Upward social comparison on network and alcohol dependence in college students: the chain mediation impacts of relative deprivation and anxiety Authors: Shuna Li, Kaijie Zhang, Kaizhen Kong, Hua Niu, Zhenli Jiang, Hongbo Guo Published: 2024 ABSTRACT BackgroundNowadays, contemporary society has placed growing emphasis on the alcohol drinking behavior of college students.MethodsThis study employed a cross-sectional design to assess the relationship between upward social comparison (USC) on network and college students drinking behavior, while exploring the mediation impact of relative deprivation (RD) and anxiety. A total of 329 college students (Mage = 21.37 years old; 16.11% women) were enrolled to complete a set of questionnaires evaluating USC on network, RD, anxiety and alcohol dependence.ResultsThe results showed that (1) USC on network, RD, anxiety and alcohol dependence were significantly, positively correlated in pairs, and (2) RD and anxiety exhibited a significant chain mediating role between USC on network and alcohol dependence, and USC on network positively predicted alcohol dependence via the mediating effect of anxiety.ConclusionsThis study reveals that USC on network positively predicts alcohol dependence via the mediating effect of anxiety, and that RD and anxiety play a chain-mediating role between USC on network and alcohol dependence. Introduction Drinking poses a significant health risk, particularly among college students who frequently engage in this behavior [1]. During college, students are in a critical period of physical and mental development, making them susceptible to external influences that can increase various health risks [2, 3]. With the advent of the information age, upward social comparison (USC) on network has become prevalent among college students [4]. Previous studies have shown that negative emotions such as depression and anxiety are closely related to USC on network [5], which may, in turn, may predict adverse behaviors like alcohol dependence [6]. Issues related to relationships, mental health, and emotions are often associated with a higher frequency or volume of drinking [7, 8]. Therefore, this study focuses on college students as subjects to investigate the underlying mechanisms connecting USC on network and alcohol dependence. Understanding the psychological mechanisms behind alcohol dependence could be instrumental in addressing this issue among college students. Theoretical basis The basic principle of social comparison emphasizes that individuals need an accurate and stable self-evaluation. Typically, people assess themselves using impartial and non-social criteria. In the absence of such objective data, they rely on other information as reference to make comparisons with others [9]. USC involves comparing oneself to those perceived as “better” in some way, which can be detrimental to individual development. Although some studies have indicated that USC can have positive effects, others have highlighted its negative consequences [9–11]. Researchers believe that USC is directly linked to various health-related and behavioral problems [12–14]. USC on network represents a new form of social comparison, but its impact on alcohol consumption behavior is still unclear, and the underlying mechanisms warrant further investigation. Relative deprivation (RD) theory explains RD as a perception of being disadvantaged, where individuals compare their situation to a standard or reference point and perceive a sense of deprivation [15]. Social comparison is a core psychological process of RD [16], with individuals evaluating their own situation and status mainly by comparing themselves to others. RD encompasses both cognitive and emotional components; vulnerable groups often feel deprived of basic rights, leading to feelings of anger and resentment [17], which can adversely affect their physical and mental health [15, 18]. USC on network and alcohol dependence Based on the principle of social comparison, USC suggests that it is easier for individuals to form self-evaluations by comparing themselves to others in better circumstances or with superior abilities. Researchers have identified two distinct effects based on the motivations and expected outcomes of social comparison: the assimilation effect and the contrast effect [19]. The assimilation effect occurs when self-evaluation aligns with that of the comparison target, while the contrast effect occurs when self-evaluation moves in the opposite direction. Upward comparisons are more common in networks, where internet users pay more attention to the characteristics of individuals who outperform them. This often forms feelings of stress and difficulty in coping, a phenomenon known as the contrast effect [20]. Numerous studies have examined the negative impacts of upward comparisons, and found that USC on network is associated with problematic behaviors and predict health issues such as drinking [13, 21]. For instance, Litt reported a significant relationship between social comparison and alcohol-related negative consequences, with high social comparison predicting higher reports of such consequences [14]. Therefore, this study hypothesizes that USC on network may be associated with an increased likelihood of drinking problems. Hypothesis 1 posits that USC on network can positively predict alcohol dependence. The mediation impact of RD The basic principle of social comparison suggests that the contrast effect of USC affects individuals’ cognition, emotion and behavior [9, 22], potentially influenced by RD [23, 24]. USC often predicts feelings of RD, which can bring negative outcomes [25]. For example, USC on network is negatively correlated with college students' life satisfaction through the mediating effect of RD [24], and USC on network can predicted non-suicidal self-injury behavior through RD [26]. These findings support the perspective of this study, suggesting that RD may predict negative cognition or behavior during the USC process. Thus, we propose the following hypothesis 2: RD may mediate the relationship between USC on network and alcohol dependence. The mediating role of anxiety Based on the contrast effect of social comparison, USC negatively impacts the positive development of individuals [9]. Current studies agree that USC can significantly positively predict anxiety, as can USC on network [22, 27, 28]. And research on young adults show that anxiety is a robust predictor of drinking and related problems [29, 30]. So we speculate that USC may have negative effects on individuals through the mediating role of anxiety. And some studies back this up, they found that negative emotion is a mediating factor in how USC on network influences individuals’ negative behaviors such as impulsive buying [31, 32]. Therefore, hypothesis 3 is proposed: anxiety has a mediation impact between USC on network and alcohol dependence. The chain mediation impacts of RD and anxiety The USC on network may influence alcohol dependence among college students through the chain mediation of RD and anxiety. According to the social identity theory's positive distinctiveness principle, individuals are fundamentally motivated to seek positive self-esteem by comparing themselves with others, establishing their status advantage, and thus achieving high self-esteem and self-satisfaction [33]. The advent of the information age has significantly expanded the avenues and scope for such comparisons. However, engaging in negative USC on network can lower an individual’s self-esteem and satisfaction [34], and ultimately promote the emergence of RD feelings [35]. This frustration may predict various types of behavior such as aggression and social deviance [36]. Previous research has found that individuals with RD are prone to persistent and distressing states of anxiety [37], and may attempt to escape this frustrating reality through deviant behaviors such as smoking, alcohol dependence, and binge eating [38–40]. Hence, college students who perceive higher RD due to the influence of USC on network are more likely to experience anxiety, increasing the likelihood of alcohol dependence. Based on the above analysis, hypothesis 4 is proposed: RD and anxiety exert a chain mediation impact between USC on network and alcohol dependence. Method Subjects The research sample comprised 3,096 first-year students (1,752 men) from a university located in an eastern city of China, who participated in a school-organized psychological health assessment. Under the supervision of their teachers, students completed the assessment on computers upon entering the university. With the consent of the students, their drinking behavior over the past year was assessed through the question “Have you consumed alcohol in the past year?” A total of 364 students who reported drinking behavior were required to complete subsequent assessments. According to the Alcohol Dependence Scale scoring criteria, a score of 1 or higher indicates varying degrees of alcohol dependence. After excluding samples with an ADS score of 0, those with patterned responses, and those with excessive missing data, 329 valid samples (276 men) were collected, resulting in a recovery rate of 90.38%. The mean age of all participants was 21.37 ± 2.54 years. Procedures Utilizing cluster sampling, a survey was conducted among first-year college students of Shandong University of Science and Technology in Qingdao, China. By inquiring whether the participants had consumed alcohol within the past year, those who had consumed alcohol were screened out for further formal investigation and research. The study protocol was approved by the Institutional Review Board of the same university. All participants signed informed consent prior to data collection. Measures USC on network The Upward comparison subscale of Iowa-Netherlands Comparison Orientation Measure (INCOM) which is compiled by Gibbons and Buunk [36], and translated by Bai, Liu and Liu [40], was applied by this study. In order to make the investigate more targeted, the questions of the scale were limited to the network background by Lian [41, 42]. Such as this scale detect the behavior of USC on network of the subjects by investigating the frequency of individuals compared with those who are more successful than themselves in the network. The questionnaires contained 6 items, and scored as follows: highly inconsistent (1) to highly consistent (5). High scores indicated a strong USC on network. This scale has been extensively applied in China [43]. Zhang and Liu also employed this method for measurement, and good reliability and validity was obtained [28]. The Cronbach's α of USC on network was 0.92. The confirmatory factor analyses indicated that the items could represent the corresponding factors (χ2/df = 3.24, CFI = 0.99, NFI = 0.98, IFI = 0.99, RMSEA = 0.08). RD This research used the RD scale, as compiled by Ma [44]. The scale consisted of 4 items, which were scored as follows: highly inconsistent = 1, highly consistent = 6. An example item was “I always feel that others have possessed what should belong to me.”Higher scores indicated higher levels of RD. The scale showed excellent reliability and validity. It has been widely used among college student groups in China [44, 45]. The Cronbach’s α of RD was 0.76. The confirmatory factor analyses indicated that the items could represent the corresponding factors (χ2/df = 2.70, CFI = 0.99, NFI = 0.99, IFI = 0.99, RMSEA = 0.07). Anxiety Anxiety was measured using the 20-item Chinese version of the Self-Rating Anxiety Scale which was compiled by Zung [46]. An example item was “I feel more nervous and anxious than usual.” Responses were obtained on a 4-point Likert scale ranging from 1 (almost never) to 4 (almost always). Higher scores represent greater degree of anxiety. The scale has demonstrated satisfactory reliability and validity [47]. The Cronbach’s α of anxiety was 0.88. The confirmatory factor analyses indicated that the items could represent the corresponding factors (χ2/df = 2.48, CFI = 0.91, NFI = 0.86, IFI = 0.91, RMSEA = 0.07). Alcohol dependence This study used the Alcohol Dependence Scale (ADS) which was selected from the ‘Mental Health Rating Scale Manual’ edited by Wang to estimate the level of alcohol dependence [48]. The scale comprised of 25 items. Scores range from 0 to 47, where 0 indicates the absence of alcohol dependence, 1–13 indicate mild, 14–21 moderate, 22–30 substantial, and 31–47 severe alcohol dependence. The total score of which was the sum of each item in ADS. An example item was “Do you always want to drink?”, with higher scores indicating greater alcohol dependence. The scale has demonstrated satisfactory reliability and validity [49]. The Cronbach’s α of alcohol dependence was 0.85. The confirmatory factor analyses indicated that the items could represent the corresponding factors (χ2/df = 3.82, CFI = 0.79, NFI = 0.74, IFI = 0.80, RMSEA = 0.09). Data analysis The extreme values and multicollinearity within the sample datasets were examined. The findings of collinearity assessment indicated that the variance inflation factor VIF of all predictor variables was less than 10 and the tolerance was greater than 0.1, indicating the absence of multicollinearity. Then, 0.0001 was selected as the threshold for Mahalanobis distance p-value, and the variable lower than this value was considered to be multivariate extreme value. Consequently, five multivariate extreme values were identified and ultimately retained to accurately represent the distribution of the sample. Firstly, the correlations among variables were evaluated by Pearson correlation analysis. If a significant correlation exists between the variables, the analysis can proceed. Furthermore, independent samples t-tests were used to analyze each variable to compare gender differences. Secondly, the regression coefficients of upward comparisons on network-alcohol dependence, USC on network-RD-anxiety-alcohol dependence in turn were verified. Thirdly, the chain mediating effect of each variable was determined using the Macro PROCESS in SPSS. The presence of chain mediation can be confirmed if the 95% confidence interval (CI) of the bootstrap analysis does not overlap with the value of 0. Results Common method variance Harman’s one-factor test was carried out in this study. The findings demonstrated that there were fourteen factors with an eigenvalue of > 1, while the first factor accounted for 15.81% of the total variability, falling short of the 40% threshold [50]. This indicates that the issue of common method bias is relatively minor. Correlations and descriptive statistics The descriptive statistics and correlation analysis showed that USC on network was positively correlated with RD, anxiety and alcohol dependence. RD was positively correlated with anxiety and alcohol dependence. Anxiety was positively correlated with alcohol dependence. The results displayed in Table 1 suggest that it is appropriate to proceed with further mediating effect analysis. Additionally, to further control the impact of gender differences on alcohol dependence and consequently mitigate the influence of these individual factors on the overall study, an independent samples t-test were performed to investigate potential gender (men vs. women) differences in the key study variables (as shown in Table 2). An independent samples t-test indicated that there was a significant difference in RD (t = 2.24, p < 0.05), alcohol dependence (t = 3.74, p < 0.001) and there was no gender difference in terms of USC on network (t = -0.49, p = 0.63), SAS (t = 0.09, p = 0.93) between men and women. Table 1The result of data’s description and correlation analysisMSD12345Age21.372.541USC on network14.925.820.011RD12.433.57-0.050.42**1Anxiety31.907.920.070.31**0.34**1Alcohol dependence3.014.110.010.22**0.15**0.20**1M Mean, SD Standard Deviation, USC Upward Social Comparison, RD Relative Deprivation *p < 0.05. **p < 0.01. ***p < 0.001Table 2Gender difference analysis of USC on network, RD, Anxiety and Alcohol dependence (N = 329)Male(n = 276)Female(n = 53)tpMSDMSDUSC on network14.865.9615.285.05-0.490.63RD12.633.5611.433.502.24 < 0.05Anxiety31.928.0931.817.020.090.93Alcohol dependence3.214.411.961.483.74 < 0.001M Mean, SD Standard Deviation, USC Upward Social Comparison, RD Relative Deprivation*p < 0.05, **p < 0.01 and ***p < 0.001 Table 1The result of data’s description and correlation analysisMSD12345Age21.372.541USC on network14.925.820.011RD12.433.57-0.050.42**1Anxiety31.907.920.070.31**0.34**1Alcohol dependence3.014.110.010.22**0.15**0.20**1M Mean, SD Standard Deviation, USC Upward Social Comparison, RD Relative Deprivation *p < 0.05. **p < 0.01. ***p < 0.001 The result of data’s description and correlation analysis M Mean, SD Standard Deviation, USC Upward Social Comparison, RD Relative Deprivation *p < 0.05. **p < 0.01. ***p < 0.001 Table 2Gender difference analysis of USC on network, RD, Anxiety and Alcohol dependence (N = 329)Male(n = 276)Female(n = 53)tpMSDMSDUSC on network14.865.9615.285.05-0.490.63RD12.633.5611.433.502.24 < 0.05Anxiety31.928.0931.817.020.090.93Alcohol dependence3.214.411.961.483.74 < 0.001M Mean, SD Standard Deviation, USC Upward Social Comparison, RD Relative Deprivation*p < 0.05, **p < 0.01 and ***p < 0.001 Gender difference analysis of USC on network, RD, Anxiety and Alcohol dependence (N = 329) M Mean, SD Standard Deviation, USC Upward Social Comparison, RD Relative Deprivation *p < 0.05, **p < 0.01 and ***p < 0.001 Test for the mediation model This study successfully fulfilled the statistical criteria necessary for evaluating the regulatory influences of RD and anxiety. The sampling process was repeated 5000 times using the bootstrapping method, employing the SPSS macro program compiled by Hayes. Subsequently, a 95% unbiased corrected CI was constructed. If the 95% CI of the standardized path coefficient does not contain 0, it indicates a significant mediation effect. Model 6 from the PROCESS plug-in was utilized to employ the chain mediation model, taking into account two mediation variables. Considering that there are significant differences in alcohol dependence and relative deprivation between different gender samples, this study controlled for demographic factors (e.g., age and gender). Following adjustment for age and gender, the regression model data demonstrated that the standardized regression coefficient of USC on network and alcohol dependence is β=0.22, with a significance level of p<0.001. Therefore, USC on network has a significant positive effect on alcohol dependence. Upon including RD and anxiety in the regression equations, the results of the regression analysis indicated that USC on network had a significant positive prediction of RD (β=0.42, p<0.001), anxiety (β=0.20, p<0.001), and alcohol dependence (β=0.17, p<0.01). In addition, RD positively predicted anxiety (β=0.27, p<0.001), while anxiety positively predicted alcohol dependence (β=0.15, p<0.05), as presented in Table 3. Table 3The chained mediation models of relative deprivation (RD) and anxiety in the relationship between Upward social comparison (USC) on network and alcohol dependence after controlling for the effects of gender and age (n = 329)Predictor variableOutcome variableRR2FβtSE95%CIModel 1 USC on networkRD0.440.2025.95***0.42***8.400.07[0.20, 0.32] Gender-0.14**-2.730.48[-2.27, -0.37] Age-0.06-1.180.03[-0.22, 0.06]Model 2 USC on networkAnxiety0.400.1615.45***0.20***3.570.08[0.12, 0.42] RD0.27***4.710.13[0.34, 0.84] Gender0.030.491.11[-1.64, 2.72] Age0.081.660.16[-0.05, 0.58]Model 3 USC on networkAlcohol dependence0.280.085.67***0.17**2.820.04[0.04, 0.20] RD0.010.180.07[-0.13, 0.15] Anxiety0.15*2.500.03[0.02, 0.14] Gender-0.11*-2.120.60[-2.47, -0.09] Age-0.004-0.070.09[-0.18, 0.17]All estimated coefficients are standardized.SE Standard Error, CI Confidence Interval, USC Upward Social Comparison, RD Relative Deprivation*p < 0.05, **p < 0.01 and ***p < 0.001 Table 3The chained mediation models of relative deprivation (RD) and anxiety in the relationship between Upward social comparison (USC) on network and alcohol dependence after controlling for the effects of gender and age (n = 329)Predictor variableOutcome variableRR2FβtSE95%CIModel 1 USC on networkRD0.440.2025.95***0.42***8.400.07[0.20, 0.32] Gender-0.14**-2.730.48[-2.27, -0.37] Age-0.06-1.180.03[-0.22, 0.06]Model 2 USC on networkAnxiety0.400.1615.45***0.20***3.570.08[0.12, 0.42] RD0.27***4.710.13[0.34, 0.84] Gender0.030.491.11[-1.64, 2.72] Age0.081.660.16[-0.05, 0.58]Model 3 USC on networkAlcohol dependence0.280.085.67***0.17**2.820.04[0.04, 0.20] RD0.010.180.07[-0.13, 0.15] Anxiety0.15*2.500.03[0.02, 0.14] Gender-0.11*-2.120.60[-2.47, -0.09] Age-0.004-0.070.09[-0.18, 0.17]All estimated coefficients are standardized.SE Standard Error, CI Confidence Interval, USC Upward Social Comparison, RD Relative Deprivation*p < 0.05, **p < 0.01 and ***p < 0.001 The chained mediation models of relative deprivation (RD) and anxiety in the relationship between Upward social comparison (USC) on network and alcohol dependence after controlling for the effects of gender and age (n = 329) All estimated coefficients are standardized. SE Standard Error, CI Confidence Interval, USC Upward Social Comparison, RD Relative Deprivation *p < 0.05, **p < 0.01 and ***p < 0.001 This research performed a mediation analysis to examine whether RD and anxiety mediated the effect of USC on network on alcohol dependence. The mediation effect analysis revealed that anxiety significantly affected the relationship between USC on network and alcohol dependence. RD and anxiety exhibit a significant chain mediation impact between USC on network and alcohol dependence. The standardized mediation effect value was 0.05, and the effect size was 0.23. Notably, the 95% CI of the standardized path coefficient of the indirect effects 2 and 3 did not contain the value of 0, suggesting that the mediating effects of these two paths are significant. To be specific, the indirect effect 2 was obtained from USC on network to anxiety to alcohol dependence (the effect value and size were 0.03 and 0.14, respectively); while the indirect effect 3 was obtained from USC on network to RD to anxiety to alcohol dependence (the effect value and size were 0.02 and 0.09, respectively). However, the 95% confidence intervals of the standardized path coefficient of the indirect effect 1, which was obtained from USC on network to RD to alcohol dependence contained 0. Hence, the mediating effect of this path is not significant (Table 4). Table 4Mediation effects of RD, AnxietyStandardEstimateBoot SEBoot LLCIBoot ULCITotal Ind0.050.020.020.10Ind10.0050.02-0.030.05Ind20.030.010.010.06Ind30.020.010.010.03Ind1: USC on network → RD → Alcohol dependence; Ind2: USC on network → Anxiety → Alcohol dependence; Ind3: USC on network → RD → Anxiety → Alcohol dependence. All estimated coefficients are standardized. SE Standard Error, CI Confidence Interval, LLCI Lower limit of the 95% CI, and ULCI Upper limit of the 95% CI; and Number of bootstrap samples for percentile bootstrap confidence intervals is 5,000 Table 4Mediation effects of RD, AnxietyStandardEstimateBoot SEBoot LLCIBoot ULCITotal Ind0.050.020.020.10Ind10.0050.02-0.030.05Ind20.030.010.010.06Ind30.020.010.010.03Ind1: USC on network → RD → Alcohol dependence; Ind2: USC on network → Anxiety → Alcohol dependence; Ind3: USC on network → RD → Anxiety → Alcohol dependence. All estimated coefficients are standardized. SE Standard Error, CI Confidence Interval, LLCI Lower limit of the 95% CI, and ULCI Upper limit of the 95% CI; and Number of bootstrap samples for percentile bootstrap confidence intervals is 5,000 Mediation effects of RD, Anxiety Ind1: USC on network → RD → Alcohol dependence; Ind2: USC on network → Anxiety → Alcohol dependence; Ind3: USC on network → RD → Anxiety → Alcohol dependence. All estimated coefficients are standardized. SE Standard Error, CI Confidence Interval, LLCI Lower limit of the 95% CI, and ULCI Upper limit of the 95% CI; and Number of bootstrap samples for percentile bootstrap confidence intervals is 5,000 Discussion USC on network and alcohol dependence Research supports the contrast effect of social comparison, confirming that USC on network is a significant predictor of alcohol dependence. Hypothesis 1 was validated by our findings, consistent with previous findings [51]. For instance, Dana’s research indicated that college students who frequently compare themselves to others are more likely to be influenced by perceived drinking norms, which significantly predicts their alcohol consumption. College students, in the emerging stage of adulthood, undergo cognitive and social role changes, gradually gaining autonomy and experiencing fewer constraints on their daily behaviors. This period is marked by a high incidence of risky behaviors [52]. Social media provides ample material for social comparison, and those with a high tendency for social comparison are particularly interested in others’ behaviors and often feel uncertain about themselves [41]. Thus, when exposed to alcohol-related content online, their behavior is more strongly influenced by their reference group. This finding contributes to the advancement of social comparison theory. This study introduces the concept of social comparison into the realm of online networking environments. Given the widespread use of the Internet, particularly the popularity of short videos, college students are increasingly exposed to a plethora of positive information regarding others’ life situations, appearances, and abilities, prompting them to engage in USC on network more frequently [31, 53]. Therefore, it is important to further explore the specific effects of USC on individual psychology and behavior within the online environment. The present study revealed a positive association between USC on network and alcohol dependence among college students. While previous research has primarily focused on self-evaluation, self-esteem, life satisfaction, and compulsive buying behavior in relation to social comparison [54–56], little attention has been given to its impact on alcohol dependence. Therefore, this study broadens the application scope of social comparison theory. The mediating role of RD and anxiety This study indicated that the mediating impact of RD on the relationship between USC on network and alcohol dependence was not significant, contrary to hypothesis 2. Existing research has identified a significant relationship between RD and alcohol dependence, and suggested that negative affective evaluation fully mediates the relationship between RD and unsafe behavior intentions [57]. Thus, we speculate that RD may indirectly affect alcohol dependence through other factors. This might explain why hypothesis 2 was not supported. However, it was found that anxiety could act as a mediator in the association between USC on network and alcohol dependence, thus confirming hypothesis 3. Engaging in USC on social networks can make individuals feel inferior to others, triggering anxiety stemming from concerns about social belonging and validation [43]. The anxiety may then lead to increased alcohol consumption as a coping mechanism to alleviate feelings of stress [58]. Overall, the findings further elucidate the role of distinct psychological mechanisms in the association between USC on network and alcohol dependence. Specifically, it is revealed that anxiety has a more substantial mediating influence on the relationship between USC on network and alcohol dependence compared to RD. Additionally, these results offer a novel perspective on understanding the etiology of alcohol dependence. Previous research suggests that beyond biogenetic and environmental factors [59, 60], an individual’s psychological state, particularly anxiety, plays a pivotal role in promoting alcohol dependence. This discovery implies innovative strategies for preventing and intervening in alcohol dependence by providing psychological support and anxiety management skills to individuals prone to anxiety during social comparisons. The chain mediating role of RD and anxiety According to the study results, hypothesis 4 is confirmed, while hypothesis 2 is not. This indicates that RD and anxiety can mediate the relationship between USC on network and alcohol dependence. The findings also suggest that the emergence of RD further exacerbates anxiety levels among college students. This could be attributed to RD inducing negative cognitive patterns, which heighten anxiety experiences, consistent with previous research [37]. Specifically, college students often experience a sense of RD when comparing themselves to individuals with certain advantages online [61]. This feeling of RD can form negative thinking patterns and eventually increase levels of anxiety among college students [37, 62]. Therefore, college students may resort to maladaptive behaviors, such as developing alcohol dependence, as temporary means of alleviating their anxiety [63]. In summary, this study provides new insights into how social comparison processes influence alcohol dependence, specifically indicating that RD and anxiety mediate the relationship between online USC and alcohol dependence. Additionally, the findings support the theoretical framework of “stimulation-cognitive-emotion-behavior” [64]. According to the cognitive-evaluative theory of emotion, cognitive processes may precede emotion generation. Individuals first make a cognitive assessment of environmental stimuli and then develop an emotional response based on this evaluation [65]. Our results substantiate that RD is an antecedent variable for anxiety-related emotions. This aligns with the findings of Nadler [37], who found that RD can trap individuals in negative thought patterns and increase anxious emotions. Furthermore, while our study demonstrates a significant chain mediating effect between RD and anxiety, the sole mediating effect of RD was not significant. This suggests that USC on network does not influence alcohol dependence through RD alone but rather through the negative emotions generated by RD, which subsequently influence behavior. Main findings and limitations This study contributes to a better understanding of the association between USC on network and alcohol dependence, as well as the underlying mechanisms. First, this study evaluated the impact of the psychological process of USC on network on alcohol dependence. Therefore, this study enhances the existing research on the correlation between USC on network and alcohol dependence, and offers a novel theoretical framework for future research in this area. Other studies can use this framework to explore the mechanisms by which USC influences alcohol dependence. Second, the study results verified that RD and anxiety had a chain mediation impact between USC on network and alcohol dependence and clarified precedence relationship between RD and anxiety. Altogether, the findings of this study provide several practical implications for college students, suggesting that addressing and alleviating negative emotions can help reduce alcohol dependence. Nevertheless, it is important to acknowledge the limitations of this study. Firstly, given the cross-sectional design of this study, a causal relationship among the variables was not able to be established. Future research based on the longitudinal design is needed to better understand the relationship among these variables. Second, this study solely focused on college students in eastern China as the research subjects. Further investigations should be conducted on individuals with a wider age span and regional scope to validate our results. Third, data were collected through questionnaire survey measures, and thus the results might suffer from the influences of common method variance and inaccurate reporting. In the future, other research methods, such as task measurement, should be employed to obtain more accurate and objective research results. Fourth, we did not focus on the development process of alcohol dependence. There is a transition period from alcohol use to alcohol dependence, and understanding the factors at play during this period is worth further exploration. Conclusion This study reveals that USC on network positively predicts alcohol dependence via the mediating effect of anxiety, and that RD and anxiety play a chain-mediating role between USC on network and alcohol dependence. These findings suggest the importance of interventions that address both RD and anxiety to prevent alcohol dependence among college students who experience frequent USC.