0.588] Sig.	32.75%	0.323

Notes: Sig. denotes a significant direct effect at 0.05; N.Sig. denotes a non-significant direct effect at 0.05. Bootstrapping based on n = 5000 subsamples.

• • • . p < 0.001.

ns. : not significant (based on t(4999), one-tailed test).

t(0.001, 4999) = 3.092.

To test our mediation hypotheses (H2–H4), we have applied the analytical approach described by Nitzl et al. (2016). We specify and contrast the indirect effects through the mediators (i.e., routinization and infusion) (Table 5). We also examine the total (c) and direct (H1: c′) effects of the independent variable (i.e., frequency of use) on the dependent variable (i.e., social integration). We have chosen the bootstrapping procedure to test the indirect effects using both percentile and bias corrected confidence intervals. As Table 5 show, frequency of use has a significant total effect on social integration. When mediators are introduced (Fig. 1), frequency of use no longer has a significant direct effect on social integration (H1: c′). Therefore, H1 is not supported. In addition, all the indirect effects of frequency of use on social integration are significant. This means that H2 to H4 have been supported. This way, the analyses reported show that routinization mediates the relationship between frequency of use and social integration (H2: a₁b₁). The results also show that infusion mediates the path between frequency of use and social integration (H3: a₂b₂). Finally, we find that frequency of use is positively associated with higher routinization and infusion, which relates to higher levels of social integration (H4: a₁a₃b₂). Consequently, this means that routinization and infusion fully and jointly mediate the influence of frequency of use on social integration. This is also supported by applying the variance accounted for (VAF) index (Hair et al., 2017), which determines the size of the indirect effect in relation to the total effect. When the VAF has an outcome above 80%, a full mediation can be assumed. This occurs when we assess the total indirect effect of frequency of use on social integration (VAF = 96.6%).

Table 5. Summary of mediating effect tests.

Total effect of FU on SI (c)		Direct effect of FU on SI			Indirect effects of FU on SI
Coefficient	t Value		Coefficient	t Value		Point estimate	Bootstrap 95% CI		VAF
							Percentile	BC
0.425***	9.892	H1:c′	0.014ns	0.267	Total	0.410	[0.343; 0.479]	[0.343; 0.478]	96.6%
					H2:a1b1 (via R)	0.170	[0.101; 0.240]	[0.101; 0.240]	40.0%
					H3:a2b2 (via I)	0.117	[0.063; 0.176]	[0.101; 0.240]	27.6%
					H4:a1a3b2 (via R + I)	0.123	[0.080; 0.172]	[0.079; 0.171]	29.0%

Notes: FU: frequency of use. R: routinization. I: infusion. SI: social integration.

• • • . p < 0.001.

ns. : not significant (based on t(4999), one-tailed test).

t(0.001, 4999) = 3.092.

BC: bias corrected. VAF: Variance accounted for.

5. Discussion

This study has tackled SNS usage at different levels of sophistication, considering a usage range from frequency of use to deeper usage behaviors such as routinization and infusion. Our results show these types of usage are interrelated in such a way that routinization and infusion fully mediate the effect of frequency of use on social integration. In addition, this paper demonstrates that diverse kinds of usage exert different influences on social integration as a gratification (e.g., social self-esteem).

5.1. Theoretical implications

Using a technological diffusion approach (Cooper and Zmud, 1990 ;  Saga and Zmud, 1994), we have distinguished three post-adoption behaviors or successive stages of SNS usage; that is, frequency of use (acceptance), routinization and infusion. This paper is the first that tackles the study of the factors that can influence the infusion variable in the context of SNSs (cf. Hassandoust et al., 2016). In addition, our contribution sheds light on how diverse types of SNS usage exert different influences on social integration, seen as an individual performance outcome or personal net benefit. On the one hand, a routinized behavior – as a consequence of an exploitation learning process (Po-An Hsieh & Zmud, 2006) – is more important than the initial use of an SNS. On the other hand, when members, by an exploration learning process (Burton-Jones & Straub, 2006), use an SNS to its fullest extent, it allows them to capitalize the SNS’ potential, prompting an enhancement of their social integration as a positive individual impact (Cooper and Zmud, 1990 ;  Jones et al., 2002). In fact, infusion represents the best predictor of social integration in our research model (it clarifies 32.75% of its explained variance), surpassing the influence of frequency of use and routinization on the final endogenous variable.

Besides, this study has helped to demystify the complex relationships between these different stages of SNS usage. Indeed, the structural model shows that all path coefficients between frequency of use, routinization and infusion are significant. Consequently, such post-adoptive behaviors are interrelated not only in a sequential process, but also in parallel (Hall & Loucks, 1977); that is, frequency of use is significantly associated with routinization and infusion. Going deeper into this issue, we have shown that routinization and infusion sequentially mediate the relationship between frequency of use and social integration. Our results further exhibit that this mediation is complete. Accordingly, this demonstrates the relevance of extended use or higher levels of usage over initial acceptance (frequency of use) in order to exploit the fullest potential of an SNS. Certainly, in the presence of routinization and infusion, the direct impact of frequency of use on social integration is completely irrelevant.

5.2. Practical implications

Our empirical results provide strong support for the arguments that post-adoption behaviors lead the Tuenti member into developing a growing social integration. Routinized behaviors and infusion predispose members to a higher participation in community activities, peoples feelings of attachment to them and expectations of continuity. Post-adoption behaviors affect the purpose of social communicating in a specific SNS, and provide scholars with a growing understanding of their interaction effects on the relationships between the community drivers and the sense of attachment to their SNS. Our results seem to indicate that analyzed constructs help to foster a desire to develop a stable relationship, a willingness to make short-term sacrifices to maintain the relationship, and a confidence in the stability of the relationship.

Community managers should seek complementary ways to foster their members’ continuance interaction and active participation. On the one hand, community managers need first of all to ensure that an SNS is technically sound, with good security practices put in place to minimize the risks for the users. Perceived risk is an important factor initially affecting user intention to socially integrate into SNSs. Our research also suggests the expanding of technology acceptance models by considering both the affective and the cognitive dimensions of users’ attitude. Managers interested in the development of feelings of being supported by others in an SNS, an active participation in it, and, as a result, a sense of belonging to it should develop strategies and mechanisms based on relational drivers that encourage interaction in terms of quantity or quality (e.g., participating in SNSs includes updating others on their activities and whereabouts, sharing photo and archiving events, getting updates on activities by friends, displaying a large social network, sharing the same interests or passions, among others) and the strength of the relationships between members.

Furthermore, because the SNS is an important part of the members life, highly-routinized members have strong motivations to avoid dissatisfaction and emphasize social integration. In this regard, fulfilling, gratifying, useful and easy access to community features – via technical adjustments and modifications to rules and procedures – to enhance technical and operational knowledge (e.g., self-efficacy) will indeed reinforce the spontaneous tendency to go back to a preferred SNS. Regarding the users’ control level, virtual environments must also be designed to be optimally stimulating to use, with the aim of evoking compelling experiences in order to increase cumulative satisfaction. Post-adoption behaviors – with daily procedures for dealing with SNSs – could then be one type of switching costs as they will become essential if the members question the relationship.

Finally, in this context of the routinization of online social network users, a necessary strategy determines that the SNS and its members be omnipresent in the members life through all kinds of mobile devices. These guarantee immediateness, broadened personal information and proximity location. It is thus indispensable for the community users to visualize their meetings as routine social experiences and even advance toward extended, integrating or emerging uses (cf. Saga and Zmud, 1994; Schwarz and Chin, 2007 ;  Sundaram et al., 2007).

5.3. Limitations

An initial potential limitation that readers might think about this paper is about the SNS selected to test our model, that is, Tuenti. This is why its owner, Telefonica, has shut down this social network, transforming it in a mobile virtual network operator (Dans, 2016). Tuenti had been losing users since 2012. This phenomenon might be explained either by a replacement or switching process to other SNS (e.g., Facebook or Twitter) (Bhattacherjee et al., 2012 ;  Polites and Karahanna, 2012) or by a discontinuance process (Recker, 2016) that affects the users intention to discontinue or switch the use of the SNS, a stage that is previous to the three post-adoption behaviors upon which our attention is focused. Consequently, both the aims and the results derived from our study would not be affected by such an SNS closure.

The second limitation of this research pertains to the sample. This might limit the external validity of our findings. For instance, the specific cultural context may have affected our findings. Future research using a more sophisticated sampling design and different SNSs (e.g. Facebook or Twitter) would establish the external validity of our findings. The third limitation is the age of the data. These were collected in 2009. We considered that it would be advisable to replicate this study with more recent data from a more relevant SNS. The fourth limitation is that our model is a set of main-effect relationships. Potential contingency variables have not been included. Future research should assess the contingent relationship and test the moderating role of variables such as gender and age. Fifth, the design of the research employed was cross-sectional. However, the members’ perceptions can change over time. Accordingly, future research should measure the constructs analyzed at several points of time, taking the dynamics in user patronage behavior into account.

Acknowledgments

This research was supported by the Junta de Andalucía (Consejería de Economía, Innovación y Ciencia) Spain (Proyectos de investigación de excelencia P10-SEJ-6081/P10-SEJ-5801).

References

1. Afonso et al., 2015 C.M. Afonso, A. Schwarz, J.L. Roldán, M.J. Sánchez-Franco; EDMS use in local e-government: An analysis of the path from extent of use to overall performance; International Journal of Electronic Government Research (IJEGR), 11 (2) (2015), pp. 18–34
2. Ajzen, 1988 I. Ajzen; Attitudes, personality, and behavior; Dorsey Press, Chicago, IL (1988)
3. Algesheimer et al., 2005 R. Algesheimer, U.M. Dholakia, A. Herrmann; The social influence of brand community: Evidence from European car clubs; Journal of Marketing, 69 (7) (2005), pp. 19–34
4. Apaolaza et al., 2013 V. Apaolaza, P. Hartmann, E. Medina, J.M. Barrutia, C. Echebarria; The relationship between socializing on the Spanish online networking site Tuenti and teenagers’ subjective wellbeing: The roles of self-esteem and loneliness; Computers in Human Behavior, 29 (4) (2013), pp. 1282–1289
5. Bhattacherjee et al., 2012 A. Bhattacherjee, M. Limayem, C.M. Cheung; User switching of information technology: A theoretical synthesis and empirical test; Information & Management, 49 (7) (2012), pp. 327–333
6. Blanchard, 2007 A.L. Blanchard; Developing a sense of virtual community measure; CyberPsychology & Behaviour, 10 (6) (2007), pp. 827–830
7. Blanchard, 2008 A.L. Blanchard; Testing a model of sense of virtual community; Computers in Human Behaviour, 24 (5) (2008), pp. 2107–2123
8. Blanchard and Markus, 2004 A.L. Blanchard, M.L. Markus; The experienced sense of a virtual community: Characteristics and processes; The Data Base for Advances in Information Systems, 35 (1) (2004), pp. 65–79
9. Burton-Jones and Straub, 2006 A. Burton-Jones, D.W. Straub; Reconceptualizing system usage: An approach and empirical test; Information Systems Research, 17 (2006), pp. 228–246
10. Casaló et al., 2007 L.V. Casaló, C. Flavián, M. Guinalíu; The impact of participation in virtual brand communities on consumer trust and loyalty. The case of free software; Online Information Review, 31 (2007), pp. 775–792
11. Chin and Marcolin, 2001 W.W. Chin, B.L. Marcolin; The future of diffusion research; The DATA BASE for Advances in Information Systems, 32 (2001), pp. 8–12
12. Cooper and Zmud, 1990 R.B. Cooper, R.W. Zmud; Information technology implementation research: A technological diffusion approach; Management Science, 36 (1990), pp. 123–139
13. Dans, 2016 E. Dans; Spains Tuenti: From social network to MVNO; (2016) Retrieved 20.02.16. from https://medium.com/enrique-dans/spain-s-tuenti-from-social-network-to-mvno-a8992dc670f9#.2rcspdxpq
14. de Valck et al., 2009 K. de Valck, G. Van Bruggen, B. Wierenga; Virtual communities: A marketing perspective; Decision Support Systems, 47 (2009), pp. 185–203
15. DeLone and McLean, 2003 W.H. DeLone, E.R. McLean; The DeLone and McLean model of information systems success: A ten-year update; Journal of Management Information Systems, 19 (2003), pp. 9–30
16. Gracia and Herrero, 2004 E. Gracia, J. Herrero; Determinants of social integration in the community: An exploratory analysis of personal, interpersonal and situational variables; Journal of Community and Applied Social Psychology, 14 (2004), pp. 1–15
17. Hair et al., 2017 J.F.J. Hair, G.T.M. Hult, C. Ringle, M. Sarstedt; A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM); (2nd edition)SAGE Publications, Thousand Oaks, CA (2017)
18. Hall and Loucks, 1977 G. Hall, S. Loucks; A developmental model for determining whether the treatment is actually implemented; American Educational Research Journal, 14 (1977), pp. 263–276
19. Haron and Razzaque, 2008 H. Haron, M. Razzaque; The effect of virtual community participation on online purchase intention; D. Spanjaard, S. Denize, N. Sharma (Eds.), Proceedings, Australian and New Zealand Marketing Conference, 2008 (2008), pp. 1–7
20. Hassandoust et al., 2016 F. Hassandoust, A.A. Techatassnasoontorn, F.B. Tan; Factors influencing the infusion of information systems: A literature review; Pacific Asia Journal of the Association for Information Systems, 8 (1) (2016), pp. 1–32
21. Henseler et al., 2016 J. Henseler, G. Hubona, P.A. Ray; Using PLS path modeling in new technology research: Updated guidelines; Industrial Management & Data Systems, 116 (1) (2016), pp. 2–20
22. Herrero and Gracia, 2004 J. Herrero, E. Gracia; Predicting social integration in the community among college students; Journal of Community Psychology, 32 (2004), pp. 707–720
23. Herrero and Gracia, 2007 J. Herrero, E. Gracia; Measuring perceived community support: Factorial structure, longitudinal invariance, and predictive validity of the PCSQ (Perceived Community Support Questionnaire); Journal of Community Psychology, 35 (2007), pp. 197–217
24. Jasperson et al., 2005 J. Jasperson, P. Carter, R.A. Zmud; Comprehensive conceptualization of postadoptive behaviors associated with IT enabled work systems; MIS Quarterly, 29 (2005), pp. 525–557
25. Jones et al., 2002 E. Jones, S. Sundaram, W.W. Chin; Factors leading to sales force automation use: A longitudinal analysis; Journal of Personal Selling & Sales Management, 22 (2002), pp. 145–156
26. Koh and Kim, 2004 J. Koh, Y.G. Kim; Knowledge sharing in virtual communities: An e-business perspective; Expert Systems with Applications, 26 (2004), pp. 155–166
27. Lim and Lee, 2011 J.H. Lim, H. Lee; What makes an online brand community detain its users? A layered motivation perspective; International Journal of eBusiness and eGovernment Studies, 3 (2) (2011), pp. 115–123
28. March, 1991 J.G. March; Exploration and exploitation in organizational learning; Organization Science, 2 (1991), pp. 71–87
29. Nitzl et al., 2016 C. Nitzl, J.L. Roldán, G. Cepeda-Carrión; Mediation analysis in partial least squares path modeling: Helping researchers discuss more sophisticated models; Industrial Management & Data Systems, 116 (9) (2016)
30. Po-An Hsieh and Wang, 2007 J.J. Po-An Hsieh, W. Wang; Explaining employees’ extended use of complex information systems; European Journal of Information Systems, 16 (2007), pp. 216–227
31. Po-An Hsieh and Zmud, 2006 J.J. Po-An Hsieh, R.W. Zmud; Understanding post-adoptive usage behaviors: A two-dimensional view. Diffusion Interest Group in Information Technology (DIGIT); Proceedings. Paper 3 (2006) Available at: http://aisel.aisnet.org/digit2006/3 accessed 16.01.14)
32. Polites and Karahanna, 2012 G.L. Polites, E. Karahanna; Shackled to the status quo: The inhibiting effects of incumbent system habit, switching costs, and inertia on new system acceptance; MIS Quarterly, 36 (1) (2012), pp. 21–42
33. Polites et al., 2012 G.L. Polites, N. Roberts, J. Thatcher; Conceptualizing models using multidimensional constructs: A review and guidelines for their use; European Journal of Information Systems, 21 (1) (2012), pp. 22–48
34. Recker, 2016 J.C. Recker; Reasoning about discontinuance of information system use; Journal of Information Technology Theory and Application, 17 (1) (2016), pp. 41–66
35. Ringle et al., 2015 C.M. Ringle, S. Wende, J.-M. Becker; SmartPLS 3; SmartPLS GmbH, Boenningstedt (2015) http://www.smartpls.com
36. Roldán and Sánchez-Franco, 2012 J.L. Roldán, M.J. Sánchez-Franco; Variance-based structural equation modelling: Guidelines for using Partial Least Squares in information systems research; M. Mora, O. Gelman, A. Steenkamp, M. Raisinghani (Eds.), Research methodologies, innovations and philosophies in software systems engineering and information systems, IGI Global, Hershey, PA (2012), pp. 193–221
37. Saga and Zmud, 1994 V. Saga, R.W. Zmud; The nature and determinants of IT acceptance, routinization, and infusion; L. Levine (Ed.), Diffusion, transfer, and implementation of information technology, Amsterdam, Holland, Elsevier Science BV (1994), pp. 67–86
38. Sánchez-Franco et al., 2015 M.J. Sánchez-Franco, E.M. Buitrago-Esquinas, R. Yñiguez-Ovando; What drives social integration in the domain of Social Network Sites? Examining the influences of relationship quality and stable and dynamic individual differences; Online Information Review, 39 (1) (2015), pp. 5–25
39. Sarstedt et al., 2016 M. Sarstedt, J.F. Hair, C.M. Ringle, K.O. Thiele, S.P. Gudergan; Estimation issues with PLS and CBSEM: Where the bias lies!; Journal of Business Research, 69 (10) (2016), pp. 3998–4010
40. Schwarz and Chin, 2007 A. Schwarz, W.W. Chin; Looking forward: Toward an understanding of the nature and definition of IT acceptance; Journal of the Association for Information Systems, 8 (2007), pp. 230–243
41. Sundaram et al., 2007 S. Sundaram, A. Schwarz, E. Jones, W.W. Chin; Technology use on the front line: How technology enhances individual performance; Journal of the Academy of Marketing Sciences, 35 (2007), pp. 101–112
42. von Loewenfeld, 2006 F. von Loewenfeld; Brand communities—Erfolgsfaktoren und ökonomische Relevanz von Markengemeinschaften; Deutscher Universitäts-Verlag, Wiesbaden (2006)