The increasing integration of social media into everyday life has sparked significant interest in its psychological effects, particularly its relationship with subjective happiness. Subjective happiness is defined as an individual’s cognitive and affective evaluation of their overall life satisfaction and emotional well-being (Lyubomirsky and Lepper, 1999). As social media platforms become central to communication, entertainment, and self-expression, researchers have attempted to assess whether these digital interactions enhance or undermine users’ happiness.

A substantial body of research suggests that excessive or passive social media use can negatively affect well-being. Passive consumption—such as endlessly scrolling through curated posts—has been linked to increased feelings of envy, loneliness, and reduced self-esteem, all of which may contribute to lower happiness (Twenge, 2017; Mathers et al., 2009). Such outcomes are often attributed to upward social comparisons, where users evaluate their lives against idealized portrayals of others.

Conversely, other studies have found that active and purposeful engagement with social media can enhance happiness. Users who interact with others through comments, messages, or content sharing may experience increased social support, emotional connection, and a stronger sense of community (Doğan, 2016; Hatuka and Zur, 2020). For instance, when individuals use social media to maintain relationships, express themselves creatively, or consume inspiring content, they may feel more connected and emotionally fulfilled.

Despite these contrasting findings, much of the existing research remains generalized, focusing on overall social media use without distinguishing between different types of content or user motivations. As digital influencers gain prominence on platforms like Instagram, YouTube, and TikTok, they represent a distinct category of content producers whose interactions with followers are often personal, emotionally engaging, and aspirational. However, limited research has examined how motivations for following influencers such as entertainment, escapism, or self-expression—relate to subjective happiness.

Moreover, the influence of cultural context on this relationship is frequently overlooked. Since happiness is a culturally sensitive construct, the emotional outcomes of media use may vary depending on the societal norms and values that shape users’ expectations and behavior. Thus, understanding how specific motivations tied to influencer engagement affect happiness, especially within culturally restrained societies,presents a valuable and underexplored area of research.

This study addresses this gap by investigating how users’ motivations for following digital influencers influence their subjective happiness, applying the lens of the Uses and Gratifications Theory (UGT) and considering cultural moderators.

Katz et al. (1973) proposed Uses and Gratifications Theory (UGT), UGT provides a helpful framework that give deeper understanding why individuals select certain media. UGT considers audiences to be active decision-makers who seek out media to satisfy specific psychological and social needs, as opposed to passive consumers. People use media for different reasons, such as entertainment, information gathering, and self-status, or maintaining relationships with others.

UGT is important when we look at social media, and More accurately, why individuals follow influencers. Unlike traditional mass media, which is mostly one-way and entertainment-driven, social media allows users to personalize their experience to a much greater extent (Sheldon and Bryant, 2016; Molnar, 2021). UGT helps explain the range of motivations behind that engagement, from curiosity and inspiration to emotional support and a sense of community.

The motivations for the following influencers can be grouped into several key categories:

Entertainment: Users seek enjoyable content that amuses or relaxes them, especially as a way to unwind or cope with stress (Bae, 2023; Cook et al., 2022; Freitas et al., 2021).

These motivations are not mutually exclusive and often overlap depending on individual user needs and cultural contexts (Price and Applebaum, 2022; Basedow and Kuitunen-Paul, 2022). For example, a user might follow a travel influencer for both escapism and practical tips on destinations, while also identifying with the influencer’s values and aesthetics.

Influencers, in turn, cater to these needs by crafting content that aligns with audience expectations, offering tutorials, behind-the-scenes insights, emotional narratives, or real-time engagement through live sessions (Aldlimi et al., 2025). This dynamic, interactive relationship aligns strongly with UGT’s premise that users turn to media that gratifies their unique personal and emotional needs (Schaap and de Rooij, 2025).

However, the effectiveness of these gratifications—particularly their impact on subjective happiness—may be shaped by external factors, including cultural values and social norms. In culturally restrained societies, where the pursuit and expression of personal gratification may be discouraged, these motivations may not yield the same psychological benefits. Therefore, while UGT is instrumental in identifying why users follow influencers, its integration with cultural dimensions theory can provide a more holistic understanding of how these motivations translate into emotional outcomes like happiness.

Hofstede’s (2011) cultural dimensions theory remains a cornerstone for understanding how values, behaviors, and social interactions differ across cultures. The six dimensions—Power Distance, Individualism vs. Collectivism, Masculinity vs. Femininity, Uncertainty Avoidance, Long-Term vs. Short-Term Orientation, and Indulgence vs. Restraint (IVR)—offer a solid framework for exploring the ways in which societies shape people’s worldviews and day-to-day choices. Each of Hofstede’s six cultural dimensions reflects core values that shape how people see the world, interact with others, and make everyday decisions.

One dimension that has not gotten as much attention—yet is especially relevant in the digital age—is Indulgence versus Restraint (IVR). In cultures that lean toward indulgence, there’s a strong focus on enjoying life, fulfilling personal desires, and openly expressing happiness. On the other hand, restraint-oriented cultures emphasize self-control, discourage seeking pleasure for its own sake, and tend to follow stricter social rules around what’s considered acceptable behavior. These cultural leanings do not just affect individuals—they also influence how people consume media and engage online (Hofstede, 2011; Todea and Harin, 2024).

This is particularly interesting when looking at how people interact with digital influencers. In more indulgent societies, influencer content—whether it’s lifestyle tips, entertainment, or just feel-good posts—is often embraced as a source of joy and escape. It’s easier for users in these cultures to feel a genuine boost in happiness when engaging with such content (Hatamleh et al., 2023a). But in cultures where restraint is the norm, like Jordan, things play out a bit differently. Even if people follow influencers for fun or to feel socially connected, cultural expectations might hold them back from fully enjoying that experience. As Aleqedat (2021) notes, this can water down the emotional benefits that influencer content might otherwise bring.

The sixth and most recently introduced cultural dimension, elaborated in our 2010 publication, adopts Minkov’s concept of Indulgence versus Restraint. This dimension, derived from recent World Values Survey data, complements the earlier framework by addressing aspects not covered by the original five dimensions, particularly those related to happiness research. It shows a weak negative correlation with Long-Term versus Short-Term Orientation (Li et al., 2022).

Indulgence refers to societies that permit relatively free gratification of basic and natural human desires associated with enjoying life and having fun (Heydari et al., 2021). In contrast, restraint characterizes societies that regulate the fulfillment of needs through the imposition of strict social norms and expectations. Data on this dimension have been collected for 93 countries and regions, with validation studies confirming its relevance across diverse cultural settings (Todea and Harin, 2024).

Societies characterized by indulgence typically report a higher proportion of individuals who describe themselves as very happy, coupled with a stronger sense of personal agency and life control. In these societies, freedom of speech is valued highly, leisure time is considered essential, and positive emotions are more frequently recalled. In addition, indulgent societies are also found to have more births among well-educated people, more participation in sport activities, and more cases of obesity where food is in abundance. More permissive sexual norms are also adopted in wealthier indulgent societies and keeping a strict national order is not held in top priority (Todea and Harin, 2024).

On the other hand, in societies with constraints, fewer people are very happy, and individuals tend to experience a greater perception of external control in their lives (Hofstede, 2011). Freedom of speech is viewed as less important, and recreation is ranked lower. Positive emotional episodes are less likely to be remembered, and birth rates tend to remain low even in educated communities. Participation in sports tends to be less widespread, and despite the availability of food, obesity rates are generally lower. In affluent restrained cultures, stricter sexual norms prevail, and there is a stronger emphasis on societal order, often reflected by a higher number of police officers per capita (Aleqedat, 2021).

Geographically, indulgent cultures are predominantly found in South and North America, Western Europe, and parts of Sub-Saharan Africa. By contrast, restraint tends to characterize societies in Eastern Europe, Asia, and much of the Muslim world (Hatamleh et al., 2023a). Mediterranean Europe is in a median location between the two extremes of culture (Chue and Yeo, 2023).

This variable deepens our insights into cultural diversity since it brings to the fore the matter of societies differing in regulating human desires and seeking happiness in addition to differing in values and behaviors.

This research employed a quantitative approach to research, commonly utilized to explore variable interdependencies through numerical data and statistical analysis (Creswell, 2014). The quantitative design allows for hypothesis testing and for drawing inferences about a larger population from a sample in a systematic and objective manner. The quantitative approach is best suited for studies designed to measure theoretical constructs such as social media motivations, subjective happiness, and cultural moderation through standardized measures (Hair et al., 2014). Detection of intricate relationships such as moderation effects is also achieved with a heightened statistical precision (Bryman, 2016).

The target group in this research were Jordanian university students who are a group of individuals constantly exposed to influencer posts and are active users of digital media. University young adults are known to be early adopters of new platforms of emerging media and have complex consumption patterns of media. They are a good population to investigate digital citizenship and well-being (Kircaburun and Griffiths, 2018). Hatamleh et al. (2023a) confirms that university students the most active and largest segment use social media platform.

The convenience sampling technique was utilized to select participants. The technique samples respondents who are easily accessible and willing to participate. The non-probability sampling technique is widely used in social media and psychology studies due to its efficiency and accessibility in sampling unique groups (Etikan et al., 2016). Although convenience sampling is a limitation to the generalizability of a study, it is a good and viable technique to be utilized in exploratory and test theory studies in cases where it is employed in clearly defined boundary contexts.

The data were collected through a self-administered online questionnaire, a time- and cost-effective means of gathering large volumes of data from scattered populations (Bryman, 2016). The questionnaire was sent to students in various Jordanian universities through institutional emails, social media groups, and student networks. The purpose of the study was explained to the respondents, their confidentiality assured, and their right to withdraw at any time notified. Electronic informed consent was gained prior to completing the questionnaire.

The survey contained questions assessing users’ motivations to follow digital influencers (i.e., entertainment, escapism, self-expression), cultural restraint perception, and subjective well-being. The measures for all constructs utilized scales previously validated in other studies to ensure content validity and reliability (Khan, 2017; Whiting and Williams, 2013; Lyubomirsky and Lepper, 1999; Al Omoush et al., 2012).

In order to test hypothesized relationships and interaction effects of moderation, we employed Partial Least Squares Structural Equation Modeling (PLS-SEM), facilitated through SmartPLS 4. PLS-SEM is particularly well-suited to exploratory studies and models incorporating complex latent constructs and interaction effects (Hair et al., 2017). It permits simultaneous measurement model (construct validity and reliability) and structural model (path relations) analysis and is well suited to theory testing and model development in behavioral science (Sarstedt et al., 2019).

Bootstrapping with 5,000 subsamples was used to evaluate path coefficients and cultural restraint’s moderation effect. The guidelines for model fit in composite reliability, average variance extracted (AVE), and discriminant validity were followed Henseler et al. (2015) and Hair et al. (2021).

The study followed set ethical guidelines for conducting studies on human subjects. The participation in the survey was voluntary, and all respondents were clearly told about the aim of the study, confidentiality of responses, and their right to withdraw from the study at any time without penalty.

An informed consent form was presented at the onset of the online survey. The respondents were asked to agree to it before continuing. No personally identifiable data were gathered to maintain respondents’ anonymity and confidentiality.

As Arabic is both the official and local language in Jordan, we translated the survey questionnaire into Arabic to facilitate full understanding among respondents. Translation was carried out by a qualified bilingual professional so that accuracy and cultural sensitivity would be assured.

Since the study entailed no physical or psychological risk and was confined to anonymous survey data, formal Institutional Review Board (IRB) approval was unnecessary according to the ethical regulations of the associated institution. This is in line with norms of minimal-risk survey studies in social science (Bryman, 2016; BPS, 2021).

All data were safely stored and solely utilized for academic purposes. The investigation adhered to the ethical guidelines set forth in the Declaration of Helsinki in ensuring respect, integrity, and transparency in human-subjects research.

To better understand the characteristics of the sample, participants were asked to respond to several demographic questions. These included:

1. Gender

A total of 650 valid responses were collected from university students across several Jordanian universities. The sample consisted of 388 females (59.7%) and 262 males (40.3%). With respect to age, the largest group of respondents were between 21 and 23 years old (45.5%), followed by those aged 18–20 years (38.6%), while 15.8% were in the 24–25 age group.

Regarding academic level, 30.2% of respondents were in their third year, 28.2% in their second year, 23.7% in their fourth year or above, and 18.0% were first-year students. In terms of type of university, 417 respondents (64.2%) attended public universities, while 233 respondents (35.8%) were enrolled in private universities.

Patterns of social media use revealed that 41.5% of participants reported daily usage of 4–6 h, while 27.7% indicated using social media for more than 6 h per day. Additionally, 24.0% reported using social media between 1 and 3 h daily, and only 6.8% reported using it for less than 1 h per day.

These findings offer a detailed description of the young adult student population in Jordan, in terms of their varied academic levels, sustained use of social media, and representative demographic structure in terms of gender. This type of sample is suitable for studies examining digital behavior and psychological consequences in Arab societies (Kircaburun and Griffiths, 2018; Hatamleh et al., 2023c).

This survey employed validated and well-recognized measurement scales to evaluate the primary constructs: Motivations to follow digital influencers, Cultural Restraint, and Subjective Happiness. The items were measured on a Likert scale of 7 points from strongly disagree (1) to strongly agree (7). For testing the scales’ reliability, Cronbach’s alpha (α) was estimated, and all of them were above the criteria value of 0.70, demonstrating good internal consistency (Hair et al., 2021).

The Motivations for following digital influencers construct was theorized to be a second-order reflective construct with five first-order dimensions corresponding to the hypotheses of this study:

Entertainment

These dimensions are grounded in the Uses and Gratifications Theory (UGT), which emphasizes that individuals actively engage with media to fulfill psychological and social needs (Katz et al., 1973; Whiting and Williams, 2013). The measurement items were adapted from prior validated studies (Khan, 2017; Buzeta et al., 2020; Whiting and Williams, 2013; Gao et al., 2017).

In addition to Motivations for following digital influencers, the study measured:

Cultural Restraint, using a scale adapted from Al Omoush et al. (2012), which reflects the extent to which individuals regulate emotional expression and behavior on social media.

The results showed that all constructs demonstrated high reliability and validity, as reflected in Cronbach’s alpha, Composite Reliability (CR), and Average Variance Extracted (AVE). Specifically:

Motivations for following digital influencers (second-order) showed excellent reliability (α = 0.947; CR = 0.949),

These results confirm that the measurement model is robust and well-suited for hypothesis testing (see Table 1 and Figures 2, 3).

Table 2 presents the outer loadings of all measurement items on their respective constructs, reflecting how strongly each item contributes to its latent variable in the measurement model. Outer loadings are critical indicators of indicator reliability in structural equation modeling (SEM), and loadings above 0.70 are generally considered acceptable, while loadings above 0.80 are considered strong (Hair et al., 2021).

The items measuring Cultural Restraint (CR1–CR4) demonstrated excellent outer loadings, ranging from 0.855 to 0.886, confirming that these items strongly represent the construct. This suggests that participants consistently perceived these items as valid indicators of cultural restraint in the context of digital media use.

For the second-order construct of Motivations for Following Digital Influencers, the five first-order dimensions (Entertainment, Escapism, Information-Seeking, Self-Expression, and Social Interaction) also showed acceptable to excellent item loadings:

Entertainment (E1–E4) had loadings between 0.772–0.876, indicating good reliability, although its second-order loadings were slightly lower (0.633–0.728), suggesting some variation in how entertainment items contribute to the higher-order construct.

The Subjective Happiness construct (SB1–SB4) displayed strong loadings for most items (SB1: 0.846, SB3: 0.879, SB4: 0.920), though SB2 (0.599) fell slightly below the 0.70 threshold. While one lower-loading item can be acceptable, future research may consider refining or rephrasing this item to improve scale performance (Hair et al., 2021).

The interaction/moderation effect between Cultural Restraint and Motivations (Cultural Restraint × Motivations for following digital influencers) reported a fixed loading of 1.000, as expected in PLS when creating interaction terms using the product indicator approach (Henseler and Fassott, 2009).

Overall, the measurement model shows excellent indicator reliability and supports convergent validity, with nearly all items exceeding the recommended loading thresholds. This justifies moving forward to the structural model analysis and hypothesis testing. The slightly lower loading on SB2 does not undermine the scale’s overall reliability but suggests it could be reviewed in future scale refinements.

These measurement results, together with the reported composite reliability and AVE values, confirm that the model is statistically robust and well-suited for testing the hypothesized relationships (see Table 2).

Table 3 presents the Heterotrait-Monotrait ratio (HTMT) values, which are used to assess the discriminant validity of the measurement model. Discriminant validity determines whether a construct is truly distinct from other constructs in the model, ensuring that each latent variable captures unique phenomena not reflected by other variables (Henseler et al., 2015).

The HTMT ratio is calculated by comparing the average correlations across constructs to the average correlations within constructs. As a rule of thumb, an HTMT value below 0.85 (conservative) or below 0.90 (liberal) indicates adequate discriminant validity, while values above these thresholds may signal problematic overlap between constructs (Hair et al., 2021).

In this study, all HTMT values were below the recommended thresholds, confirming that the constructs are sufficiently distinct. Although some motivational pairs (such as Self-Expression ↔ Entertainment = 0.838) approached the conservative cutoff, they remained within acceptable limits given their theoretical relatedness (See Table 3).

Table 4 reports the variance inflation factor (VIF) values, which are used to assess multicollinearity among the measurement items in the model. Multicollinearity refers to high correlations between predictors, which can distort regression estimates and weaken the reliability of statistical results (Hair et al., 2021).

As a general rule, VIF values:

Below 3.3 are considered acceptable and indicate no serious multicollinearity problems.

Above 5 signal potential multicollinearity concerns that require further investigation.

Above 10 are typically seen as problematic.

In this study, most VIF values ranged between 1.2 and 4.0, with the highest observed VIF at 4.054 (SI3), staying below the common threshold of 5. This indicates that multicollinearity is not a major concern in the measurement model. The interaction term (Cultural Restraint × Motivations for following digital influencers) has a fixed VIF of 1.000, which is standard in moderation models.

Overall, these results confirm that the indicators contribute uniquely to their constructs and that the structural estimates are not distorted by collinearity.

Table 5 reports the R-square (R²) and adjusted R-square values, which indicate the amount of variance explained by the predictors for each dependent variable. In structural equation modeling, R² measures how well the independent variables account for the variance in the dependent variable, while the adjusted R² accounts for the number of predictors, providing a more conservative estimate (Hair et al., 2021).

For this analysis, dependent variable Subjective Happiness has an R² of 0.718 and an adjusted R² value of 0.716, demonstrating that predictors such as motivations to follow digital influencers and cultural restraint account for nearly 72% of subjective happiness variability. This is a strong explanatory power and implies that the model is very effective in measuring the psychological consequences of social media consumption.

The first-order motivational facets (Entertainment, Escapism, Information Seeking, Self-Expression, Social Interaction) have good R² values between 0.627 and 0.781, promising that the second-order construct (Following digital influencers’ motivations) indeed explains variation in these facets.

In general, these findings validate the model’s stability and capacity to meaningfully explain subjective happiness and associated motivational variables.

Table 6 presents the Q² predict, Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) values, which are used to assess the predictive power and accuracy of the model.

Q² predict evaluates the model’s out-of-sample predictive relevance, with values above 0 indicating good predictive ability (Hair et al., 2021).

For dependent variable Subjective Happiness, the model attained a Q² predict of 0.710, an RMSE of 0.543, and an MAE of 0.419. These findings show that the model possesses excellent predictive significance and good accuracy in subjective happiness outcome forecasting.

The other motivational dimensions (Entertainment, Escapism, Information Seeking, Self-Expression, Social Interaction) also show high Q² predict values (ranging from 0.625 to 0.779) and acceptable RMSE and MAE values, further supporting the model’s predictive strength.

Overall, these results confirm that the model not only explains but also accurately predicts key outcomes related to digital influencer engagement and subjective happiness.

Table 7 reports the f² effect size, which measures the strength of the contribution of each predictor to the dependent variable. According to Cohen (2013), f² values can be interpreted as:

0.02 = small effect

For Subjective Happiness, the results show:

Motivations for following digital influencers → Subjective Happiness (f² = 2.187) → a very large effect, indicating that motivations play a major role in explaining happiness outcomes.

For the motivational dimensions, the f² values are also very large, especially for:

Self-Expression (f² = 3.574).

These results confirm that the motivational constructs have a strong influence on both the first-order dimensions and subjective happiness, highlighting their importance in the model (see Table 7).