The relationships between short video addiction, subjective well-being, social support, personality, and core self-evaluation: a latent profile analysis

Abstract

Background

The proportion of short video addiction (SVA) is increasing among different populations, and its impact on the adolescent group has attracted research attention in recent years. To better explore the heterogeneity of individuals addicted to short videos and the antecedents of their addiction, the present study used quantitative research to conduct a latent profile analysis (LPA) of college students’ SVA from the perspective of internal and external resources.

Methods

Referring to the Interaction of Person-Affect-Cognition-Execution (I-PACE) theory, this study explores the relationship between subgroups of SVA on the one hand and subjective well-being (SWB), social support (SS), core self-evaluation (CSE), and extraversion on the other using latent profile analysis. The research participants consisted of 694 college students aged between 18–25 years. The research tools used were the SVA Scale for College Students, Subjective Well-Being Scale, Perceived Social Support Scale, Core Self-Evaluation Scale, and the Big Five Personality Traits Scale.

Results

The results revealed three subtypes of adolescent SVA: high (28.8%), medium (56.6%), and low (14.6%), demonstrating significant heterogeneity among groups. These groups also showed significant differences in terms of their internal and external resource scores. Specifically, the high-addiction group scored significantly lower on extraversion and SS than the other two groups. The medium-addiction group had the lowest score on CSE among the three groups. The low-addiction group scored significantly higher on SS, CSE, and SWB compared with the high-addiction and medium-addiction groups.

Conclusion

This study reveals that students with low extraversion, low SS and CSE scores tend to become addicted to short videos. SVA has negatively affected these personal traits. Our study explored the heterogeneity of SVA among college students from an individual-centered perspective, providing empirical support for scientifically and reasonably guiding the use of short videos among college students.

Introduction

The vigorous development of new technologies, such as artificial intelligence (AI) and 5G, has catalyzed the transformation of media forms [1]. Short videos, typically disseminated through mobile smart devices and reliant on mobile social platforms, have emerged in recent years, breaking through temporal, spatial, and audience constraints and rapidly becoming a focal point due to their ability to provide unique user experiences [2]. Among numerous communication channels, short video platforms have become the preferred choice for the majority of Internet users [3]. Since June 2023, the number of short video users in China has reached 1.026 billion, marking an increase of 14.54 million compared to December 2022 and accounting for 95.2% of the overall Internet population (CNNI 52nd China Internet Development Status Report, 2023). TikTok’s global user base is projected to reach 1.677 billion by the end of 2023 with 3.5 billion downloads, marking a 79.1% increase since 2020 [4]. Meanwhile, Twitter has 372.9 million users, YouTube has 2.527 billion, and Instagram has 1.628 billion [4].

At the same time, the popularity of short videos has led to reports of excessive use among college students. This overuse has disrupted their daily lives, including but not limited to lower emotional management, reduced efficiency in studying or working, elevated levels of anxiety, and depression [5]. As one of the groups considered as frequent users of short videos, college students are more prone to excessive use of short videos due to their weaker self-control and discernment abilities, coupled with external influences from their social environment, thereby causing video dependence or even addiction [6]. An increasing number of studies have proven that the ease of use and entertainment characteristics of short videos strengthen people’s tendency for excessive use, making it difficult to restrain their impulses [7]. The excessive use of short videos may also lead to addictive behaviors. However, compared to other forms of behavioral addiction (e.g., gaming addiction or Internet addiction), short videos require longer durations and sustained engagement to drive addictive behavior, leading to fundamental differences between them [7]. Given that research on SVA is still in its early stages both domestically and internationally, investigating the influencing factors of SVA and the relationships between these factors is crucial [8].

As early as 1998, researchers have already proposed seven symptoms for diagnosing Internet addiction based on the substance addiction criteria in DSM-IV. These include loss of control, tolerance, withdrawal symptoms, excessive frequency and intensity of Internet use, unsuccessful attempts to reduce Internet time, excessive time spent on Internet activities, and severe impairment of daily life [9]. However, opinions vary among scholars regarding the reasons for SVA. Thus far, few studies have focused on the uniqueness of SVA behavior, mostly borrowing from similar concepts in adjacent fields, such as social media addiction [10], smartphone addiction [11], and online game addiction [12]. While some studies have addressed the ‘obsession’ or ‘problematic use’ of short videos, they have not provided clear diagnostic criteria for what constitutes addiction.

Based on a research by Young and the compilation of a smartphone addiction scale by Liang Yongchi, scholars have proposed four dimensions of SVA: withdrawal, escapism, inefficacy, and inefficiency [13]. Referring to these dimensions, some scholars have defined “SVA” as a new form of Internet addiction, characterized by a chronic or cyclical state of obsession caused by the repeated use of short video apps (e.g., TikTok), resulting in strong and persistent cravings and psychological dependence [14]. However, considering individual distinctions, scholars have long reported characteristic behavioral differences among groups with varying degrees of social media addiction [15]. Therefore, to better explore the heterogeneity of individuals addicted to short videos and the antecedents of their addiction, the present study conducted a latent profile analysis (LPA) of college students’ SVA from the perspective of internal and external resources.

Theoretical background

Addiction to short video and subjective well-being

Subjective well-being (SWB) is an important comprehensive psychological indicator used by evaluators to assess individuals’ overall quality of life (QoL) according to their own standards [16]. In 1960, Wilson elevated the study of SWB to theoretical heights. Referring to his work, “Correlates of Avowed Happiness” (1967) as a reference point, some scholars roughly divided research on SWB into two stages: before 1967, which is dominated by qualitative research, and after 1967, which is characterized by a combination of qualitative and quantitative research [17]. Since then, research on well-being has reached a relatively mature level.

Among numerous studies, the influencing factors of SWB can mainly be divided into basic human needs (e.g., external wealth and health) and psychological needs (e.g., internal emotional needs and respect from others). These two types of needs are mutually independent, and their fulfillment can enhance individuals’ SWB. However, existing research indicates that the impact of external factors on SWB is often limited. Demographic factors, such as income, marriage, and education level, only account for a small part of the variance in happiness measurement [18].

The development of media technology has made people increasingly reliant on various forms of media for their understanding of the external world. However, as the media becomes the people’s primary or even sole source of information, the portrayal of social reality in various media content can strongly shape or even distort people’s perceptions of reality. By understanding the reality presented by the media, certain beliefs and attitudes of individuals may change [19], thereby altering the standards by which individuals evaluate their quality of life and ultimately influencing their SWB. Some studies suggest that the use of online social media can strengthen connections with others and provide social support (SS), positively impacting SWB [20]. However, as a form of media, short videos can shape the SWB of college students, potentially causing cognitive dissonance and leading to maladaptive use of short videos. Research also shows that the improper use of short videos and other social media can have negative effects on people’s physical and psychological well-being, thereby affecting their SWB [21]. For example, individuals with SVA are more likely to experience anxiety [22], depression [23], and other psychological issues. Thus, there exists a negative correlation between levels of SWB and SVA.

Subjective well-being, social support, and addiction to short videos

Social Support (SS) is the perception of the kind of support individuals believe they can receive or are receiving from external sources [24]. Previous research on SS and SWB has mostly explored and analyzed the relationship between the sources, characteristics, and nature of SS [25]. According to different reference standards, SS can be classified into four categories: material, emotional, esteem, and companionship support [26]. Furthermore, different types of social relationships provide varying kinds of SS [27]. Perceiving SS, which meets the individual’s needs, is an important coping resource for the psychological health of college students [28].

Existing research indicates a significant correlation between SS and various indicators of SWB. For example, through a decade-long longitudinal study, scholars found that SS could still predict SWB even after controlling for income [29]. Different aspects of SWB can be predicted by various aspects of SS. For example, perceived support significantly predicts life satisfaction and negative emotions [30]. Other studies also reported a positive correlation between SS and SWB among college students [31]. Such a positive correlation has also been reflected in other studies [32], thereby demonstrating a strong correlation between SS received and SWB. SS is considered an important predictor of SWB [33], and some scholars have proposed testing the relationship path model between interpersonal relationships (self-centered social networks and perceived SS, PSS) and SWB [34].

In recent years, with the rapid growth of the Internet and social media users, online channels have become increasingly important platforms through which people can give and receive SS [35]. Meanwhile, as short videos emerge as the fastest-growing social network in the post-pandemic era, they have already become an important channel for college students’ social interaction, serving as one of their significant sources of SS. However, research indicates that although college students can obtain some form of online SS, Internet addicts cannot fundamentally reduce their sense of isolation caused by the lack of real-life SS through the support obtained from the Internet [36]. Moreover, in accordance with the cognitive-behavioral model, SS is considered an important proximal condition affecting Internet addiction [37].

Subjective well-being, personality, core self-evaluation, and addiction to short videos

Even after considering external circumstances, there is still considerable individual variability in SWB, a variation that can be explained by personality traits [18]. Researchers often directly study the relationship between personality traits and SWB, and these studies consistently demonstrate that personality is one of the strongest and most stable predictors of SWB [38]. Among numerous personality traits, neuroticism and extraversion show the strongest correlation with happiness. For example, empirical studies have reported the significant impact of extraversion on the SWB of college students (Xue Zheng et al., 2003). Scholars believe that extraverts exhibit different behaviors compared to introverts, and such behavioral differences lead to positive happiness outcomes [18]. Furthermore, individuals with extraverted personalities often have stronger sociability, implying better interpersonal harmony and, therefore, higher levels of happiness [39].

Thus far, conflicting results have been obtained regarding the relationship between extraversion and the Internet or social media addiction. For example, Kuss argues that individuals with low extraversion often encounter obstacles in real-world social activities, and in the online world, their social activities are more relaxed and enjoyable, naturally leading them to avoid real-world interactions. However, some scholars have found that individuals with high extraversion are more prone to addiction. For example, Gosling reported that high extraversion not only predicts the frequency of college students’ use of Facebook but also predicts the extent of their addiction to this social media platform. In Wang’s survey, the regression coefficient of extraversion on Internet social addiction also reached 0.10. Regardless of which viewpoint, this indicates a strong correlation between extraversion and addiction.

As an important indicator of personality variables, core self-evaluation (CSE) is the basic evaluation individuals hold of their own abilities and worth [40]. The four basic traits included in CSE are self-esteem, general self-efficacy, neuroticism (emotional stability), and locus of control. In accordance with the motivational mechanism of CSE, individuals with positive self-concepts tend to pursue approach goals, while those with negative self-concepts tend to pursue avoidance or protective goals. Moreover, the most recent studies also showed that CSE has a positive predictive effect on SWB [41]. Furthermore, individuals with high CSE have positive self-concepts, thereby generating more positive approach goals and higher motivation for approach success, which is an important mechanism for people with high CSE to experience happiness [42]. Personality traits have been identified as risk factors for various behavioral addictions [43] and are often studied as mediating or moderating variables for related addictions, such as smartphone addiction and social media addiction [22]. General self-efficacy in CSE is also considered an important factor in addiction; in fact, researchers have found that it is related to the motivation to initiate or resist addictive behavior [44]. On the one hand, social media addiction often accompanies setbacks, such as declining academic performance and hindered social interactions, which can lead college students to self-denial and subsequently reduce their CSE [45]. On the other hand, information from social networking sites can trigger college students to engage in upward social comparison, which can negatively affect the CSE of addicted individuals [46]. The compulsiveness caused by addiction makes individuals perceive a decrease in self-control and damage to CSE [47]. At the same time, individuals with low self-evaluation may perceive social media and other online platforms as safer places to express themselves than those with high self-evaluation (Forest & Wood, 2012), thus deepening the level of addiction.

Theoretical support

This study aims to explain addictive behaviors related to short videos based on the Interaction of Person-Affect-Cognition-Execution (I-PACE) model. This model assumes that the development of addictive behavior results from the interactions between predisposing variables, emotional and cognitive responses to specific stimuli, and executive functions [48]. Research based on satisfaction theory suggests that one of the reasons for social media addiction is the inadequate satisfaction of the psychological need for belonging [49]. Individuals with high belonging needs but lacking a sense of satisfaction may eventually overuse social media, which can also apply to short videos. Lower levels of life satisfaction in the cognitive dimension are associated with lower SWB, leading to negative emotions and discomfort. As argued by cognitive dissonance theory, individuals typically adopt either changing or rationalizing behaviors to reduce this discomfort. However, both methods can considerably influence an individual’s cognition. Moreover, individuals may even process information in a biased manner during the rationalization of their behaviors, further leading to irrational behavior.

Based on the updated I-PACE model [50], the addiction process can be divided into early and late stages, and an individual’s core characteristics (e.g., degree of extraversion in personality) may serve as predisposing variables for addiction. These specific predisposing variables are considered features of different specific addictive behaviors. In the early stages of addiction, individuals may experience problems with the problematic use of short videos, which may jeopardize significant interpersonal relationships (e.g., with friends and/or family), leading to reduced SS. At the same time, investing a significant amount of time in short videos can lead to a decrease in individual work performance and academic achievement, which in turn, can directly affect individuals’ CSE. This discomfort may lead individuals to rationalize their addictive behavior by emphasizing the benefits of short videos, leading to maladaptive cognition. This maladaptive cognition makes individuals more anxious, irritable, and uneasy in their interactions with real-life situations, thereby suppressing SWB. Consequently, the decrease in SWB leads to negative emotional and cognitive responses in individuals, thereby increasing their attention and impulsivity towards these stimuli, specifically manifested in addiction to short videos. Emotional and cognitive responses can lead individuals to make decisions in specific ways, guided by two interacting systems: the impulsive response system and the reflective deliberation system. Individuals with addictive tendencies are believed to increasingly rely on the impulsive response neural system, with decreased brain control over impulses and desires during the addiction process, which ultimately leads to SVA.

Method

Questionnaire and sample collection

This study adopted random sampling to select college students for sampling surveys. A total of 750 questionnaires were distributed randomly. The distribution of questionnaire issuance was conducted from August 29, 2023, to September 29, 2023. Students participated in an online survey through a professional electronic questionnaire website (https://www.wjx.cn/). In addition, all participants were informed that the questionnaire was anonymous and that their online participation was entirely voluntary. After deleting samples that took too short to answer and samples with exactly the same answer, ultimately, 694 valid questionnaires were collected from the total 750 questionnaires, yielding an effectivity rate of 92.53%.

In terms of demographic characteristics, there were 311 male participants and 383 female participants. The age range of the students was between 18 and 25 years old, with 463 individuals aged 18–20 years, accounting for 66.70% of the total sample. Regarding the duration of short video usage, the majority of participants reported an average usage time of 1–3 h. Please refer to Table 1 for detailed data.

Table 1.

Descriptive statistics of demographic variables

Variables	Category	Numbers	Percentage
Age	18–20	463	66.70%
	20–25	211	30.40%
	Over 25	20	2.90%
Gender	Male	311	44.80%
	Female	383	55.20%
Average usage duration	1-2H	251	36.17%
	2-3H	252	36.31%
	3-5H	169	24.35%
	Over 5H	22	3.17%

H Hour

Research tools

Variables

Short Video addiction (SVA)

This study used the Short Video Addiction Scale (SVAS) for College Students developed by Qin Haoxuan [13]. Originally adapted from Liang Yongchi’s Mobile Phone Addiction Inventory (MPAI), this scale also incorporates the diagnostic criteria for Internet addiction developed by Young [51] based on pathological gambling standards. These criteria have been widely used by scholars worldwide to measure the levels of Internet addiction among college students and have undergone reliability and validity testing. The formal scale consists of 14 items and organized into four dimensions (withdrawal, escapism, loss of control, and inefficiency). Higher scores indicate more self-reported symptoms of short video addiction. In our study, the Cronbach's α coefficients for the total scale and sub-scales were 0.82, 0.79, 0.76, 0.68, and 0.69, indicating a high level of internal consistency and reliability.

Subjective Well-Being (SWB)

Subjective Well-Being refers to individuals’ overall emotional and cognitive evaluations of their QoL, serving as an important comprehensive psychological indicator of personal life quality. This scale was proposed by Diener [16] and later revised into a Chinese version by Xing Zhanjun [52]. The scale consists of 20 items rated on a 6-point Likert scale (1 = strongly disagree, 6 = strongly agree). In our study, the Cronbach's α coefficient for the SWBS was found to be 0.82 for the total scale, indicating a high level of internal consistency and reliability.

Social Support (SS)

“Social Support” is a concept that is closely related to feelings of loneliness and mental health within social networks. Sarason [53] defines SS as individuals’ perceptions of the support they can receive or expect from the external environment. Perceived SS, as measured by the Perceived Social Support Scale (PSSS) developed by Blumenthal [54], emphasizes individuals’ self-understanding and feelings regarding the support obtained from various aspects of society. The scale consists of 12 items rated on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree) and organized into three dimensions (family support, friends support, and significant others support). Higher scores indicate more self-reported well-being. Due to the different original sample group (patients with coronary artery disease), comfirmatory factor analysis was conducted, and the fitting indexes were as follows:χ² = 158.12, DF = 51, RMSEA = 0.06, RMR = 0.07, GFI = 0.94, AGFI = 0.94, CFI = 0.98, IFI = 0.98. In our study, the Cronbach's α coefficients for the total scale and sub-scales were 0.95, 0.89, 0.89, and 0.87, indicating a high level of internal consistency and reliability.

Core Self-Evaluation (CSE)

Core Self-Evaluation is the fundamental evaluation individuals make of their own abilities and worth, playing a significant predictive and moderating role in college students’ mental health. The Core Self-evaluation Scale (CSES) was developed by Judge [40] and has been translated and revised by Dai Xiaoyang [55] for the Asian cultural context and comprises 10 items rated on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). Higher scores indicate more self-reported core self-evaluation. In our study, the Cronbach's α coefficients for the one-dimension scale were 0.77, indicating a high level of internal consistency and reliability.

Personality

“Personality” consists of observable behavioral traits and internal, unobservable characteristics [55]. The Chinese version of the Ten-Item Personality Inventory (TIPI-C) [56], was based on the original TIPI scale developed by Samuel D. Gosling [57]. We adopted two items rated on a 7-point Likert (1 = strongly disagree, 7 = strongly agree), which involved the extroversion dimension of the personality. Higher scores indicate more self-reported extroversion. The original sample group was consistent with our sample group. In our study, the Cronbach's α coefficient for the for the extraversion sub-scale was found to be 0.72, indicating a high level of internal consistency and reliability.

Data analysis tools

This study used MPLUS 8.3 for the Latent Profile Analysis(LPA) and SPSS 27.0 for descriptive statistics of demographic variables, analysis of variance, and logistic regression analysis. First, a correlation analysis was conducted using SPSS software to examine the means, standard deviations, and Pearson correlation coefficients of the variables involved. Then, LPA was performed to further explore the antecedents of SVA. Four variables influencing SVA (SWB, CSE, SS, and extraversion of personality) were used as manifest indicators. The LPA models were fitted by sequentially increasing the number of latent classes affecting SVA, and the participant classes were determined based on the analysis data. Finally, based on the LPA results, the differences in the main variables among the different types were analyzed.

Results

Correlation analysis of the variables

The means, standard deviations, and correlation coefficients of the main variables in this study are presented in Table 2. Among these, SVA, CSE, and SWB are correlated with each other in a pairwise manner, except for extraversion and SS.

Table 2.

The means, standard deviations, and correlation coefficients of the main variables

	M	SD	1	2	3	4	5	6	7	8
1 age	1.36	0.54	1
2 gender	1.55	0.50	−0.07	1
3 average usage duration	1.95	0.85	0.03	−0.04	1
4 extroversion	7.85	3.02	−0.06	0.01	−0.01	1
5 short video addiction	3.46	1.99	0.04	−0.15**	0.27***	0.01	1
6 Social Support	57.93	17.05	−0.04	0.10**	−0.06	0.44***	−0.16***	1
7 Core Self-Evaluation	32.68	6.07	−0.03	−0.04	−0.08*	0.12**	−0.19***	0.27***	1
8 Subjective Well-Being	3.89	0.67	0.01	−0.01	−0.09*	0.19***	−0.16***	0.41***	0.66***	1

M Mean, SD Standard Deviation

*P<0.05 **P<0.01 ***P<0.001

Latent profile analysis

The LPA study used SWB, CSE, SS, and extraversion as the four dimensions to examine model fit using information criteria AIC, BIC, and sample-adjusted BIC, along with the entropy index. Likelihood ratio test (LMR-LRT) and bootstrap likelihood ratio test (BLRT) were employed as model fit indices. The lower values of the three information criteria indicate better model fit, while the entropy index ranges between 0 and 1, in which values closer to 1 indicate more accurate latent classification. An entropy value less than 0.60 implies a classification error rate exceeding 20%, while an entropy value of 0.80 indicates a classification accuracy rate exceeding 90% [58]. LMR and BLRT are primarily used for model comparison; if the p-values of these two indices reach significance, this indicates that an n-category model is superior to an n-1 category model [59].

The data results shown in Table 3 indicate that during the testing process from 2 to 6 categories, in comparison with the information criteria, the 4-category model outperforms the 3-category model, and the 5-category model outperforms the 4-category model. However, in the test between the 5- and 4-category models, the LMRT value is not significant, indicating that the 5-category model is not superior to the 3-category model. From the perspective of the entropy index, the highest value is observed for the 3-category model, indicating the most accurate classification. Additionally, both the LMRT and BLRT values for the 3-category model are significant. Considering all factors, the 3-category model is thus selected as the optimal model.

Table 3.

Fit indices for various profile categories

Class	AIC	BIC	aBIC	Entropy	p-LMRT	p-BLRT	Proportion
1	9862.430	9907.855	9876.103	—	—	—	—
2	9428.712	9501.391	9450.588	0.789	0.0000	0.0000	0.702/0.298
3	9172.687	9272.621	9202.767	0.804	0.0000	0.0000	0.146/0.566/0.288
4	9100.169	9227.358	9227.358	0.769	0.0691	0.0000	0.118/0.267/0.258/0.357
5	9044.144	9198.588	9090.632	0.740	0.2132	0.0000	0.261/0.118/0.249/0.269/0.102
6	8998.126	9179.825	9052.818	0.784	0.0545	0.0000	0.020/0.117/0.267/0.068/0.278/0.251

The selected models AIC, BIC and ABIC are the information evaluation indexes; p−LMR and p−BLRT are the model fit test indexes

To convert the scores of the scales into standardized scores (z-scores) and analyze the scores of different categories of individuals on each indicator’s z-scores based on their performance on the four influencing factors in the 3-category model (Fig. 1), it was found that the mean differences in the scores of the three latent categories of students on SVA were significant, thus indicating different characteristics. This finding suggests heterogeneity among college students in terms of SVA. The first category of individuals had the highest SVA index, while the scores on the other four manifest variables were relatively low. These students exhibited high characteristics of SVA and were therefore named “high-addiction type,” accounting for 14.6% of the total sample. The second category of individuals had a medium SVA index, and among the other manifest variables, the self-evaluation score was the lowest among the three categories of individuals. They were thus named “medium addiction type,” comprising the largest proportion of the sample at 56.6%. The final category of individuals had the lowest SVA index, while the scores on the other four manifest variables were relatively high. They were named “low-addiction type,” comprising 28.8% of the total sample.

Fig. 1.

The response characteristics of different latent categories in each item

ANOVA results showed significant differences among people with different degrees of SVA in terms of extraversion, SWB, CSE, and SS. Posthoc tests (Table 4) revealed that the high-addiction group scored significantly lower on extraversion and SS than the other two types, while the differences in SWB and CSE scores (Fig. 1) between the high- and medium-addiction groups were not significant. The medium-addiction group scored significantly lower than the low-addiction group in SS, SWB, and CSE, with no significant difference in extraversion scores. Furthermore, the low-addiction group scored significantly higher than the other two groups in SS, SWB, and CSE, but showed no significant difference in extraversion than the medium-addiction group.

Table 4.

Differences in various dimensions among various profile categories

	1	2	3	F	η2	Post-hoc test
1 Extroversion	4.86 ± 2.05	8.36 ± 2.90	8.68 ± 2.71	75.73***	0.18	1 < 2 = 3
2 Social support	26.15 ± 8.19	60.02 ± 10.88	69.87 ± 8.99	660.58***	0.66	1 < 2 < 3
3 Core Self-Evaluation	30.91 ± 3.98	29.60 ± 4.19	39.62 ± 4.14	398.05***	0.54	1 = 2 < 3
4 Subjective Well-Being	3.55 ± 0.59	3.60 ± 0.46	4.68 ± 0.55	330.03***	0.49	1 = 2 < 3
5 Short video addiction	4.36 ± 1.57	3.51 ± 1.99	2.93 ± 2.05	18.31***	0.05	1 > 2 > 3

*P<0.05 **P<0.01 ***P<0.001

The relationship of the different profile categories to the internal and external resources

To further explore the heterogeneities of different categories of SVA in various indicators, one-way analysis of variance (ANOVA) was conducted. The results showed significant differences among different categories of SVA in terms of SS and SVA. The results showed that there were significant differences between Category 1 and Category 2 in terms of extraversion and between Category 2 and Category 3 in terms of CSE and SWB, which were statistically significant (*** = P < 0.001). Posthoc tests (Table 4) revealed that the high-addiction type had the lowest scores in the extroversion and social support dimensions, while the medium addiction type had the lowest core self-evaluation scores among the three categories. The low-addiction type had the highest scores in short video addiction and the lowest scores in extroversion social support.

Discussion

Latent categories and characteristics of SVA

The study found significant grouping characteristics in college students’ addiction to short videos. By comparing the LPA indicators (AIC, BIC, and sample-corrected BIC, and entropy index) and model fit indices (LMR-LRT and BLRT), a 3-category model was identified as the best model. This aligns with a previous research categorizing college students’ mobile phone addiction into high-, low-, and no-addiction groups [60]. Based on the research results, the three different latent subtypes were named as high-addiction type (14.6%), medium addiction type (56.6%), and low-addiction type (28.8%). ANOVA results showed significant differences among people with different degrees of SVA in terms of extraversion, SWB, CSE, and SS.

Based on the Posthoc tests (Table 4) results, the high-addiction subtype had the highest score in SVA, indicating a deep level of addiction. Their levels of extraversion, CSE, perceived SS, and SWB were the lowest among the three types. These individuals tend to be introverted, resulting in difficulties in maintaining harmonious interpersonal relationships with others, receiving less SS, developing the tendency to escape from reality, and having low life satisfaction. The medium-addiction subtype scored at a medium level on SVA, with the lowest CSE scores among the three groups and accounting for the largest proportion of the sample. This suggests that over half of the individuals are in a state of medium addiction to short videos, with low basic evaluations of their abilities and values. Influenced by reality, they hold negative self-concepts and tend to pursue relaxed and enjoyable activities, but are nevertheless constrained by real-world interactions, preventing the further deepening of their addiction. The low-addiction subtype had a shallow level of SVA, with the highest scores in extraversion, CSE, perceived SS, and SWB among the three types. This finding indicates that individuals with higher quality of life, more love and support, and positive self-concepts generally have low levels of SVA.

Relationship between Latent Categories of SVA and Intrinsic and Extrinsic Resources (Social Well-Being, Social Support, Personality, and Core Self-Evaluation)

The data analysis results indicate that there are significant differences in SWB, extraversion, CSE, and SS among individuals in different latent categories of SVA. Based on the LPA and ANOVA results, individuals in the high-addiction category scored the lowest on extraversion, with no significant differences between the medium- and low-addiction categories. Even after considering external resources, there remains considerable individual differences in SWB, which can be explained by personality traits [18]. Personality is one of the strongest and most stable predictors of SWB [38], especially extraversion, which has a strong correlation with well-being [18]. The variance analysis in the present study showed that the high-addiction group scored the lowest on extraversion, indicating that individuals with low extraversion are more likely to be addicted to short videos. This finding is consistent with some previous studies, as individuals with low extraversion often face social obstacles in real life, leading to lower levels of happiness due to the lack of harmonious interpersonal relationships. Therefore, they tend to have more relaxed social interactions in the virtual world, leading to increased frequency of social network use, prolonged exposure to short videos, and ultimately, addiction. Therefore, extraversion is considered a negative predictor of SVA, consistent with the results of the data analysis in this study.

On CSE, the medium-addiction group scored the lowest, with no significant difference from the high-addiction group. Individuals with low CSE usually have negative self-concepts, tend to pursue avoidance or defensive goals, and engage in more relaxed and enjoyable activities. The decrease in CSE can be attributed to setbacks, such as declining academic performance and interpersonal obstacles, which naturally lead to thoughts of escapism and excessive use of short videos. In addition, addictive behaviors are typically rationalized under the influence of cognitive dissonance, leading to maladaptive cognition that, in turn, inhibits SWB. Under the influence of impulsive response neural systems, individuals’ control over impulses and desires decreases, ultimately leading to SVA.

When SVA is low, SWB scores are high, whereas when addiction is high, SWB scores are low (Fig. 1). In accordance with the satisfaction theory [49] and the I-PACE model [48], a decrease in life satisfaction leads to a decrease in SWB, resulting in negative emotions and feelings that reduce individual satisfaction and increase attention and impulsivity toward short videos. Based on cognitive dissonance theory, individuals tend to rationalize their behavior to reduce negative emotions and feelings. Thus, for individuals with addictive tendencies, decision-making and behavior depend on impulsive response neural systems, leading to decreased control over impulses and desires, ultimately resulting in SVA [23].

Regarding SS, the higher the addiction level, the lower the score for SS. Studies have shown that SS predicts SWB [29]. Furthermore, a decrease in SS predicts the emergence of negative emotions and a decrease in life satisfaction [31], thus indirectly predicting the occurrence of SVA. Based on the updated I-PACE model [50], individuals jeopardize important interpersonal relationships in the early stages of addiction due to problematic use of short videos. In turn, this leads to a decrease in SS, resulting in negative emotions and a decrease in SWB. Therefore, a decrease in SS is considered one of the reasons for SVA.

Based on the above reasoning, this study believes that SVA is the result of the combined effects of intrinsic and extrinsic resources (SWB, SS, extraversion, and CSE). Therefore, based on previous research and data analysis results, extraversion, CSE, SWB, and SS are considered negative predictors of SVA, consistent with previous studies on factors related to SVA, and with the results of the data analysis in the present study.

Negative effects of SVA on college students at different stages

In the initial stage of addiction, corresponding to the low-addiction category in this study, both intrinsic and extrinsic resource indicators have relatively high scores, with the lowest score observed in extraversion. Personality, as a core characteristic, may serve as a triggering variable for addiction. Therefore, individuals in this category have potential risks of SVA. In the middle stage of addiction, corresponding to the medium-addiction category in this study, individuals in this category have the lowest scores in CSE. This finding may be due to setbacks encountered in real life leading to self-doubt in individuals, and prolonged exposure to such situations can have a negative impact on the CSE of addicted individuals [46], which could lead to a decrease in perceived self-control and, ultimately, addiction. When individuals are in a state of high-addiction, as observed in the high-addiction category in this study, both intrinsic and extrinsic resource scores are relatively low, with the lowest score observed in SS. This may be due to the fact that during addiction, individuals become engrossed in short videos, neglecting the external world, such as their academic or work responsibilities and interpersonal relationships, resulting in a decrease in SS received by this individual.

Conclusion

The present study used LPA to explore SVA among college students using five indicators: intensity of usage, SWB, extraverted personality, CSE, and SS.The study investigated meaningful specific differences among low, medium, and high levels of SVA types, identified three latent subgroups with different combinations of indicators (i.e., SWB, SS, CSE, and personality), and examined the influencing factors of different subgroups. The study proposed that SWB, extraverted personality, CSE, and SS are important predictive variables of SVA. Furthermore, the study extended the application of the I-PACE model in the empirical literature in examining the process of SVA and elucidated the reasons for SVA based on cognitive dissonance theory. The findings of the study help provide a theoretically sound model and explore the possibility of identifying potential addictive individuals. Specifically, college students with low SWB, extraverted personality, CSE, and SS are more likely to engage in high-risk SVA behaviors to balance their cognition, satisfy themselves through the use of short videos, and compensate for their deficiencies. These findings are theoretically reasonable and meaningful for the field.

To date, there have been relatively few studies employing LPA to analyze SVA. Previous research has primarily focused on understanding the psychological mechanisms through which SVA affects SWB, namely, how individuals’ more frequent and intense engagement in short video usage ultimately increases or decreases their feelings of happiness and satisfaction, as well as the mediating or moderating effects of various variables. Thus far, much of the analysis has been conducted from a variable-centered perspective, without considering the heterogeneity of the data, specifically the differences among latent subgroups within the data.

Future scope

This study explored the heterogeneity of SVA among college students from an individual-centered perspective, providing empirical support for scientifically and reasonably guiding the use of short videos among college students. Future research could consider conducting longitudinal surveys on the same group of individuals and incorporate more boundary conditions into the models. Furthermore, considering the influence of different social environments on college students, it would be beneficial to analyze the demographic heterogeneity of SVA, such as gender and nationality.

Furthermore, given that this study found that college students with medium addiction receive the lowest level of SS, efforts should be made to help them improve their social relationships. In other words, college students may need more care and support from family and friends in their real lives. Through communication and interaction, efforts can be made to understand their inner feelings, provide psychological support, and help them better cope with the anxiety and pressure brought by society. Finally, developing reasonable plans for short video usage, enhancing self-management, and avoiding excessive and indiscriminate use can also be emphasized as part of these efforts. We also suggest clinicians tailor specific interventions according to the different subtypes proposed in this study. As for the high addiction group (the lowest extraversion scores), social skills training and increased social engagement are recommended to bolster their extraversion. For the medium addiction group (lower social support, subjective well-being, and core self-evaluation), family, school, and community support are vital to enhancing their social support and perceived well-being. Cognitive behavioral therapy is also recommended to aid college students in improving their understanding of short videos.

Authors’ contributions

Conceptualization, J.D.; methodology, J.D.; investigation, Y.Z. and Z.H; resources, J.D., Y.Z. and Z.H.; data curation, Y.Z., Z.H. ,Y.X; writing-original draft preparation, Z.H. ,Y.Z.and J.D.; writing-review and editing, J.D.; supervision, J.D.; project administration, J.D.; funding acquisition, J.D. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by “the Fundamental Research Funds for the Central Universities (HUST)”, Grant Number 2023WKYXZX006.

Data availability

The data provided in this study are available upon request from the corresponding authors. Data in the administrative Excel file are not publicly available because the information may compromise the privacy of study participants.

Declarations

Ethics approval and consent to participate

This study conforms to the principles outlined in the Declaration of Helsinki and has been approved by the Ethics Committee of the School of Education of Huazhong University of Science and Technology (No. 20230802). Informed consent was obtained from all the participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.