Effects of social attachment on social media continuous usage intention: The mediating role of affective commitment

Abstract

BACKGROUND:

Social attachment has been identified as a key antecedent motivating users’ social media involvement. However, there is a scarcity of research investigating whether and how three dimensions of social attachment exert impacts on users’ continuous usage intention of social media.

OBJECTIVE:

Based on structural equation model analysis, the current research clarifies the relationships between social attachment, affective commitment and social media continuous usage intention, which unveils the underlying mechanism through which three dimensions of social attachment influence users’ continuous usage intention of social media.

METHODS:

A survey was conducted with 536 informative responses obtained from TikTok public users for hypothesis testing analysis.

RESULTS:

Results indicate that three dimensions of social attachment (social connections, social dependence and social identity) are all positively related to users’ continuous usage intention of social media. Affective commitment partially mediates the relationship between social attachment and users’ continuous usage intention of social media.

CONCLUSIONS:

The current research makes an in-depth study about the underlying mechanism whereby social attachment exerts impacts on social media continuous usage intentionand provides several managerial and theoretical implications. Future research directions are discussed as well.

Keywords

Social attachment social media continuous usage intention affective commitment

Maosheng Yang is currently a Ph.D. candidate at Beijing Jiaotong University. He received a MA in Business Management from Guizhou University. He has a fruitful professional career with one of the big multinational companies in China, where he held a Marketing Manager position responsible for the strategic brand. In the research fields, he has published papers on SSCI/SCI and other international journals. And his research interests mainly involve brand positioning, competitive strategy, and human resources management in a cross-cultural environment.

Dr. Shangui Hu is an associate professor at Anhui University of Technology. He received his Ph.D in Management from University of Science and Technology of China. His research interests include cross-cultural studies, information management, E-commerce and human resources management in cross-cultural settings. Dr. Hu has published papers at SSCI/SCI indexed journals including Information Technology and People, Electronic Commerce Research and Applications, International Journal of Intercultural Relations and other international conferences.

Bagna Essohanam Kpandika holds an MA in Business Management from Beijing University of International Business and Economics, China in 2017. He has a rich experience in Procurement Management, Project Management and Marketing from Huawei Subsidiary in Togo and Startimes Headquarters in Beijing, China, and also held a position of CEO in a sales company in 2014. He currently is a Ph.D. candidate in Beijing Jiaotong University, China from 2017. His research interests include Corporate management, Corporate Social Responsibility and Human Resources Development.

Dr. Lei Liu is an associate professor of business administration at Jiangxi University of Science and Technology. He received a doctorate degree in sociology of technology from Tsinghua University, and holds a master degree in Philosophy of Science and Technology from Guangxi University. He has participated in many international academic seminars and national cooperation projects, and published papers at SSCI/SCI indexed journals. In addition, he is the director of the Innovation and Entrepreneurship and Regional Industry Coordinated Development Research Center.

1 Introduction

Users’ continuous usage of social media still remains vital and challenging for e-business prosperity and has aroused burgeoning research efforts in recent years [1, 2]. For business practitioners, users’ continuous usage of social media is deemed to be more important than the initial usage, because the long-term business value is largely decided by users’ loyalty and social media continuous usage. Previous research investigated the new emergence from the marketing perspective [3, 4], and more scholars attempt to apply interdisciplinary approaches to research users’ social media behaviors from the perspectives of information system and interactive relationships between users and social media platforms [5, 7].

Among all research of social media behaviors, continuous intention of social media usage attracts increasing scholarly interests [8 , 11]. Relevant research has explored antecedents influencing users’ continuous usage intention of social media including users’ satisfaction [12, 13], and previous usage experience [14, 15]. Among all theoretical lens explicating users’ social media behaviors, scholars contend that technology acceptance models and extension theories demonstrate powerful explanations for users’ continuous use. The technology acceptance theory argues that users’ cognitive beliefs (perceived usefulness and perceived ease of use) are the key variables driving their willingness of continuous usage [16 , 18]. The Information System Continuous Use Model (ISCM) further points out that the users’ intention to use information system (IS) depends on their satisfaction and perceived usefulness in the early stage [5].

Although scholars have endeavored to illustrate users’ continuous usage intention from the perspective of cognitive paradigm within the framework of corresponding technology acceptance model or information system continuous use model, there is still a scarcity of research about the underlying mechanism underpinning how users’ continuous usage intention is generated [19].

In order to unveil the underlying mechanism which affects users’ continuous usage intention, Carroll and Ahuvia [20] made a call on investigating the roles of attachment and love. Scholars argue that users show loyalty to a particular social media platform, not switching to other providers of similar services or content of social media, not because of their lack of motivation or conversion of inertia, but rather due to the affective commitment to that social media platform [21, 22]. Therefore, social attachment and affective commitment may be the key elements explicating users’ loyalty to social media and deserve further investigation. However, there is paucity of research investigating how social attachment and affective commitment are connected with users’ continuous intention of social media usage.

In fact, some scholars have discussed the social media continuous usage intention from the perspective of attachment, but there are still some limitations in previous studies. First, scholars have regarded attachment as a major factor driving users’intention of social media continuous usage and reached the conclusion that attachment is positively related to users’ continuous usage intention. However, they regarded attachment as a general concept driving continuous usage intention [19]. Therefore, the effects of different dimensions of attachment exerted on the continuous usage intention still deserve more scholarly attention. Secondly, prior research mainly focused on the effect of attachment on the social media continuous usage intention, while ignoring how attachment affects the social media continuous usage intention. In this regard, those studies have not answered well to the questions why users’ continuous usage underscores the value of developing social software and acquiring competitive business advantages.Thirdly, the majority of previous research on social media continuous usage intention was conducted with samples from western countries. As a second largest economy, more empirical studies are necessary so that a much clearer picture about how to enhance social media continuous usage intention could be developed.

To address the above gaps, a research framework is designed and the structural equation model is adopted to unveil the potential mechanism influencing users’ continuous usage intention of social media. To be more specific, we explore the relationships between the different modes of social attachment (social connection, social dependence and social identity), affective commitment and users’ social media continuous usage intention (see Fig. 1 for the theoretical model).

Fig. 1

Research framework.

The remainder of the paper is organized as follows. The next section is about literature review and hypothesis development. After that, research methodology is demonstrated, followed by data analysis result. And the part of conclusion ends the whole paper.

2 Literature review and hypothesis development

2.1 Social attachment and continuous usage intention

As a typical character attachment, social attachment has the basic attributes and functions of attachment, paying attention to the emotional characteristics and behavioral tendencies that individuals form with specific objects in their growth [23]. In this study, social attachment is a unique and affective bond between users and social media [24]. According to the basic point of view of character attachment theory, users with a strong sense of attachment regard social media as an integral part of their lives, and are willing to pay extra time, energy, and money and other personal resources for using social media, thereby generating sustained and positive usage behaviors [21, 25]. It is acknowledged that social attachment theory can well explain the bond between users and social media [26] which mainly contains three important aspects: social connections, social dependence and social identity [24]. It indicates that individuals who have more close friends on the social media platform have a stronger attachment to the social media. Social connections provide networks for individuals to establish and maintain interpersonal interactions [27]. In this regard, users keep connected with other users in the virtual world and exchange information, share ideas and offer feedback about online shopping experiences. Those sharing common interests are bounded together and build up affective and personal relationships, which can stipulate the users ’ willingness to use social media [28, 29]. Social dependence reflects the users’ dependence on a certain social media or a sort of characterized function. We argue that users who demonstrate more dependence on a social media platform are supposed to show more willingness to continuously use that platform [23, 28]. Social identity is the degree to which an individual feels connected to social media and belongs to social media as a whole. In this regard, users are expected to have sense of belonging to that social media community and thus develop strong intention of continuously using social media platforms as well [30]. Overall, based on the analysis of prior research, we hypothesize that:

H1: Social connection (H1a), Social identity (H1b) and Social dependence (H1c) are positively related to users’intention of social media continuous usage intention.

2.2 Affective commitment and continuous usage intention

Affective commitment refers to the identification, participation and affective attachment of the relationship between users and specific targets (such as virtual communities, social platforms) [31]. Affective commitment affects the users’ intention to use social media and long-term affective relationships between users and social media platforms. On one hand, social media satisfies users’ cognitive and emotional needs by searching for relevant information and interacting with community members [25]. On the other hand, users’ emotional bonds with social media have been encouraged as well [27]. Thus users’ affective commitment to the social media platforms have been finalized [32]. Relevant research has identified that affective commitment is an important prerequisite for generating users’ willingness and behavior for social media continuous usage. Zhou et al. [33] adopted the commitment dedication constraint framework to discuss the issues on customers’ continuous usage of information system. The structural equation model analysis with 438 experienced users found that the impact of affective commitment on users’ continuous usage intention of information system services is stronger than that of satisfaction. Furthermore, users’ continuous usage of information system is determined jointly by two mechanisms, affective commitment (attraction) and computational commitment (lock), among which affective commitment plays a more central role. Therefore, affective commitment is supposed to be a key predictor of users’ continuous involvement in information systems or virtual communities [34, 35]. Based on the above analysis, we hypothesize that:

H2: Affective commitment is positively related to users’ intention of social media continuous usage intention.

2.3 Social attachment and affective commitment

Emotional commitment is the emotional tendency to maintain continuously valuable relationship. It can encourage individuals to maintain this relationship and serves as the basic element of long-term relationship success. It is developed on the basis of personal dedication. That is to say, the individual hopes to maintain long-term mutually beneficial relationship [36], reflecting the individual’s sense of participation and belonging in the desirable relationship. Due to the good attitudes, influences, emotions, and perceptions of experiences in valuable relationship, emotional commitment has always been considered as the most important factor to predict continuous intention [32]. Therefore, some scholars argue that affective commitment is a desire-based attachment that attracts users to maintain long-term relationship with support of social media [33]. Previous research has identified that affective commitment is an important predictor determining users’ continuous usage of virtual communities and information systems [25]. Affective commitment reflects the degree of affective bond that users perceive to connect with community members emotionally. During the experiences of using social media usage, users have developed affective bond with social media due to perceived usefulness and satisfaction with functionalities created by social media [25, 27]. In order to attach self to the group identity, users enhance their affective bond to the degree of affective commitment to social media usage [25, 33]. Given the user’s openness and self-identification features, social media provides users with various opportunities to benefit them from multiple sources, including social connections, social dependence and social identity. The basic principle of social attachment’s impact on affective commitment is to satisfy users’ personal needs stemmed from their work (i.e social connection) to their social life (i.e social dependence) and finally to spiritually-oriented applications (social identity) [37]. When expecting a high level of future value, users are inclined to establish and maintain a relatively long-term relationship with preferred social media [19]. During the process of social media application, users’ affective commitment to the social media has thus been developed [38, 39].

Based on the aforementioned, we make the following hypothesis:

H3: Social connections (H3a), Social identity (H3b) and Social dependence (H3c) are positively related to users’ affective commitment.

2.4 Mediating role of affective commitment

Relevant research has identified both the direct and indirect associations between social attachment and affective commitment [40]. Taking the perspective of the attachment theory, scholars have identified that social attachment is significantly associated with affective commitment, which positively responds to users’ continuous use of social media [41]. During the process of users’ social attachment with a social commerce website, they are inclined to acquire emotional support when interacting with members of that social media community built by that social media platform [35, 39]. Received support from community members brings happiness and satisfaction to users and thus contributes to growth of affective commitment, thereby positively influencing their intention of social media continuous usage [35, 42]. Related research has also found that the attachment relationship between users and social media will affect the social attachment between users and social media platforms through the partial mediating effect of the emotional commitment to social media usage [5, 43]. Therefore, we argue that the social attachment stimulated by the use of social media may have an indirect positive effect on the user’s social attachment through the mediating effect of emotional commitment.

Based on the above statement, we hypothesize that

H4: Affective commitment plays a mediating role between social connections (H4a), social identity (H4b), social dependence (H4c) and users’ intention of social media continuous usage intention.

3 Research methodology

3.1 Participants and procedure

Participants were recruited through the online platform TitTok in mainland China, which achieves increasing popularity and provides functions equivalent to Amazon Mechanical Turk (MTurk). In order to stimulate participants’ interest, a monetary reward of $6 was offered. The online platform has been successfully used for collecting primary data in published social media research [44]. Research has also demonstrated that the platform allows researchers to collect data with good quality and reliability [44].

More importantly, the study selected TikTok mass users as the research samples. We chose this local social media for the current research for several reasons. First, Tiktok, as the largest leading social platform in China, has been proved to be feasible and reliable in previous studies [44, 45]. Secondly, TikTok is a local representative APP, which is a leader in the market segments of social software and possesses the typical characteristics required by the research work. If research findings are supported by the data of TikTok, the current study can just provide management suggestions to improve the corporate brand strategy of the local social software platform. Therefore, from this perspective, the use of TikTok public users to conduct data surveys is compatible with our original research purpose. A questionnaire was delivered to samples from TitTok mother group users [45, 46]. The feasibility sampling method was used to collect the data.

The whole research period lasted 6 months. Participants were asked to complete general background information and then a formal survey was conducted. Survey links were sent to participants’ mobile phones. We recruited 536 participants who were engaged in a variety of industries (e.g., finance, health care, professional services, education). And 520 returned their responses with an overall response rate of 89.7%.Among 520 participants, female accounted for 74.6% and male accounted for 25.4%. The average age is 26.36 and 29.6% participants were under 20 years old, 31.5% between 21 to 30 years, and 19.6% between 30 to 40 years. Over 41 years old accounted for 19.3%. Unmarried participants accounted for 51% and the married accounted for 49%.

3.2 Measures

The variable measurements were selected from existing SSCI journal papers, and necessary modifications were made to meet the status quo of samples using TikTok.

To measure affective commitment, we used four items proposed by Zhou et al. [32]. The sample items included “When I use TikTok, I immerse myself unconsciously” and “I have a deep affection for TikTok”.

A five-item scale was adapted to measure social media usage, which was originally developed by Thong et al. [47]. The sample items included “I share the experience of using TikTok with my friends” and “ When I have the need, I use TikTok”.

A 16-item scale, proposed by Kyle, et al. [48] and Yang, et al. [24], were adopted to measure three dimensions of social attachment, among which 5 items were used to measure social connection, 5 items for measuring social identity and 6 for assessing social dependence. The sample items included “TikTok has a special connection with me (for example, following stars, entertainment)”, “TikTok has become a part of my life or work” and “TikTok can show oneself or express emotions”.

In addition, we also listed 7 items including participants’ gender, age, marital status, education, occupation, monthly consumption level and time spent on social media usage as control variables because those items are closely connected with TikTok usage.

3.3 Data analysis method

After deleting the invalid responses, the research applied Mplus7.0 to examine relevant CFA, convergence validity, discriminant validity, composite reliability, average variance extracted, and path coefficients to ensure the match between research model and statistics. As Bagozzi and Yi [49] indicated, SEM was developed from covariance structure analysis to examine the relationships between different variables at the same time. And this analysis method fits well with the current research.

4 Data analysis result

4.1 Confirmatory factor analysis

In this study, the evaluation and reduction of CFA measurement model variables were revised based on the two-stage model proposed by Kline [50]. If the measurement model fit is acceptable, the full SEM model report can be followed.

There are five constructs in this research model, namely continuous usage intention, affective commitment, social connection, social dependence and social identity. According to the standards proposed by Anderson and Gerbing [51], CFA analysis was performed on all constructs. Following the standard proposed by Hair, Anderson, Tatham and Black [52], Nunnally and Bernstein [53], and Fornell and Larcker [54], convergent validity was tested with standardized factor loadings greater than 0.50, the composite reliability higher than 0.60, and the average variance extracted was higher than 0.50. Therefore, according to the results shown in Table 1, each variable of this study has a good reliability and polymerization validity.

Table 1
Confirmatory factor analysis

Construct Item Significant test of parameter estimation Item Reliability Composite Reliability Convergence Validity

Unstd. S.E. Unstd./S.E. P STD. SMC CR AVE

AC AC1 1.000 0.816 0.666 0.892 0.676

AC2 1.126 0.053 21.276 0.000 0.897 0.805

AC3 1.024 0.053 19.296 0.000 0.828 0.686

AC4 1.051 0.064 16.419 0.000 0.740 0.548

CI CI1 1.000 0.672 0.452 0.863 0.558

CI2 0.883 0.069 12.823 0.000 0.729 0.531

CI3 0.902 0.066 13.617 0.000 0.785 0.616

CI4 1.043 0.071 14.631 0.000 0.807 0.651

CI5 0.974 0.075 12.953 0.000 0.736 0.542

SC SC1 1.000 0.747 0.558 0.853 0.538

SC2 0.945 0.065 14.456 0.000 0.704 0.496

SC3 0.960 0.067 14.285 0.000 0.735 0.540

SC4 1.053 0.070 15.058 0.000 0.769 0.591

SC5 0.881 0.065 13.482 0.000 0.712 0.507

SD SD1 1.000 0.707 0.500 0.880 0.550

SD2 0.964 0.073 13.263 0.000 0.704 0.496

SD3 1.181 0.074 15.957 0.000 0.817 0.667

SD4 1.196 0.083 14.468 0.000 0.763 0.582

SD5 1.133 0.080 14.125 0.000 0.737 0.543

SD6 1.136 0.085 13.388 0.000 0.714 0.510

SI SI1 1.000 0.775 0.601 0.880 0.648

SI2 1.063 0.055 19.180 0.000 0.879 0.773

SI3 1.015 0.055 18.612 0.000 0.856 0.733

SI4 0.891 0.061 14.645 0.000 0.697 0.486

Construct	Item	Significant test of parameter estimation	Item Reliability	Composite Reliability	Convergence Validity
AC	AC1	1.000				0.816	0.666	0.892	0.676
	AC2	1.126	0.053	21.276	0.000	0.897	0.805
	AC3	1.024	0.053	19.296	0.000	0.828	0.686
	AC4	1.051	0.064	16.419	0.000	0.740	0.548
CI	CI1	1.000				0.672	0.452	0.863	0.558
	CI2	0.883	0.069	12.823	0.000	0.729	0.531
	CI3	0.902	0.066	13.617	0.000	0.785	0.616
	CI4	1.043	0.071	14.631	0.000	0.807	0.651
	CI5	0.974	0.075	12.953	0.000	0.736	0.542
SC	SC1	1.000				0.747	0.558	0.853	0.538
	SC2	0.945	0.065	14.456	0.000	0.704	0.496
	SC3	0.960	0.067	14.285	0.000	0.735	0.540
	SC4	1.053	0.070	15.058	0.000	0.769	0.591
	SC5	0.881	0.065	13.482	0.000	0.712	0.507
SD	SD1	1.000				0.707	0.500	0.880	0.550
	SD2	0.964	0.073	13.263	0.000	0.704	0.496
	SD3	1.181	0.074	15.957	0.000	0.817	0.667
	SD4	1.196	0.083	14.468	0.000	0.763	0.582
	SD5	1.133	0.080	14.125	0.000	0.737	0.543
	SD6	1.136	0.085	13.388	0.000	0.714	0.510
SI	SI1	1.000				0.775	0.601	0.880	0.648
	SI2	1.063	0.055	19.180	0.000	0.879	0.773
	SI3	1.015	0.055	18.612	0.000	0.856	0.733
	SI4	0.891	0.061	14.645	0.000	0.697	0.486

Note, Unstd., Unstandardized factor loadings. Std, Standardized factor loadings. SMC, Square Multiple Correlations. CR, Composite Reliability. AVE, Average Variance Extracted.

4.2 Discriminant validity

The discriminant validity analysis is to examine whether different two variables shut in the statistics are different or not. This study uses the latest discriminant validity analysis method, namely confidence interval method [55]. The confidence interval method is used to confirm the confidence interval of the correlation coefficient between variables. If it fails to include “1”, then it is completely correlated, indicating that the facets have different validity. As Hancock and Nevitt [56] suggested, bootstrap test was conducted in this study. 95% confidence interval of the correlation coefficient does not involve 1 (see Table 2 below), which shows the good discriminant validity between all the variables.

Table 2
Distinct validity

Parameter Estimate Bias-corrected 95% Percentile 95%

Lower Upper Lower Upper

SC< –>SI 0.748 0.622 0.843 0.621 0.842

SC< –>SD 0.725 0.618 0.813 0.618 0.813

SC< –>CI 0.690 0.600 0.782 0.596 0.781

SC< –>AC 0.455 0.343 0.552 0.343 0.552

SI< –>SD 0.751 0.678 0.821 0.676 0.818

SI< –>CI 0.716 0.636 0.788 0.633 0.787

SI< –>AC 0.474 0.380 0.568 0.375 0.567

SD< –>CI 0.734 0.663 0.801 0.663 0.801

SD< –>AC 0.472 0.372 0.570 0.365 0.562

AC< –>CI 0.480 0.376 0.564 0.376 0.564

Parameter	Estimate	Bias-corrected 95%	Percentile 95%
SC< –>SI	0.748	0.622	0.843	0.621	0.842
SC< –>SD	0.725	0.618	0.813	0.618	0.813
SC< –>CI	0.690	0.600	0.782	0.596	0.781
SC< –>AC	0.455	0.343	0.552	0.343	0.552
SI< –>SD	0.751	0.678	0.821	0.676	0.818
SI< –>CI	0.716	0.636	0.788	0.633	0.787
SI< –>AC	0.474	0.380	0.568	0.375	0.567
SD< –>CI	0.734	0.663	0.801	0.663	0.801
SD< –>AC	0.472	0.372	0.570	0.365	0.562
AC< –>CI	0.480	0.376	0.564	0.376	0.564

Note, SC, Social Connection. SI, Social Identity. SD, Social Dep-endence. AC, Emotional commitment. CI, continuance intention.

4.3 Research hypothesis test

The purpose of research hypothesis testing is to understand the impacts of the independent variable on the estimated value of the dependent variable proposed in the theoretical model. In the current research, Table 3 suggested that social connection (β= 0.167, p < 0.001), social identity (SI) (β= 0.249, p < 0.001) and social dependence (SD) (β= 0.413, p < 0.001) were positively related to the continuous intention. Thus H1a, H1b and H1c were supported. In addition, affective commitment (AC) (β= 0.231, p < 0.01) was also found to be positively related to continuous usage intention. Thus H2 was confirmed. Further, the statistics showed that social connection (SC) (β= 0.216, p < 0.05), social identity (SI) (β= 0.239, p < 0.05) and social dependence (SD) (β= 0.237, p < 0.05) were all significantly associated with affective commitment (AC). Therefore, H3a, H3b and H3c were supported as well.

Table 3
Hypothesis testing

DV IV Unstd S.E. Unstd./S.E. P Std. R² Hypothesis supported?

AC SC 0.216 0.102 2.113 ^* 0.198 0.290 supported

SI 0.239 0.109 2.188 ^* 0.206 supported

SD 0.237 0.109 2.179 ^* 0.189 supported

CI SC 0.167 0.071 2.357 ^* 0.180 0.657 supported

SI 0.249 0.076 3.263 ^* 0.252 supported

SD 0.413 0.080 5.185 ^* 0.387 supported

AC 0.093 0.040 2.341 ^** 0.109 supported

DV	IV	Unstd	S.E.	Unstd./S.E.	P	Std.	R²	Hypothesis supported?
AC	SC	0.216	0.102	2.113	^*	0.198	0.290	supported
	SI	0.239	0.109	2.188	^*	0.206		supported
	SD	0.237	0.109	2.179	^*	0.189		supported
CI	SC	0.167	0.071	2.357	^*	0.180	0.657	supported
	SI	0.249	0.076	3.263	^***	0.252		supported
	SD	0.413	0.080	5.185	^***	0.387		supported
	AC	0.093	0.040	2.341	^**	0.109		supported

Note, ^*p-value < 0.05. ^**p-value < 0.01. ^***p-value < 0.001.

4.4 Model fit analysis

This study used goodness-of-fit indices to investigate the measurement model’s goodness-of-fit. The purpose is to evaluate whether or not it can explain the actual observed data or if there is a gap between the theoretical model and actual observed data [57]. The first part is “Absolute Fit Measures”, which includes MLχ²/df, Normed Chi-square (χ²/df), GFI, AGFI, SRMR and AGFI, MLχ²and degree of freedom. The maximum of MLχ²/df is 3 [58]. GFI and AGFI should be over 0.9 [58, 59]. SRMR and RMSEA should be below 0.08 respectively, and a value close to 0 indicates that the model fits very well [59 –61]. Comparative fit indexes include NNFI and CFI. Bentler [59] suggested that the value of comparative fit indexes should be higher than 0.9.

The measurement model used in this research met all the required criteria, which suggested that the current research model adequately matches well with the data, as shown in Table 4.

Table 4
Measurement and structural model goodness-of-fit indices

Model fit Criteria Model fit of research model Result

MLχ² The small the better 817.322

DF The large the better 282.376

Normed Chi-sqr (χ²/DF) 1 < χ²/DF < 3 2894 Supported

RMSEA <0.08 0.075 Supported

SRMR <0.08 0.068 Supported

TLI (NNFI) >0.9 0.953 Supported

CFI >0.9 0.906 Supported

GFI >0.9 0.935 Supported

AGFI >0.9 0.916 Supported

Model fit	Criteria	Model fit of research model	Result
MLχ²	The small the better	817.322
DF	The large the better	282.376
Normed Chi-sqr (χ²/DF)	1 < χ²/DF < 3	2894	Supported
RMSEA	<0.08	0.075	Supported
SRMR	<0.08	0.068	Supported
TLI (NNFI)	>0.9	0.953	Supported
CFI	>0.9	0.906	Supported
GFI	>0.9	0.935	Supported
AGFI	>0.9	0.916	Supported

4.5 Mediating effect analysis

In order to test the mediating effects of affective commitment, this study also employed structural equation model analysis. First, the bootstrap estimation was conducted to estimate the standard error of the mediating effect, and then the significant level of the mediating effect was examined. The results (Table 5) showed that the overall effect of social connection on continuous intention is 0.187. The standard error was 0.095. And the Z value was 1.972, which was greater than the benchmark of 1.96. At the 95% confidence level, the lower limit of the confidence interval obtained by the Bias-corrected estimation method was 0.18 and the upper limit was 0.78, which excluded 0. The result indicated that mediating hypothesis was supported. In the same way, the indirect effect and the direct effect were also confirmed. Therefore, affective commitment partially mediates the relationship between social attachment and users’ continuous usage of social media. In this regard, H4a, H4b and H4c were all supported.

Table 5
Indirect effect analysis table of intermediary model

Effect Point Estimate Product of coefficients Bootstrap Bias-corrected 95%

SE Z-Value p-value Lower Upper

Total effect SC ⟶ CI 0.187 0.095 1.972 ^ 0.029 0.418

Indirect effect SC ⟶ AC ⟶ CI 0.020 0.015 1.967 ^ 0.001 0.063

Direct effect SC ⟶ CI 0.167 0.092 1.918 ^ 0.010 0.382

Total effect SI ⟶ CI 0.271 0.104 2.615 ^ 0.063 0.465

Indirect effect SI ⟶ AC ⟶ CI 0.022 0.016 1.991 ^ 0.001 0.067

Direct effect SI ⟶ CI 0.249 0.104 2.401 ^ 0.047 0.441

Total effect SD ⟶ CI 0.435 0.100 4.361 ^* 0.245 0.637

Indirect effects SD ⟶ AC ⟶ CI 0.022 0.016 2.349 ^ 0.001 0.069

Direct effect SD ⟶ CI 0.413 0.100 4.124 ^*** 0.224 0.608

Effect	Point Estimate	Product of coefficients	Bootstrap Bias-corrected 95%
Total effect SC ⟶ CI	0.187	0.095	1.972	^**	0.029	0.418
Indirect effect SC ⟶ AC ⟶ CI	0.020	0.015	1.967	^**	0.001	0.063
Direct effect SC ⟶ CI	0.167	0.092	1.918	^**	0.010	0.382
Total effect SI ⟶ CI	0.271	0.104	2.615	^**	0.063	0.465
Indirect effect SI ⟶ AC ⟶ CI	0.022	0.016	1.991	^**	0.001	0.067
Direct effect SI ⟶ CI	0.249	0.104	2.401	^**	0.047	0.441
Total effect SD ⟶ CI	0.435	0.100	4.361	^***	0.245	0.637
Indirect effects SD ⟶ AC ⟶ CI	0.022	0.016	2.349	^**	0.001	0.069
Direct effect SD ⟶ CI	0.413	0.100	4.124	^***	0.224	0.608

Note, ^**p-value < 0.01. ^***p-value < 0.001.

5 Discussions

5.1 Research results

The research confirmed that the three dimensions of user’s social attachment all exert significant impacts on the continuous intention.

To be more specific, social connection is positively related to user’s continuous usage intention. The research finding is consistent with the previous research [28, 29]. Based on prior experiences, affective reactions and cognitive representations, users interact with social media platforms. In this regard, users’ emotional connection is developed [23] and thereby users’ continuous usage intention is influenced through affective bonds. Second, social dependence was found to be significantly related to users’ continuous usage intention. This has also confirmed the previous research findings [27, 29]. As Lewicka [63] indicated, developing and maintaining users’ social dependence not only contributes to growth of users’ continuous usage intention, but also stimulates users’ positive social attitudes toward social media platforms. Finally, social identity is also positively related to users’ continuous usage intention. The research result is in line with previous investigations [28, 30]. In a virtual world, users interact with others by establishing a virtual identity. Similar to the real world, social media users are motivated by sense of belonging and self-demonstration when they surf the virtual world. This reflects users’ demand of social identity, which increases users’ communications and emotional bonds with others, and thus improves their continuous usage intention.

In regard to affective commitment, three dimensions of social attachment are all significantly related to affective commitment, which confirms the prior research as well [34]. When exposed to social media environment, users tend to interact frequently with others. During the process of interpersonal interactions, users’ affective bonds with other community members are created and strengthened through the social media platform. During the process of emotional appraisals, established affective bonds increase users’ understandings and favoritism about social media, and thus affective commitment is fostered. Likewise, social dependence is also positively associated with affective commitment. Goel et al. [64] argued that products and relevant services are supposed to satisfy users’ utilitarian purposes. When users’ expectations are satisfied, their emotional distress and skepticism about product’s quality are mitigated, and their dependence on social media is naturally increased. Further, when users perceive the potential value of functionalities and uniqueness of social media, their confidence is also boosted by appropriate products and perceived satisfaction. In this regard, their affective commitment to social media is developed. Accordingly, the current research has also confirmed that social identity contributes to affective commitment. Social media provides a virtual environment and support for users’ interaction with others. With regard to the increasing popularity of social media, a wide range of platforms create opportunities for social interactions and self identifications between individuals. When users’ interaction is responded and self identity is accepted by other members of communities, their social identity is developed in that social media community. We contend that users’ social identity and affective commitment will be increased when they are involved more in the activities supported by social media platforms.

The current research has also confirmed the effect of users’ affective commitment on continuous usage intention. The findings keep consistent with previous research [19, 38]. Affective commitment is expected to accumulate resources for users’ learning activities when they use social media to interact with others. In case users’ attitudes and affective reactions may shift between social media platforms, affective commitment should be considered a priority except that users’ basic needs are satisfied by social media.

More importantly, the research has also identified the mediating role of affective commitment, which unveils the underlying mechanism through which social attachment exerts impacts on users’ continuous usage intention. The findings are also consistent with prior research [19, 42]. The research findings suggest that affective commitment is an efficacious channel through which three dimensions of social attachment (social connection, social dependence and social identity) are connected with users’ continuous usage intention.

5.2 Theoretical and practical implications

The current research makes an in-depth study about the underlying mechanism whereby social attachment exerts impacts on social media continuous usage intention. Based on research findings, the current research makes several important contributions.

The current research makes important theoretical contributions.

First, this article takes the interdisciplinary perspective of combining social media continuous usage intention and social attachments. Existing research on continuous usage intention currently focuses on marketing, while social attachment belongs to the field of psychology. This research fills the gap of previous related research perspectives by combining social media continuous usage intention with social attachment and this method unveils whether and how social attachment impacts social media continuous usage intention. And this approach serves as a new avenue for future research.

Second, the current research extends the application of attachment theory. This study divides social attachment into three categories, namely, social connection, social dependence, and social identity and examines the underlying mechanism whereby different components of social attachment influence social media continuous usage intention. The research findings enrich the literature on social attachment and attach the theory to the field of information management.

Third, this study contributes to literature on affective commitment as well. As a connecting bond between social attachment and social media continuous usage intention, affective commitment was found to play an enabling role. To the best of our knowledge, this study is among the first trials to identify the role of affective commitment in bridging social attachment and social media continuous usage intention. And this contributes to a part of efforts made for unveiling the mechanism driving social media continuous usage intention.

Except for the theoretical contributions, the current research also offers practical implications for business practitioners.

First, practitioners should endeavor to improve users’ social attachment because three dimensions of social attachment have been identified to be able to enhance users’ social media continuous usage intention. Faced with the fierce market competition, social media providers should lay much emphasis on the development of users’ social attachment by offering services to satisfy their hedonic, functional and representative demands. And necessary investments should be made to contribute to affective bonds between users and social media platforms, which help users to satisfy self, realize self and get rich in self. In this way, users’ cognitive framework would be re-structured and integrated with social media. Thus users are willing to invest personal resources for developing social attachment and continuous intention of using social media as a return.

Second, practitioners should make sufficient efforts to increase users’ affective commitment, thereby increasing users’ continuous usage intention of social media. In light of functioning similarities possessed by various social media platforms, more investments should be made to increase uniqueness of social media platforms in case users may shift their interests to other ones. Strategies to further loyal users’ interests and favoritism should be considered. Social media providers should increase the attractiveness of platform design, style, content and offer more conveniences for users’ usage to search for what they are most interested in. In this regard, users’ social attachment could be enhanced. In addition, social media providers should endeavor to bring more easiness and opportunities for users to communicate with others to help users establish social identity and belongingness to that social media community. Necessary measures should be taken as well for those active or potential users to encourage them to continuously use the platform. Timely and intimate gifts and individualized services could be offered to stimulate users’ loyalty about the social media usage. Those measures are supposed to be effective to increase users’ affective commitment, thereby encouraging users’ continuous usage intention of social media.

5.3 Limitations and future research directions

Although our study provides useful insights, it has several limitations which could be addressed in the future research efforts.

First, the current research has only explored the mediating role of affective commitment. More empirical research is required to further explore boundary conditions for social media usage in social media environment. In light of the importance of the research topic, future studies should attempt to extend our proposed model by incorporating other potential social commerce factors including moderating and mediating elements.

The second limitation is about an inherent problem of a common method bias which was caused by a single questionnaire to measure all constructs at the same time in this study. Although Harman one-factor testing revealed a low possibility of common method bias, a marker-variable technique should be designed whether common method bias inflates or deflates the results within the cross-sectional data. The power of the proposed model in predicting the relationships between social attachment, affective commitment and social media continuous usage intention may be further validated in the future studies by measuring users’ real social media behaviors.

Third, the current research adopts the cross-sectional design. All variables in the research model were measured at the same time and their relationships can only be inferred to be associative, not causal ones. Future longitudinal or experimental design could be alternatives to solve this problem by working out the potential causal relationships between variables.

Finally, the research findings should be cautiously generalized because of the research samples. Although all variables measurement has demonstrated good validity and reliability and no serious common bias problem, the research was conducted with samples of Chinese social media users. Sample groups from other countries could be selected from various social media platforms including Facebook, Twitter to duplicate the research findings of the current study.

Footnotes

Acknowledgments

The research was financially supported by 2019’ National Education Sciences Planning Project of China under Grant BBA190019 and Anhui Provincial Natural Science Foundation under Grant 1908085MG238.

Author contributions

CONCEPTION: Shangui Hu and Maosheng Yang

INTERPRETATION OR ANALYSIS OF DATA: Maosheng Yang and Lei Liu

PREPARATION OF THE MANUSCRIPT: Maosheng Yang and Bagna Essohanam Kpandika

REVISION FOR IMPORTANT INTELLECTUAL CONTENT: Shangui Hu, Lei Liu and Maosheng Yang

Appendix

Measure items:

Affective commitment

1. When I use TikTok, I immerse myself unconsciously

2. I have a deep affection for TikTok

3.TikTok gives me a strong sense of belonging

4.TikTok is very attractive to me

Social media continuous usage intention

1. I share the experience of using TikTok with my friends

2. When I have the need, I use TikTok

3. In the future, I will continue to use TikTok

4. I recommend TikTok to my friends

5. I use TikTok to communicate with friends and share information

Social connections

1. TikTok has a special connection with me (for example, following stars, entertainment)

2. TikTok has become a part of my life or work

3. TikTok can show oneself or express emotions

4.TikTok brought me happiness and joy

5.TikTok can meet my virtual social needs

Social dependence

1. I like TikTok better than similar social software (such as Kuaishou)

2. I get more satisfaction from TikTok

3. Compared with similar social software (such as Kuaishou), using TikTok is more important

4. Compared with similar social software (such as Kuaishou), TikTok is irreplaceable

5. TikTok is my favorite local social software

6. If possible, I would like to spend more time on TikTok

Social identity

1. TikTok is of great significance to me (such as gaining recognition and respect/being an anchor or influencer)

2. I really like TikTok

3. I strongly agree with TikTok

4.TikTok can reflect my personality and style

5. I have a crush on TikTok

References

, MIN

, 윤민석. An Empirical Study on the Continued Use and Information Sharing of UGC through the Flow Theory: Focusing on TikTok in China. The Journal of Internet Electronic Commerce Research. 2019;19(6):19–40.

Mouakket

. The role of personality traits in motivating users’ continuance intention towards Facebook: Gender differences. The Journal of High Technology Management Research. 2018;29(1):124–40.

Lin

, Lu

. Why people use social sites Networking: An Empirical Study Integrating Network externalities and Motivation Theory. Computers in Human Behavior. 2011;27(3):1152–61.

Rajabion

, Nazari

, Bandarchi

, Farashiani

, Haddad

. Knowledge sharing mechanisms in virtual communities: A review of the current literature and recommendations for future research. Human Systems Management. 2019;48(4):365–84.

Lin

, Featherman

, Sarker

. Understanding factors affecting users’ social networking site continuance: A gender difference perspective. Information & Management. 2017;54(3):383–95.

Behringer

, Sassenberg

. Introducing social media for knowledge management: Determinants of employees’ intentions to adopt new tools. Computers in Human Behavior. 2015;48:290–6.

Magsamen-Conrad

, Dillon

. Mobile technology adoption across the lifespan: A mixed methods investigation to clarify adoption stages, and the influence of diffusion attributes. Computers in Human Behavior.. 2020;112(20):1–61.

Joo

, Choi

. Understanding users’ continuance intention to use online library resources based on an extended expectation-confirmation model. Electronic Library. . 2016;34(4):554–71.

Hsu

, Chang

, Chuang

. Understanding the Determinants of Online Repeat Purchase Intention and Moderating Role of Habit: The Case of Online Groupbuying in Taiwan. International Journal of Information Management. 2015;35(1):45–56.

10.

Zhao

, Ni

, Zhou

. What factors influence the mobile health service adoption? A meta-analysis and the moderating role of age. International Journal of Information Management. 2018;43:342–50.

11.

Nabavi

, Taghavi-Fard

, Hanafizadeh

, Taghva

. Information technology continuance intention: a systematic literature review. International Journal of E-Business Research. 2016;12(1):58–95.

12.

Hamari

, Keronen

. Why do people play games? A meta-analysis. International Journal of Information Management. 2017;37(3):125–41.

13.

Oghuma

, Libaque-Saenz

, Wong

, Chang

. An Expectation Confirmation Model Continuance of at intention to use Mobile Instant Messaging. Telematics & Informatics. 2016;33(1):34–47.

14.

Kim

, Malhotra

, Predicting system usage from intention and past use: Scale issues in the predictors, Decision Sciences. 2015;36(1):187–96.

15.

Cheng

, Jin

. Analysis on the current situation of domestic IT/IS users’ continuous use research. Information Science. 2018;36(4):171–6.

16.

Taylor

, Todd

. Decomposition and crossover effects in the theory of planned behavior: A study of consumer adoption intentions. International Journal of Research in Marketing. 1995;12(2):137–55.

17.

Naseri

, Rajabion

, Zaravar

, Khordehbinan

, Moradi

. A new model for evaluating the impact of ergonomic architectural design of gym, social network and digital knowledge on the desire of students for sport activities. Human Systems Management. 2019;38(40):347–55.

18.

Naseri

, Rajabion

, Zaravar

, Khordehbinan

, Moradi

19.

Zhao

, Wang

. Discrimination of the concept of attachment in information system and the preliminary study of its effect. Information Science. 2016;34(11):60–5.

20.

Carrollb

, Ahuvia

. Some antecedents and outcomes of brand love. Marketing Letters. 2006;17(2):79–89.

21.

, Browne

, Wetherbe

. Why Do Internet Users Stick with a Specific Web Site? A Relationship Perspective. International Journal of Electronic Commerce. 2006;10(4):105–41.

22.

Chen

. Research on the governance mechanism of e-commerce platforms to seller user companies: from the perspective of brand attachment theory. Journal of Shanghai University of Finance and Economics. 2019;21(02):106–23.

23.

Zhao

, Zhou

, Yu

, Liu

. Based on the loyalty perspective of virtual community continuous use model construction-beyond the cognitive judgment paradigm. Prediction. 2012;31(5):14–20.

24.

Yang

, Zhang

, Gao

, Hu

, Liu

. Development of Social Attachment Multidimensional Scale and Validity Test of Its Reliability. Social Psychology.. 2020;51(4):255–72.

25.

, Liu

, Zhang

, Wang

. Proactive personality and cross-cultural adjustment: Roles of social media usage and cultural intelligence. International Journal of Intercultural Relations. 2020;74:42–57.

26.

Jiang

, Dong

. Research on the continuous mechanism of the relationship between consumers and online stores based on attachment theory. Commercial Economics and Management. 2015;4:53–62.

27.

, Gu

, Liu

, Huang

. The moderating role of social media usage in the relationship among multicultural experiences, cultural intelligence, and individual creativity. Information Technology and People. 2017;30(2):265–81.

28.

Tsai

. Place attachment and tourism marketing: Investigating international tourists in Singapore. International Journal of Tourism Research. 2012;14(2):139–52.

29.

Ramkissoon

. Authenticity, satisfaction, and place attachment: A conceptual framework for cultural tourism in African island economies. Development Southern Africa. 2015;32(3):292–302.

30.

Brown

, Smith

, Assaker

. Revisiting the host city: An empirical examination of sport involvement, place attachment, event satisfaction and spectator intentions at the London Olympics. Tourism management. 2016;55:160–72.

31.

Hwang

, Lin

, Shin

. Knowledge system commitment and knowledge sharing intention: The role of personal information management motivation. International Journal of Information Management. 2018;39:220–27.

32.

Zhou

, Jin

, Fang

, Vogel

. Toward a theory of perceived benefits, affective commitment, and continuance intention in social virtual worlds: cultural values (indulgence and individualism) matter. European Journal of Information Systems. 2015;24(3):247–61.

33.

Zhou

, Fang

, VogeiL

, Jin

, Zhang

. Attracted to or locked in? predicting continuance intention in social virtual world services. Journal of Management Information Systems. 2012;29(1):273–306.

34.

Chou

, Hung

. Understanding knowledge outcome improvement at the post-adoption stage in a virtual community. Information Technology & People. 2016;29(4):774–806.

35.

Chen

, Lu

, Wang

. Customers’ purchase decision-making process in social commerce: A social learning perspective. International Journal of Information Management. 2017;37:627–38.

36.

Kim

, Son

. Out of dedication or constraint? a dual model of post-adoption phenomena and its empirical test in the context of online services. MIS Quarterly. 2009;33(1):49–70.

37.

Fredrickson

. What good are positive emotions? Review of General Psychology. 1998;2(3):300–19.

38.

Zhu

, Zhang

, Yao

. Research on the impact of perceived value on mobile short video attachments: Based on the perspective of mediating effect of network belonging and network privacy concerns. Journalism. 2019;7:68–82.

39.

Chen

, Lu

, Wang

, Zhao

, Li

. What drives content creation behavior on SNSs? A commitment perspective. Journal of Business Research. 2013;66(12):2529–35.

40.

Hwang

, Lin

, Shin

. Knowledge system commitment and knowledge sharing intention: The role of personal information management motivation. International Journal of Information Management. 2018;39:220–7.

41.

Mende

, Bolton

. Why Attachment Security Matters: How Customers’ Attachment Styles Influence Their Relationships With Service Firms and Service Employees. Journal of Service Research. 2011;14(3):285–301.

42.

Choi

. Information systems attachment: an empirical exploration of its antecedents and its impact on community participation intention. Journal of the American Society for Information Science and Technology.. 2013;64(11):2354–65.

43.

Chen

, Nguyen

, Oncheunjit

. Understanding continuance intention in traffic-related social media Comparing a multi-channel information community and a community-based application. Internet Research. 2019;30(2):539–73.

44.

, Bai

XCZ

, Fiske

. Admired Rich or Resented Rich? How Two Cultures Vary in Envy. Journal of Cross-Cultural Psychology. 2019;49(7):1114–43.

45.

Wang

. Humor and camera view on mobile short-form video apps influence user experience and technology-adoption intent, an example of TikTok (DouYin). Computer in Human Behavior. 2020;110:1873–7692.

46.

Paolacci

, Chandler

. Inside the Turk: Understanding Mechanical Turk as a participant pool. Current Directions in Psychological Science. 2014;23:184–8.

47.

Thong

JYL

, Hong

, Tam

. The effects of post-adoption beliefs on the expectation-confirmation model for information technology continuance. Human-Computer Studies. 2006;64(9):799–810.

48.

Kyle

, Graefe

, Manning

. Testing the dimensionality of place attachment in recreational settings. Environment and Behavior. 2005;37:153–77.

49.

Bagozzi

, Yi

. On the evaluation for structural equation models. Journal of the Academy of Marketing Science. 1998;16:74–94.

50.

Kline

. Principles and practice of structural equation modeling (2nd ed.). New York: Guilford Press 2005.

51.

Anderson

, Gerbing

. Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin. 1998;10(3):411–23.

52.

Hair

, Anderson

, Tatham

, Black

. Multivariate data analysis (5th ed.), Upper Saddle River, NJ: Prentice Hall 1988.

53.

Nunnally

, Bernstein

. Psychometric theory (3rd ed.), New York: McGraw-Hill 2016.

54.

Fornell

, Lacker

. Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research. 1994;18(1):39–50.

55.

Torkzadeh

, Koufteros

, Pflughoeft

. Confirmatory analysis of computer self-efficacy. Structural Equation Modeling. 2003;10(2):263–75.

56.

Hancock

, Nevitt

. Bootstrapping and the identification of exogenous latent variables within structural equation models. Structural Equation Modeling. 1999;6:394–9.

57.

Jackson

, illaspy

, Purc-Stephenson

. Reporting practices in confirmatory factor analysis: An overview and some recommendations. Psychological Methods.. 2009;14(1):6–23.

58.

Bentler

. EQS structural equations program manual, Los Angeles: BMDP Statistical Software 1989.

59.

Bentler

. Comparative fit indices in structural models, Psychological Bulletin. 1990;107(2):238–46.

60.

Bollen

, Stine

. Bootstrapping goodness-of-fit measure in structural equation models. Sociological Methods and Research. 1992;21:205–29.

61.

Hair

, Black

, Babin

, Anderson

. Multivariate data analysis: A global perspective (7th ed.), Pearson Education International 2010.

62.

Bentler

, Bonett

. Significance tests and goodness-of-fit in the analysis of covariance structures. Psychological Bulletin. 1980;88(3):588–606.

63.

Lewicka

. Place attachment: How far have we come in the last 40 years? Journal of Environmental Psychology. 2011;31(3):207–30.

64.

Goel

, Johnson

, Junglas

, Ives

. From Space to Place: Predicting Users’ Intention to Return to Virtual Worlds. MIS Quarterly. 2011;35(3):749–71.