https://orcid.org/0000-0002-8542-7614
Figures
Abstract
Background
Most studies on social media usage and parasocial relationships (PSRs) have been conducted in WEIRD (Western, Educated, Industrialized, Rich, and Democratic) societies, potentially overlooking the unique cultural, social, and economic factors present in non-WEIRD contexts. Examining these phenomena in a non-WEIRD setting is essential for a comprehensive understanding of social media’s global impact.
Methods
Secondary data from 574 participants in Qatar who followed Instagram influencers were analyzed using Bayesian analyses aided by Markov Chain Monte Carlo (MCMC) algorithms to examine the relationships between social media usage time, PSRs, and demographic factors.
Findings
The analysis results show that, regarding linear effects, a stronger parasocial relationship with Instagram influencer(s) is associated with higher daily social media usage time. Meanwhile, being male, being older, and having higher incomes all have negative associations with daily social media usage time. When parasocial relationships and the three demographic factors are seen in their interactions, negative associations with social media usage were also found in a similar pattern. To elaborate, among those with high parasocial relationship degrees, females, young people, and poor people tend to use social media for more hours each day.
Conclusions
This study highlights that demographic factors such as gender, age, and income in their interactions with parasocial relationships are associated with social media usage time within the non-WEIRD social context of Qatar. The findings underscore the necessity of considering the specific local cultural settings when studying social media behaviors.
Citation: Jin R, Le T-T (2025) Attachment beyond the screen: The influences of demographic factors and parasocial relationships on social media use in Qatar. PLoS One 20(6): e0326685. https://doi.org/10.1371/journal.pone.0326685
Editor: Andrea Fronzetti Colladon, Roma Tre University: Universita degli Studi Roma Tre, ITALY
Received: September 16, 2024; Accepted: June 3, 2025; Published: June 20, 2025
Copyright: © 2025 Jin, Le. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data is available online at: https://osf.io/jp9e5/?view_only=e4abb1409801403c81cad3c4f7dbbf40.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Social media platforms have become a cornerstone of modern communication, enabling users to connect, share, and consume content with others on a global scale. In 2023, it was estimated that there were over 4.9 billion social media users around the globe, which accounted for approximately 61.2% of the global population [1]. Among these social media platforms, Facebook leads the popularity globally with around 2.9 billion monthly active users, followed by YouTube and WhatsApp, each with over 2 billion users [2]. Famous content creators and influencers such as Kim Kardashian, PewDiePie, and Charli D’Amelio have hundreds of millions of followers, exerting far-reaching impacts in consuming, gaming, pop cultures, or even political views among the younger generation [3–5].
1.1. Excessive social media use and its outcomes
While social media offers numerous benefits, it is crucial to address the challenges that have emerged alongside its growth. One of the significant issues that emerged after the explosive growth of social media is the excessive amount of time users spend on these platforms. This aspect has gained attention widely in the Western-educated, Industrialized, Rich, and Democratic (WEIRD) context. In the United States, people average 37 min to 2 hours and 16 min per day of social media usage time [6]. In Germany, before COVID-19, people averaged 2.74 hours of social media usage time, and the COVID-19 restriction extended the average usage time to 3.74 hours [7]. Similarly, in Portugal, a study stated that Portuguese users average 2.5 hours of daily social media usage time [8]. Long-time use of social media has been found to be associated with several problems, including decreased productivity, and sleep disturbances [9,10]. Long time social media use can also bring about depression and other mental health disorders among vulnerable groups [11]. Furthermore, increased time spent on social media was suggested to be related to other conduct problems such as strong alcohol use among adolescents [12].
1.2. Social media use and associated demographic factors
Understanding the impact of excessive social media use requires examining the demographic factors that influence usage patterns. Demographic variables such as age, gender, and income significantly impact social media use time, shaping how different groups engage with digital platforms. Younger users, particularly adolescents, and young adults, are more likely to spend extensive time on social media [13–16]. On the other hand, females are more likely than men to use social media to strengthen their social ties [14–16]. Lastly, income is also reported to be associated with social media addiction, as one study concluded that lower-income people tend to be more likely to suffer from social media addiction [14].
1.3. PSRs and social media use
In addition to demographic factors, the nature of relationships formed on social media, particularly PSRs, plays a critical role in user behavior and psychological outcomes. PSRs refer to the one-way attachment that individuals form with media figures, such as influencers and celebrities [17]. PSRs can offer comfort to individuals and help form certain identities among fans, which could be helpful for those experiencing loneliness or social anxiety [18]. However, they also pose significant psychological risks. First of all, individuals with low self-esteem may develop unhealthy attachment patterns, which would be detrimental to their mental health conditions [19]. Second, PSRs can also be linked to issues such as social isolation and emotional distress [20]. Furthermore, PSRs might lead to unrealistic expectations of relationships, which would bring about difficulties in forming and maintaining healthy interpersonal connections [11,21].
Given the influence of PSRs on user behavior, it is important to understand how these relationships interact with social media usage patterns. However, only a few studies probed the relationship between PSRs and social media use time. One study suggested that individuals with a high amount of social media use is associated with a high level of parasocial interaction toward Korean Pop Stars [22]. Another study also corroborated that social media usage time is a significant predictor of PSRs, particularly parasocial friendships and emotional connections with media personalities [23].
According to the Uses and Gratifications Theory (UGT), individuals would consume media content to satisfy their specific socio-psychological needs, such as the need for social connection, entertainment, or self-validation [24,25]. In the case of social media users, susceptible demographic groups may desire to gratify various social identity-based needs through interactions with influencers and celebrities, which might be associated with extensive social media usage time and PSRs.
1.6. Current study
However, findings from WEIRD contexts may not directly generalize to non-WEIRD populations, because various cultural, socioeconomic, and social norms might significantly influence the social media use pattern, and the formation and impact of PSRs. Currently, studies on social media use in non-WERID contexts offered complicated findings. For example, Chinese netizens average about 2 hours per day social media use time [26]. One study on Chinese adolescents suggested that social media use time alone is not associated with increased anxiety toward their body image. However, when seeking attention through the use of social media, the more social media use time Chinese adolescents, the greater they would suffer from social appearance anxiety [27]. On the other hand, India has more than 600 million social media users [28], one studies on Indian adolescents offered similar findings, suggesting that long time use of social media is associated with increased stress, anxiety, and depression [29]. However, few studies have been conducted on non-WEIRD populations in the Middle East. One cross-cultural study examined the mindful use of social media among Iranian and American users, finding that Iranian participants reported higher levels of mindful social media use, which were associated with lower symptoms of social media addiction [30]. Furthermore, one recent study investigated the role of PSRs with favorite food influencers among Iranian social media users. Their study found that stronger PSRs with food influencers were associated with higher levels of eating disorder symptoms, food addiction, and grazing behaviors [31]. However, the aforementioned studies primarily examined social media use and PSRs within Iran, leaving a significant gap in research on non-WEIRD populations in the broader Middle Eastern context. Qatar’s unique context offers a distinctive environment where rapid economic development and high internet penetration coexist with deeply rooted cultural and traditional norms. This blend creates a unique dynamic in how individuals engage with social media and form PSRs with influencers. Furthermore, due to the impact of COVID-19, nationwide lockdowns and social-distancing measures accelerated the adoption of digital platforms and heightened reliance on Instagram for entertainment, communication, and shopping, pushing more users to interact with influencers and branded content online [1–3]. In fact, recent estimates suggest that Instagram’s penetration in Qatar grew by over 10% between 2020 and 2022, and influencers took on a more prominent role in disseminating public information and shaping consumer habits during the pandemic [4,5]. As of January 2024, Instagram users have been reported to reach 1.7 million in 2024 (up from 1.1 million in 2023) with 35.1% female users and 64.9% male users [32,33].
As a result of the limited research on non-WEIRD Middle Eastern populations and the rapid, COVID-19–induced shifts in Instagram adoption and influencer engagement, understanding these dynamics in Qatar’s digital spaces not only contributes to the social media development in Qatar but also offers novel insights into similar regions of quickly developing or emerging economies, where technological development and modernization meet with conservative values and ideologies. To this end, the present study has the following research questions (RQs):
- RQ1: In linear relationships, how may social media usage time be associated with PSR, gender, age, and income?
- RQ2: How may social media usage time be associated with the interacted influences from PSR and the above demographic factors?
2. Methodology
2.1. Materials and variables
We use secondary data from the data article “Instagram Influencers Attributes and Parasocial Relationship: A dataset from Qatar” [34]. Research ethics approval for the data collection by [34] was obtained from the Qatar University Review Board (number QU-IRB 1195-E/19). [34] declared that participation was entirely voluntary, all respondents were systematically informed about the study’s content and objectives before participation, and all respondents gave informed consent to participate. The data collection happened in 2020, from January 29th to February 16th.
The dataset contains survey information from 574 participants living in Qatar who followed Instagram influencers. Participants were required to follow at least one Instagram influencer from the following areas of expertise: Fashion (N = 26, 4.5%), Traveling (N = 30, 5.2%), Beauty Products (N = 16, 2.8%), Food and Beverages (N = 44, 7.7%), Others (N = 272, 47.4%), and Multiple (N = 186, 32.4%). Among these participants, 38.3% of the participants were males (N = 220) and 61.7% were females (N = 354). The participants were divided into age groups: 18–24 (N = 375, 65.3%), 25–34 (N = 142, 24.7%), 35–44 (N = 45, 7.8%), 45–54 (N = 10, 1.7%), 55–64 (N = 2, 0.3%). Regarding income (measured in Qatari Riyals), there were 6 groups: < 50,000 (N = 354, 61.7%), 50,000–150,000 (N = 116, 20.2%), 150,000–250,000 (N = 48, 8.4%), 250,000–350,000 (N = 29, 5.1%), 350,000–450,000 (N = 10, 1.7%), > 450,000 (N = 17, 3%). About half of the participants spent more than 5 hours every day on social media. There were no participants who did not use social media on a daily basis. More details on the data collection process and basic statistics are available openly online in the original data article [34].
The degree of PSR was measured using the adapted scale based on the study by [35], which included 6 items. Answers were scored on a 5-point Likert scale ranging from “1” being “strongly disagree” to “5” being “strongly agree”. The Cronbach’s alpha value for the PSR scale in the dataset is 0.843 [34].
The variables used for analysis in this study are presented in Table 1.
2.2. Analysis procedure
In the present study, two analytical models for regression were constructed. Model 1 examines multiple linear relationships where time is the outcome variable. Model 1 is as follows.
The posterior distributions of time are in the form of normal distribution where is the mean value of participant
’s number of hours spent on social media every day.
is participant
’s degree of PSR.
is participant
’s gender.
is the age group that participant
belonged to.
is the annual income group that participant
belonged to. Model 1 has an intercept
and coefficients
,
,
, and
.
Model 2 examines the effects of multiple interactions between parasocial and other independent variables toward the outcome time. The two models are separated following the principle of parsimonious model construction, which helps increase the predictive power of the inference [36]. Model 2 is as follows.
Model 2 has an intercept and coefficients
,
, and
.
For statistical analysis, we used Bayesian analysis aided by Markov Chain Monte Carlo (MCMC) algorithms. The analysis procedure and result presentation followed the protocol of MCMC-aided Bayesian analytics for social sciences and psychological research [36]. The dataset used in the present study has a sample size of 574 participants. While this can be considered an acceptable sample size for the measured media-use-related parameters [34], the high skewness in demographic factors (due to the nature of digital social media use) can negatively affect inference accuracy because of the low data points available in some categories. For example, because of the higher proportion of women (61.7% vs. 38.3% men), traditional frequentist analyses can yield less stable parameter estimates under such imbalances or low data counts in specific subgroups. By contrast, a Bayesian framework aided by MCMC simulations allows for more flexible handling of skewed data, as it models parameters as probability distributions rather than fixed values. This approach “borrows strength” from more populated subgroups while explicitly accounting for uncertainty in underrepresented categories. Furthermore, the Bayesian approach treats all parameters probabilistically, and results are interpreted based on the highest probability of occurrence on credible ranges, which helps provide flexible interpretation and high predictive power [37–40].
Analytical models were checked for goodness-of-fit using Pareto-smoothed importance sampling leave-one-out (PSIS-LOO) diagnostics [41,42] to examine if simulated data fit well with the original data. Through the diagnosis run in R, if k values are all below 0.5, the model has healthy goodness-of-fit. k values above the threshold of 0.7 would indicate problematic observations that can affect the inference. Markov properties in the MCMC processes were checked using statistical indicators including the effective sample size (n_eff) and the Gelman-Rubin shrink factor (Rhat). n_eff values over 1000 are deemed sufficient for reliable inference [43], and Rhat values equaling 1 indicate good Markov chain convergence [44,45]. Convergence was also diagnosed using trace plots, Gelman-Rubin-Brooks plots, and autocorrelation plots. The analysis was conducted using the bayesvl package in R [46], using uninformative priors to minimize subjective influences. The MCMC setup was 5000 iterations (including 2000 warm-up iterations) and 4 chains.
3. Result
3.1. Model 1
The PSIS diagnostic result for Model 1 (Fig 1) shows that all k values are lower than 0.5, and there are no problematic observations that may influence the inference. The diagnosis indicates that Model 1 has a healthy goodness-of-fit.
The effective sample size (all n_eff values greater than 1000) and Gelman-Rubin shrink factor (all Rhat values equal 1) show that the Markov chains are well-converged for Model 1 (see Table 2).
The colored lines represent the Markov chains in Model 1’s trace plots (Fig 2). In each plot, the chains fluctuate around a central equilibrium after the warmup period (from 2,000th iteration), suggesting good convergence. Additionally, the Gelman-Rubin-Brooks plots show that Rhat values dropped to 1 during the warm-up period (Fig A1, Appendix). The autocorrelation plots show that problematic autocorrelation among simulated data points within the MCMC processes was quickly eliminated (Fig A2, Appendix).
Estimated posterior coefficients (see Table 2) show that parasocial is positively associated with time ( = 0.05 and
= 0.03). gender, age, and income are all negatively associated with time (
= −0.13 and
= 0.06,
= −0.18 and
= 0.04,
= −0.05 and
= 0.02). The effects have good reliability, since the posterior distributions of parasocial lie almost completely on the positive side, whereas the posterior distributions of gender, age, and income lie almost completely on the negative side (see Fig 3).
3.2. Model 2
The PSIS diagnostic result for Model 2 (Fig 4) also shows that all k values are lower than 0.5, indicating no problematic observations.
The values of effective sample size and Gelman-Rubin shrink factor are also healthy for Model 2. As shown in Table 3, all n_eff values are greater than 1000, and all Rhat values equal 1.
The trace plots (Fig 5), Gelman-Rubin-Brooks plots (Fig A3, Appendix), and autocorrelation plots (Fig A4, Appendix) all indicate that Model 2 achieved good Markov properties.
Estimated posterior coefficients (see Table 3) show that all three effects of the independent variable interactions are negative toward time ( = −0.02 and
= 0.02,
= −0.03 and
= 0.01,
= −0.01 and
= 0.01). The effects have moderate reliability, since the posterior distributions of all three parameters lie mostly on the negative side (see Fig 6).
4. Discussion
The analysis results show that, regarding linear effects, a higher PSR with one or multiple favorite Instagram influencers is associated with higher daily social media usage time. Meanwhile, being male, being older, and having higher incomes all have negative associations with daily social media usage time. When PSRs and the three demographic factors are seen in their interactions, negative associations with social media usage were also found in a similar pattern. To elaborate, among those with high PSR degrees with their favorite Instagram influencer(s), females, young people, and poor people tend to use social media for more hours each day.
The finding that a higher PSR degree with Instagram influencer(s) is associated with higher daily social media usage time is in alignment with prior studies [22,23]. Such a finding can be interpreted through a two-way influence. Firstly, when users have a strong PSR with influencers, they are naturally more inclined to engage with the influencers’ content (viewing, commenting, liking, sharing, etc.), which directs the social media platforms to automatically refine the provision of their desired content through algorithms. Thus, the platforms deliver more tailored content that aligns with the user’s interests, particularly content related to the influencers they follow [47,48]. This personalized content loop would be associated with increased social media usage, as users are drawn into a continuous cycle of engagement [27]. Reversely, as users spend more time on social media, there is a higher chance for them to come across influencers’ content that is engaging and stimulating, which helps form or reinforce the perceived connections between followers and influencers [49]. The felt attachment due to various psychological factors/reasons, as suggested in the UGT [24,25], can further strengthen a PSR through this reciprocal loop of social media content engagement.
The study findings that being male, being older, and having higher incomes have a negative association with daily social media usage time are consistent with earlier studies [14–16,50–52]. Here, we can take into account the specific cultural perspective of the studied population. One possible aspect is that when males in Qatar use social media, they also consider the socially expected masculine characteristics, as well as traditional masculine roles in the family and society, such as showcasing socioeconomic status [53], or highlighting the provider-income maker identity in the family [54,55]. In this sense, long social media usage time might be perceived as excessive leisure or gossiping, undermining male users’ ideal image (both self-image and social image). Thus, males’ social media usage time is negatively associated with their gender identity.
On the other hand, being older is negatively associated with social media usage time. Intuitively, this result is in alignment with the common notion that the older generations are not as familiar with or interested in digital services compared to younger people. However, it can also be viewed through the function of social comparison through social media [27,56]. Specifically, as the Social Comparison Theory suggests, when lacking objective measures, people tend to compare themselves to others to gain a sense of self-evaluation [57]. Younger people relatively lack concrete self-established values compared to the older age groups with more life experiences. In this case, social media is a platform to gain other forms of perceived social validation such as the number of followers, likes, or comments [27].
Lastly, higher income means more options for social interactions and entertainment. Particularly in the context of Qatar, there exists a wide range of premium entertainment services and social events for financially abundant individuals. The capability to afford these options likely decreases the desire and time available for spending on social media.
The analysis results show that among those with high PSR degrees with their favorite Instagram influencer(s), females, young people, and poor people tend to use social media for more hours each day. These findings confirmed the directions of found patterns in the demographic factors upon interacting with PSR toward social media usage time. There are a couple of noteworthy aspects that can be further considered, regarding the interactions of factors and the regional context of the studied population. Although women’s empowerment in Qatar has been on the rise in the past few decades, traditional social norms still hold certain biased views against women [58,59]. Thus, it is possible that female users may develop a stronger attachment to idolized online figures or influencers that provide a sense of social security and comfort. Regarding PSRs among younger social media users, the formation of tightly-knit fandom communities or subcultures, where members often use internet slang and unique communication patterns extensively [60] can increase the appeal toward in-group engagement. This may reinforce the sense of belongingness and commitment, and thus being associated with greater usage time. Regarding financial capability, individuals having strong PSRs are more likely to engage in behaviors such as contributing to the fan economy financially [61,62]. For those with lower income, spending time on social media and engaging in financially affordable behaviors such as retweeting influencers’ tweets [63] or watching them live-streaming [64] are ways to trade time for a sense of contribution to the PSRs.
5. Implications
The present study suggests that cultural and societal norms can have a considerable background influence in shaping the dynamics of social media usage behavior, which is in alignment with extant studies [27,65,66]. From a practical standpoint, the results suggest that interventions aimed at reducing excessive social media use should be tailored to the specific demographic and cultural context of the target population. In the case of Qatar, addressing the unique local characteristics of women, young generations, or lower-income groups might help increase effectiveness.
To be more specific, given that females with strong PSRs tend to use social media more extensively, policy makers could feature relatable female role models or influencers to promote balanced online–offline lifestyles. Such campaigns can speak directly to women’s experiences, showing positive ways of engaging with influencers (e.g., seeking inspiration without excessive scrolling) while acknowledging underlying social pressures.
Also, because the findings suggest that younger users are especially prone to both higher PSRs and higher social media usage, programs that teach digital literacy and self-regulation (e.g., how to set healthy screen-time boundaries) could be integrated into high school or university curricula. These initiatives can reduce the risks of problematic use by encouraging purposeful engagement that still supports healthy identity exploration and peer bonding.
Lastly, for lower-income individuals who may rely heavily on cost-free entertainment such as social media, policymakers and local organizations could facilitate affordable offline social activities or community events. For example, subsidized access to sports clubs, public libraries, or cultural festivals can offer enjoyable and meaningful non-digital outlets.
Future work could explore how influencer-specific characteristics—such as gender, genre, or celebrity status—shape parasocial relationships (PSRs) and social media usage. Longitudinal and qualitative approaches may further deepen our understanding of how user–influencer dynamics evolve over time, especially in non-WEIRD contexts where distinct cultural factors play a key role.
6. Limitations
This study has some limitations. The data used for analysis has high skewness in some examined parameters due to the nature of digital social media usage (such as young age). However, the employed method of MCMC-aided Bayesian analysis helped increase inference accuracy when dealing with such skewed data. Data was also collected from users who followed Instagram influencers, thus may not fully represent patterns of PSRs in other platforms. Furthermore, participants were from Qatar, which might have different psychological nuances compared to social media users in other regions of the world. Additionally, given the generic nature of the original data regarding the PSRs, the demographic factors, and the influencers’ characteristics, cautious approaches are recommended when exploring deeper into the matters based on the present study’s results. Future studies should compare and update the patterns using data from people on other platforms and regions. Qualitative research is also particularly helpful in exploring further the issues behind the relationship between social media use and PSRs.
Appendices
Acknowledgments
This paper is respectfully dedicated to Liu Tie, a cherished friend of the authors, who departed this life on December 26, 2023. His steadfast encouragement and support were invaluable. The authors are profoundly grateful for his enduring contributions. His loyal friendship remains a guiding force in their endeavors. This work serves to honor his enduring impact.
References
- 1. Statista. Number of worldwide social network users 2028. | Statista. 2023. [cited 2024 August 18]. https://www.statista.com/statistics/278414/number-of-worldwide-social-network-users/.
- 2. Kemp S. Digital 2022: Global Overview Report — DataReportal – Global Digital Insights. 2022. [cited 18 Aug 2024. ]. https://datareportal.com/reports/digital-2022-global-overview-report
- 3. Beers Fägersten K. The role of swearing in creating an online persona: The case of YouTuber PewDiePie. Discourse, Context & Media. 2017;18:1–10.
- 4.
Boffone T. The D’Amelio Effect: TikTok, Charli D’Amelio, and the Construction of Whiteness. In: TikTok Cultures in the United States. Routledge; 2022.
- 5. Jensen C. Celebrity Everyday Maker: Public Policy and the Discourse of Celebrity Surrounding Kim Kardiashian. Public Integrity. 2020;23(3):269–80.
- 6. Méndez-Diaz N, Akabr G, Parker-Barnes L. The evolution of social media and the impact on modern therapeutic relationships. Fam J. 2022;30:59–66.
- 7. Helbach J, Stahlmann K. Changes in Digital Media Use and Physical Activity in German Young Adults under the Covid-19 Pandemic - A Cross-Sectional Study. J Sports Sci Med. 2021;20(4):642–54. pmid:35321129
- 8. Côrte-Real B, Cordeiro C, Câmara Pestana P, Duarte E Silva I, Novais F. Addictive Potential of Social Media: A Cross Sectional Study in Portugal. Acta Med Port. 2023;36(3):162–6. pmid:36898203
- 9. Alonzo R, Hussain J, Stranges S, Anderson KK. Interplay between social media use, sleep quality, and mental health in youth: A systematic review. Sleep Med Rev. 2021;56:101414. pmid:33385767
- 10. Raudsepp L. Brief report: Problematic social media use and sleep disturbances are longitudinally associated with depressive symptoms in adolescents. J Adolesc. 2019;76:197–201. pmid:31563733
- 11. Ulvi O, Karamehic-Muratovic A, Baghbanzadeh M, Bashir A, Smith J, Haque U. Social Media Use and Mental Health: A Global Analysis. Epidemiologia (Basel). 2022;3(1):11–25. pmid:36417264
- 12. Brunborg GS, Burdzovic Andreas J. Increase in time spent on social media is associated with modest increase in depression, conduct problems, and episodic heavy drinking. J Adolesc. 2019;74:201–9. pmid:31254779
- 13. Abbasi IS. Social media addiction in romantic relationships: Does user’s age influence vulnerability to social media infidelity?. Personality and Individual Differences. 2019;139:277–80.
- 14. Andreassen CS, Pallesen S, Griffiths MD. The relationship between addictive use of social media, narcissism, and self-esteem: Findings from a large national survey. Addict Behav. 2017;64:287–93. pmid:27072491
- 15. Luo A, Kong W, He H, Li Y, Xie W. Status and Influencing Factors of Social Media Addiction in Chinese Medical Care Professionals: A Cross-Sectional Survey. Front Psychol. 2022;13:888714. pmid:35572263
- 16. Peris M, de la Barrera U, Schoeps K, Montoya-Castilla I. Psychological Risk Factors that Predict Social Networking and Internet Addiction in Adolescents. Int J Environ Res Public Health. 2020;17(12):4598. pmid:32604834
- 17. Tukachinsky R, Walter N, Saucier CJ. Antecedents and effects of parasocial relationships: A meta-analysis. J Commun. 2020;70.
- 18. Gleason TR, Theran SA, Newberg EM. Parasocial Interactions and Relationships in Early Adolescence. Front Psychol. 2017;8:255. pmid:28280479
- 19. Derrick JL, Gabriel S, Tippin B. Parasocial relationships and self‐discrepancies: Faux relationships have benefits for low self‐esteem individuals. Pers Relatsh. 2008;15:261–80.
- 20. Bernhold QS, Metzger M. Older Adults’ Parasocial Relationships with Favorite Television Characters and Depressive Symptoms. Health Commun. 2020;35(2):168–79. pmid:30465614
- 21. Rosaen SF, Dibble JL. Clarifying the Role of Attachment and Social Compensation on Parasocial Relationships with Television Characters. Communication Studies. 2015;67(2):147–62.
- 22. Fitri D, Ruwanti A, Cahyani FI. Korean Hallyu: Parasocial Interaction Study of Teenage K-Popers. Trend Int J Trends Glob Psychol Sci Educ. 2024.
- 23. Tatem CP, Ingram J. Social media habits but not social interaction anxiety predict parasocial relationships. J Soc Psychol Res. 2022;:198–211.
- 24. Blumler JG. The Role of Theory in Uses and Gratifications Studies. Commun Res. 1979;6:9–36.
- 25. Ruggiero TE. Uses and gratifications theory in the 21st century. Mass Commun Soc. 2000;3:3–37.
- 26. Statista. Social media in China. 2024. [cited 2024 December 5]. Available from: https://www.statista.com/topics/1170/social-networks-in-china/.
- 27. Jin R, Le T-T. Eyes on me: how social media use is associated with urban Chinese adolescents’ concerns about their physical appearance. Front Public Health. 2024;12:1445090. pmid:39145157
- 28. Statista. Social media usage in India. 2024. [cited 2024 December 5]. Available from: https://www.statista.com/topics/5113/social-media-usage-in-india/.
- 29. Taddi VV, Kohli RK, Puri P. Perception, use of social media, and its impact on the mental health of Indian adolescents: A qualitative study. World J Clin Pediatr. 2024;13(3):97501. pmid:39350908
- 30. Shabahang R, Zsila Á, Aruguete MS, Huynh HP, Orosz G. Embrace the Moment Using Social Media: A Cross-Cultural Study of Mindful Use of Social Media. Mindfulness. 2024;15(1):157–73.
- 31. Shabahang R, Kim S, Chen X, Aruguete MS, Zsila Á. Downloading appetite? Investigating the role of parasocial relationship with favorite social media food influencer in followers’ disordered eating behaviors. Eat Weight Disord. 2024;29(1):28. pmid:38647734
- 32.
DataReportal. Digital 2024: Qatar. DataReportal – Global Digital Insights. 2024. https://datareportal.com/reports/digital-2024-qatar
- 33. worldpopulationreview. Instagram Users by Country 2025. 2025. [cited 2025 April 22. ]. https://worldpopulationreview.com/country-rankings/instagram-users-by-country
- 34. Al Sulaiti S, Ben Mimoun MS, Elgohary H. Instagram influencers attributes and parasocial relationship: A dataset from Qatar. Data Brief. 2024;53:110128. pmid:38375143
- 35. Kim H, Ko E, Kim J. SNS users’ para-social relationships with celebrities: social media effects on purchase intentions. J Glob Sch Mark Sci. 2015;25:279–94.
- 36. Nguyen M-H, La V-P, Le T-T, Vuong Q-H. Introduction to Bayesian Mindsponge Framework analytics: An innovative method for social and psychological research. MethodsX. 2022;9:101808. pmid:36034522
- 37. Csilléry K, Blum MGB, Gaggiotti OE, François O. Approximate Bayesian Computation (ABC) in practice. Trends Ecol Evol. 2010;25(7):410–8. pmid:20488578
- 38. Dunson DB. Commentary: practical advantages of Bayesian analysis of epidemiologic data. Am J Epidemiol. 2001;153(12):1222–6. pmid:11415958
- 39.
Gill J. Bayesian methods: A social and behavioral sciences approach. CRC press. 2014.
- 40. Wagenmakers E-J, Marsman M, Jamil T, Ly A, Verhagen J, Love J, et al. Bayesian inference for psychology. Part I: Theoretical advantages and practical ramifications. Psychon Bull Rev. 2018;25(1):35–57. pmid:28779455
- 41. Vehtari A, Gelman A, Gabry J. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Stat Comput. 2017;27:1413–32.
- 42.
Vehtari A, Gabry J. Bayesian stacking and pseudo-bma weights using the loo package. 2019.
- 43.
McElreath R. Statistical rethinking: a Bayesian course with examples in R and Stan. 2nd ed. Boca Raton: Taylor and Francis, CRC Press. 2020.
- 44. Brooks SP, Gelman A. General methods for monitoring convergence of iterative simulations. J Comput Graph Stat. 1998;7:434–55.
- 45. Gelman A, Rubin DB. Inference from iterative simulation using multiple sequences. Stat Sci. 1992;7:457–72.
- 46.
La V-P, Vuong Q-H. Package ‘bayesvl’: Visually Learning the Graphical Structure of Bayesian Networks and Performing MCMC with ‘Stan’. 2019.
- 47.
An J, Cho H, Kwak H, Hassen MZ, Jansen BJ. Towards Automatic Persona Generation Using Social Media. In: 2016 IEEE 4th International Conference on Future Internet of Things and Cloud Workshops (FiCloudW), 2016. 206–11. https://doi.org/10.1109/w-ficloud.2016.51
- 48.
Jung SG, An J, Kwak H, Ahmad M, Nielsen L, Jansen BJ. In: Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems, Denver Colorado USA, 2017. 1748–55.
- 49. Bond BJ. Following Your “Friend”: Social Media and the Strength of Adolescents’ Parasocial Relationships with Media Personae. Cyberpsychol Behav Soc Netw. 2016;19(11):656–60. pmid:27732063
- 50. Chen B, Liu F, Ding S, Ying X, Wang L, Wen Y. Gender differences in factors associated with smartphone addiction: a cross-sectional study among medical college students. BMC Psychiatry. 2017;17(1):341. pmid:29017482
- 51. Lopez-Fernandez O. Generalised versus specific internet use-related addiction problems: A mixed methods study on internet, gaming, and social networking behaviours. Int J Environ Res Public Health. 2018;15:2913.
- 52. Zhao J, Jia T, Wang X, Xiao Y, Wu X. Risk Factors Associated With Social Media Addiction: An Exploratory Study. Front Psychol. 2022;13:837766. pmid:35496214
- 53.
Harrell DF, Vieweg S, Kwak H, Lim C-U, Sengun S, Jahanian A. Culturally-Grounded Analysis of Everyday Creativity in Social Media: A Case Study in Qatari Context. In: Proceedings of the 2017 ACM SIGCHI Conference on Creativity and Cognition, 2017. 209–21.
- 54. De Bel-Air F, Safar J, Destremau B. Marriage and family in the Gulf today: Storms over a patriarchal institution?. Arab Humanit. 2018;10.
- 55.
Doha International Family Institute. The state of Qatari families: strengths and challenges. Hamad bin Khalifa University Press (HBKU Press). 2022.
- 56. Ozimek P, Bierhoff HW. Facebook use depending on age: The influence of social comparisons. Comput Hum Behav. 2016;61:271–9.
- 57. Festinger L. A theory of social comparison processes. Hum Relat. 1954;7:117–40.
- 58.
Al-Tamimi NK. Qatari Women’s Engagement in Politics. Qatar Foundation Annual Research Conference Proceedings Volume 2016 Issue 1. Hamad bin Khalifa University Press (HBKU Press). 2016.
- 59. Shockley B, Lari NA, El-Maghraby EAA, Al-Ansari MH. Social media usage and support for women in community leadership: Evidence from Qatar. Women’s Studies International Forum. 2020;81:102374.
- 60. Permatasari SC, Karjo CH. The Influence of Fandom Language in the Word Formation of Indonesian Internet Slangs. E3S Web of Conf. 2023;388:04040.
- 61. Aw EC-X, Labrecque LI. Celebrity endorsement in social media contexts: understanding the role of parasocial interactions and the need to belong. J Consum Mark. 2020;37:895–908.
- 62.
Hung K. Celebrity and influencer in a fan economy: Unfolding the fans’ roles in enhancing endorsement effects. Multidisciplinary perspectives on media fandom. IGI Global. 2020. p. 323–40.
- 63. Kim J, Song H. Celebrity’s self-disclosure on Twitter and parasocial relationships: A mediating role of social presence. Computers in Human Behavior. 2016;62:570–7.
- 64. Lim JS, Choe MJ, Zhang J, Noh GY. The role of wishful identification, emotional engagement, and parasocial relationships in repeated viewing of live-streaming games: A social cognitive theory perspective. Comput Hum Behav. 2020;108:106327.
- 65. Barragan N, Batista A, Hall DL, Silva YN. Social Identity, Social Media Use, and Mental Health in Adults: Investigating the Mediating Role of Cyberbullying Experiences and the Moderating Effects of Gender and Age. Psychol Res Behav Manag. 2024;17:4009–20. pmid:39600921
- 66. Le T-T LT, Jin R. How East Asian colorism influences the use of skin-whitening products: The case of Chinese adolescents. Soc Behav Personal. 2024;52.