اثر حسادت کاربران و شرم بر روی مقایسه اجتماعی که در ارائه خدمات شبکه های اجتماعی رخ می دهد

Abstract In the context of the social network service environment, we explore how discrete emotions—envy and shame, in particular—may mediate the effects of social comparison on behavior intention and psychological responses. Based on the survey responses of 446 university students, the results suggest that social comparison to media figures correlates with a range of emotional responses as well as with behavioral intention and psychological responses. Envy maintained a significantly greater association with switch intention as a behavioral intention compared to shame. Conversely, shame was significantly related to burnout as a psychological response. Further, mediational analyses were consistent with the argument that envy and shame mediate the social comparison–outcome variables relationship. This research helps to illuminate the driving mechanism for the emotional effect that social comparison on social network service could elicit from a user. This predicts the nature of the behavioral and psychological outcome associated with the comparison and has implications for an enhanced understanding of the way in which the unique social network service communication environment may stimulate this process.

1. Introduction A social network service (SNS) is known as a good way for getting in touch with other people. Through SNS services, people receive news from their friends, feel that they remain connected all times, and share their interests and activities. For instance, some SNS users even wanted to share their happy moments via a campaign called “100 happy days” that was run by Instagram (i.e., one of the online photo-sharing SNS services). However, is it true that people become happy by watching other people’s happiness and delightful moments through SNS? We aim to determine whether people are able to satisfy their life in the light of others’ lives, as observed using SNS services. Numerous studies indicated that one person’s happiness and delightful moments do not necessarily positively affect the lives of others. Panger (2014) suggested that SNS is strongly affected by the Easterlin Paradox which suggests that an increase in income is not necessarily associated with happiness; in this paradox, the need to use social comparison to achieve satisfaction is assumed to decline or even disappear along with an increase in income ( Panger, 2014, p. 2096). Likewise, the use of SNS to achieve social comparison, as commonly occurs between SNS users, seems to reduce their satisfaction with life. SNS has the effect of making social comparison easier and increasing the desire to make comparisons with others. Based on cognitive emotion theory, which is a recent topic of active discussion, social comparison causes users to become stressed and can be stimulating, thereby lowering users’ satisfaction. In the present study, we suggest that social comparisons between SNS users may negatively affect those involved, eventually leading to negative behavioral intention and psychological responses.

4. Results 4.1. Model comparison The current research tested the structural model using AMOS 18.0. It provided evidence of a good model fit (χ2 = 483.741 (df = 127, p = .000), GFI = .904, CFI = .951, NFI = .934, TLI = .940, RMSEA = .079). Before evaluating the structural model, we implemented a comparison between the hypotheses model and the alternative model. The results are shown in Table 5 where the model-fit indices for the hypotheses model and the alternative model are provided. Table 5. Model comparison. Model χ2 df GFI CFI NFI TLI RMSEA Proposed model 483.741 127 0.904 0.951 0.934 0.94 0.079 Alternative model 526.922 129 0.893 0.945 0.928 0.935 0.083 Table options Against this proposed model, partially mediated by coping behaviors and based on the theory, we tested an alternative model, which suggests that the path between social comparison and the outcome variables are fully mediated by emotions. As shown in Table 5, the fit indices supported the proposed model; the chi-square value of which is significantly smaller than that of the alternative model (Δχ2 = 43.181 > χ20.05 (2) = 5.99). Even though the proposed model is more complex than the alternative model, the difference in the chi-square values is sufficiently large to offset a reduced degree of freedom. Moreover, the proposed model showed direct and indirect effect paths effect from independent variables and mediator variables to dependent variable and several fit indices show better results compared to the alternative model. Therefore, it can be concluded that the proposed model is superior to the competing model. 4.2. Test of the structural model This research aims to examine how social comparison influences SNS users’ switch intention and burnout as outcomes as well as how emotion mediates the relationship between social comparison and outcomes. In order to evaluate these effects at once, SEM (structural equation modeling) setting is necessary. The measurement model above will be transformed into a structural model for validating the proposed hypotheses. For the evaluation of this model, the coefficient of determination (R2) and the significant levels of the path-coefficient need to be observed ( Chin, 1998). Based on these data, we tested the hypotheses regarding the social comparison–emotion-outcome model. To validate our hypotheses, we used the maximum likelihood method of AMOS 18. Fig. 2 displays the standardized path coefficients, path significances, and variance (R2) for each path, all supported by the path analysis results, except H1A, H2B2, and H2C1 (also see Table 6). In more detail, for appraisal of the proposed model, the coefficient of determination (R2), and significance levels of the path coefficients need to be observed ( Chin, 1998). Regarding the variance, R2 for switch intention reached 6% and for burnout, it was 47%. Concerning the path coefficients, the social comparison positively influenced burnout. However, social comparison does not influence switch intention. Therefore, hypothesis H1B is confirmed. Test of structural model. Fig. 2. Test of structural model. Figure options Table 6. Test of structural model. Hypothesis Path Estimate C.R. P H1A Social comparison → switch intention 0.073 1.071 0.284 H1B Social comparison → burnout 0.324 5.978 ⁎⁎⁎ H2A1 Social comparison → envy 0.579 12.186 ⁎⁎⁎ H2A2 Social comparison → shame 0.427 8.462 ⁎⁎⁎ H2B1 Envy → switch intention 0.196 3.119 0.002 H2B2 Envy → burnout −0.023 −0.477 0.634 H2C1 Shame → switch intention 0.042 0.748 0.454 H2C2 Shame → burnout 0.500 10.457 ⁎⁎⁎ ⁎⁎⁎ P < 0.001. Table options For an analysis of the mediating effect of emotion, first we analyzed the mediating effect of envy. The path from social comparison to envy was positively significant, meaning that hypothesis H2A1 was supported. The paths from envy to the negative outcome variables (switch intention and burnout) were only significant for switch intention, but not for burnout. Thus, only hypothesis H2B1 was supported. It was pre-qualified that the direct path from social comparison to burnout could only be significant, so that envy performs a mediating role between social comparison and switch intention, but does not play the same mediating role for burnout. Second, the mediating role of shame for the path from social comparison to the negative consequences was partially significant. The hypothesis H2A2 was therefore supported. Only the path from shame to burnout was significant, but not the path to switch intention. Thus, only H2C2 was supported. It was also pre-qualified that the direct path from social comparison to burnout could only be significant, so that shame performed a partial mediating role between social comparison and burnout, but did not perform the same mediating role for switch intention. Furthermore, the Sobel (1982) test was performed to examine the significance of the mediating effect of emotion. The Sobel (1982) test confirmed that the path between social comparison and switch intention is mediated by envy (Z = 3.00, p = 0.002), and the path between social comparison and burnout is mediated by shame (Z = 6.53, p = 0.000). Next, the strength of the effect (f2) was calculated together with the coefficient of determination (R2) for both the full model (including the mediating variable) and the reduced model (excluding the mediating variable) to determine the appropriateness of the mediating parameter in the covariance model. Table 7 shows that when either envy or shame are input as the mediator, the coefficient of determination for burnout and switch intention can be identified as higher compared with the basic model (M1). First, when it comes to burnout, the mediation model (M4) has the best coefficient of determination compared with the basic model (M1), which only has independent variables (social comparison). Although switch intention was only slightly affected by the emotional mediators, focusing on one mediator model (M2) enabled us to confirm a different emotional role in the mediation model. Likewise, we can confirm that shame as an emotion also has a more apparent role as a mediator in one of the mediator models (M3). Table 7. The strength of effect (Chin, 1998). Variable M1 M2 M3 M4 DV Burnout IV Social comparison 0.515⁎⁎⁎ 0.33⁎⁎⁎ 0.314⁎⁎⁎ 0.324⁎⁎⁎ MV Envy 0.326⁎⁎⁎ −0.023ns Shame 0.499⁎⁎⁎ 0.5⁎⁎⁎ R2 0.265 0.337 0.474 0.479 Effect size Low High High (f2) 0.11⁎⁎⁎ 0.40⁎⁎⁎ 0.41⁎⁎⁎ DV Switch intention IV Social comparison 0.202⁎⁎⁎ 0.07ns 0.138⁎⁎ 0.073ns MV Envy 0.229⁎⁎⁎ 0.196⁎⁎⁎ Shame 0.152⁎⁎ 0.042ns R2 0.041 0.075 0.059 0.069 Effect size Low Low Low (f2) 0.04⁎⁎⁎ 0.02⁎⁎⁎ 0.03⁎⁎⁎ Note: f2 mean effect size; Chin (1998) interprets size as follows: >0.35 = high effect; >0.15 = medium effect; >0.02 = low effect. ⁎⁎ P < 0.05. ⁎⁎⁎ P < 0.001.