Differential Facebook use : differential outcomes : the role of negative social comparison in the influence of active and passive use on self-perceptions

(1)

Differential Facebook use – differential outcomes: The role of negative social comparison

in the influence of active and passive use on self-perceptions Mariia Emelianova

University of Amsterdam

Master’s Thesis

Student ID number: 10488510 Graduate School of Communication

Research Master’s programme Communication Science Supervisor: Rinaldo Kühne

(2)

Abstract

Facebook as the most popular social network globally has been extensively studied in

relationship to psychological well-being. The outcomes of these examinations, however, have been contradictory, linking Facebook use to negative, as well as positive psychological outcomes. The objective of this study is to provide a more complete understanding of the relationships between different types of Facebook use, negative social comparison, self-perceptions, life satisfaction and personal values. A survey of 232 emerging and young adults (aged 18-30) showed that passive Facebook use was related to a greater degree of negative social comparison, which in turn related negatively to self-perceived social competence, physical attractiveness and career success, confirming full mediation of this path by negative social comparison. Active Facebook use, on the contrary, was positively related to self-perceived social competence and perceived career success. Life satisfaction had an unexpected negative effect on

self-perceptions, as the data showed that the relationship between passive Facebook use and negative social comparison was stronger for the ones more satisfied with their life. The personal values achievement, hedonism, humility and stimulation had no influence on the relationship between passive Facebook use and negative social comparison. The interpretation of these findings and directions for future research are discussed.

Keywords: social networks, Facebook, passive Facebook use, active Facebook use, human values, life satisfaction, negative social comparison, self-perception

(3)

Differential Facebook use – differential outcomes: The role of negative social comparison

in the influence of active and passive use on self-perceptions

Young adults (18-29) are the main users of social networking sites among all age

categories. Among online adults 72 % use Facebook ("Social Networking Use", 2015). A recent study suggests that young adults not only spend more time on Facebook, but also get more emotionally impacted by it then older adults (Hayes, Van Stolk-Cooke, & Muench, 2015). It is thus of high relevance to study the effects Facebook use may have on young people.

Facebook has been shown to have both positive as well as negative outcomes. The current study aims to provide a deeper understanding of the mechanisms that lead to negative or positive outcomes, and to discuss the boundary conditions of these effects. On the one hand, Facebook use has been shown to lead to a decline in well-being and to other negative socioemotional consequences (Błachnio, Przepiórka, & Pantic, 2015; Burke, Marlow, & Lento, 2010; Kross et al., 2013). Psychologists and the media have even coined the term “Facebook depression” (Blease, 2015). Delving deeper into the explanations of these negative effects, other studies have shown that social comparison (Steers, Wickham, & Acitelli, 2014) and in particular negative social comparison (Haferkamp & Krämer, 2011; Vries & Kühne, 2015) may play an influential role in this process. According to Festinger (1954), when one evaluates the self as doing worse than others (e.g. engages in negative social comparison) this feeling will lead to less favorable self-perceptions. This notion has been supported by experimental (Haferkamp & Krämer, 2011) as well as correlational research (de Vries & Kühne, 2015) among young people. According to Arnett (2004), the period between 18 and 25 (and in some cases even until 29) years of age is a self-focused time of identity formation. Self-perceptions can be an essential part of identity and are thus highly relevant for young people.

On the other hand, Facebook use has also been linked to positive outcomes, such as reduced feelings of loneliness (Burke et al., 2010) feelings of support (Frison & Eggermont, 2015; Koroleva, Krasnova, Veltri & Günther, 2011) and increase in connectedness (Deters & Mehl, 2013). As an explanation for such contrasting findings, research suggests that differential Facebook use may lead to differential outcomes (Burke, Marlow & Lento, 2010; Frison &

(4)

Eggermont, 2015). Responding to the request for distinguishing the different types of Facebook use (Krasnova, Wenninger, Widjaja, & Buxmann, 2013; Vries & Kühne, 2015) the current study aims to investigate the differential effects that passive and active Facebook use have on self-perceptions of young users, while looking at the role negative social comparison plays in these relationships and taking life satisfaction and personal values into account.

Research indicates that the influence of Facebook may depend on certain boundary conditions, e.g. individual differences. Life satisfaction has been shown to be one of the factors that can protect individuals from the detrimental effects of negative social comparison (de Vries & Kühne, 2015). Additionally, while values have been considered in political communication (for instance, value framing; e.g. Brewer, 2002) and persuasive communication (Han & Shavitt, 1994), they have not received the same attention in research on social media. Values have been proposed to play an important role in media effects processes (e.g., Valkeburg and Peter), but research in this area has been quite scarce (Postman, Bruner & McGinnies, 1948; Shen and Edwards, 2005). To my knowledge, Schwartz basic human values have not been integrated in a media effects examination model, despite that this theory was used for more than twenty years (Cieciuch, Schwartz, & Vecchione, 2013). This study aims to fill this gap and test the role of values in the social media effects process.

Theoretical background

Differential Facebook Use, Social Comparison and Self-Perception

Ever since the advent of mass communication theory there have been warnings of the negative effects of media and later hopes for its positive influence. Also the user was either regarded as a passive receiver of the magic bullet of media information (DeFleur &

Ball-Rokeach, 1982) or as an empowered active user selecting the media that cater to his or her needs (Katz, Blumler & Gurevitch, 1974). This debate, however, is ongoing as new media platforms emerge and the study and empirical testing of the initial theories continues.

As Facebook can be used in different ways, resulting in different effects on young people’s well-being (Burke, Marlow & Lento, 2010; Frison & Eggermont, 2015), the current study differentiates between passive and active Facebook use.

(5)

Negative social comparison takes place when an individual compares him- or herself to another person in a negative way, implying that this other is better off (de Vries & Kühne, 2015; Festinger, 1954). In previous studies it has been shown that for emerging adults more intense Facebook use is related to a greater degree of negative social comparison on Facebook (Lee, 2014; de Vries & Kühne, 2015). To distinguish the differential impact that passive and active Facebook use may have on negative social comparison, the impact of these two types of Facebook use will be discussed next.

Passive Facebook use (browsing the information of others on Facebook, e.g. their status updates, pictures, public conversations with friends) has been shown to relate to negative psychological outcomes, such as loneliness (Burke et al., 2010), envy (Krasnova et al., 2013), depressed mood (Frison & Eggermont, 2015) and the decline in affective well-being (Verduyn et al., 2015). Moreover, passive Facebook use has also been shown to predict negative social comparison (Haferkamp & Krämer, 2011; Qiu, Lin, Leung, & Tov, 2012). This is why I hypothesize that intense passive Facebook use will lead to a higher degree of negative social comparison (i.e., the frequency of negative thoughts about the self while comparing own life to others when seeing information about others on Facebook) (H1).

The research on active Facebook use is displaying mixed outcomes. For the most part active Facebook use (posting own status updates, pictures etc.) has been known to predict positive psychological outcomes, such as reduced feelings of loneliness (Burke et al., 2010), feelings of support (Koroleva et al., 2011; Frison & Eggermont, 2015) and increase in connectedness (Deters & Mehl, 2013). In contrast, active public Facebook use also has been shown to predict depressed mood (Frison & Eggermont, 2015). This can be explained by the reaction of the author of the posted content to the feedback (or lack thereof) to the posts. Receiving a low number of likes and/or negative comments can lead to negative psychological outcomes (Frison & Eggermont, 2015). Thus, active Facebook use has been shown to have mostly positive outcomes. However, in some instances the outcomes may also be negative. As negative social comparison may play a role in these negative outcomes, especially when they have to do with receiving feedback (likes or comments) on the actively posted content, I hypothesize that the effect of active Facebook use on negative social comparison should be weaker than the effect of passive Facebook use (H2).

(6)

According to Festinger, when one evaluates the self as doing worse than others (e.g. engages in negative social comparison) this feeling will lead to less favorable self-perceptions (1954). Experimental research has demonstrated that when emerging adults view Facebook profiles (thus engage in passive Facebook use) of physically attractive or successful peers, it can have a negative impact on their self-perception in the respective domains (Haferkamp & Krämer, 2011). Additionally, correlational research suggests that negative social comparison negatively influences self-perceptions in the domains of physical appearance and social competence (de Vries & Kühne, 2015). Therefore, I hypothesize that negative social comparison will negatively predict a) perceived social competence, b) perceived physical appearance and c) self-perceived career success (H3).

Additionally, previous studies have investigated the outcomes of negative influences of Facebook use on self-perceived physical appearance: a study of young women (Fardouly, Diedrichs, Vartanian, & Halliwell, 2015) and young people and body perceptions (Manago, Ward, Lemm, Reed, & Seabrook, 2015). The study of Fardouly et al. (2015) showed that

Facebook use was positively correlated with self-objectification (which is in turn related to body dissatisfaction (Fitzsimmons-Craft & Bardone-Cone, 2012)) and that this path was mediated by appearance comparisons in general as well as by comparisons to one’s peers. That study,

however, did not distinguish between the different types of Facebook use and social comparison. The study of Manago et al. (2014) found that Facebook involvement predicted objectified body consciousness, which in turn predicted greater body shame for both young women and men. In that study the different types of Facebook uses were collapsed into one overall Facebook use measure. The current study can shed some light on these findings in a more detailed way by distinguishing the different types of Facebook use and by examining the role negative social comparison plays in it. As this has already been established in a previous study (de Vries & Kühne, 2015), I hypothesize that passive Facebook use is negatively related to self-perceptions through negative social comparison. Meaning that there is an indirect effect of passive Facebook use on a) perceived social competence, b) perceived physical attractiveness and c) self-perceived career success (H4).

As has already been pointed out previously, in general active Facebook use tends to show a positive effect on individuals’ psychological outcomes (Burke et al., 2010; Koroleva, Krasnova,

(7)

Veltri & Günther, 2011; Deters & Mehl, 2013; Frison & Eggermont, 2015). The kind of

information that individuals post about on their Facebook pages is what concerns them personally and/or interests them. As emerging adulthood is a self-focused time of identity formation (Arnett, 2004), it is therefore logical that emerging adults would post and react to topics that are most relevant to them and concern their (or their peers’) identity: social and romantic relationships, physical appearance, and work-related or recreational achievements. These could at the same time be domains for social comparison. A recent study has shown that among emerging and young adults, narcissists were found to be posting status updates with information related to their achievements as well as the work on their physical appearance - diet and exercise (Marshall, Lefringhausen, & Ferenczi, 2015). The same study also found that updating about ones achievements as well as about ones social activities positively predicted getting likes on those posts, while, e.g. updating about intellectual topics (e.g. politics and current events) was negatively associated with getting likes. Previous research has also found that getting positive feedback on one’s profile enhanced ones self-esteem and well-being (Valkenburg, Peter, & Schouten, 2006). More positive self-perceptions, in turn, have been shown to strongly relate to well-being (Diener & Diener, 1995). Taking into consideration that active Facebook use showed mostly positive influences on psychological outcomes, and that posted information about own achievements and physical appearance predicted positive feedback (likes), it was hypothesized that active Facebook use will positively predict self-perceptions of a) social competence, b) physical attractiveness and c) career success (H5).

The Guarding Influence of Life Satisfaction

As has been shown in a previous study of negative social comparison by de Vries and Kühne (2015), life satisfaction can play a protective role and moderate the influence of passive Facebook use on negative social comparison: for those reporting higher life satisfaction the link between passive Facebook use and negative social comparison becomes weaker. This can be explained by individual psychological perceptions in the processing of social information. It has been shown that those high in life satisfaction process self-relevant information (including social comparison information) more positively (Cummins & Nistico, 2002). Moreover, social

information can be interpreted in ways that lead to more positive self-perceptions (Suls, Martin, & Wheeler, 2002) or one can simply select comparison targets that are worse off than oneself and

(8)

avoid the negative influence on self-perceptions (Cummins & Nistico, 2002). Therefore we hypothesize that the positive relationship between passive Facebook use and negative social comparison on Facebook is moderated by life satisfaction: it is weaker among emerging adults who are more satisfied with their life (H6). Additionally, I hypothesize that the indirect negative relationship between passive Facebook use and self-perceptions through negative social

comparison is reduced among emerging adults who report greater satisfaction with life: for a) self-perceived social competence, b) self-perceived physical attractiveness and c) self-perceived career success (H7).

The Moderating Role of Values

According to Schwarz (2012), values can be defined as beliefs that are linked to affect and at the same time also as desirable goals that extend beyond specific situations and that serve as evaluation criteria for individuals in life. Values are ordered by importance relative to one another and the relative importance of multiple values impacts individuals’ actions. Values are assumed to form a circular motivational continuum (Cieciuch et al., 2013). Schwartz’ theory of basic values is probably the most widely used value theory, with a vast amount of empirical evidence to substantiate it. The theory has been tested in more than 75 countries in various samples, elaborated and improved during the last two decades (Cieciuch et al., 2013). Values are constructs that were initially believed to be essential parts of a national culture, as characteristics of larger social groups, such as nations and were perceived as shared meaning systems within a culture (Hofstede, 2001; Lehman, Chiu, & Schaller, 2004). However, in light of recent empirical findings of Fischer and Schwartz (2011) where values differed more within countries than between them, which made the authors question the assumption that culture determines values, Schwartz has proposed a sociological approach to values. According to Schwartz, the culture of a society is latent and external to the individual and can be noticed through the working of societal institutions: their organization and practices. These institutions implement their policies upon the individuals within societies by rewarding or punishing the desired or undesired beliefs and

behaviors, in this way implementing the values of the society. But personal individual differences account for in how far these societal norms get accepted or interiorized by the individual. As no two individuals are the same in their psychological constitution, this can explain the divergence

(9)

between values within a given society and the similarities within the values of individuals from different societies (Schwartz, 2014).

As Valkenburg and Peter suggest in their Differential Susceptibility to Media Effects Model (2013), values may play a moderating role in the path from media use to media response states. In the same article Valkenburg and Peter also suggest that individuals tend to select the media that does not diverge too much from what was previously known and close to the users, this also concerns their beliefs and cognitions. In this way, the user searches for media content congruent with his or her prior dispositions, which is then processed more efficiently, as it corresponds to the users pre-existing mental schemata. In a classical study by Postman, Bruner and McGinnies (1948) it was shown that values play an important role in the selective perception of information. It took less time for the participants to recognize words that were related to the values which were important to the individual. In a more recent study, Shen and Edwards (2005) showed that framing effects were the outcomes of an interaction between individual core values (humanitarianism and individualism) and frames of a welfare reform. These effects can be also due to the moderating effect of social context on psychological outcomes of media use. If the context (values and norms in the social environment) and media content are convergent, media effects may be amplified (Valkenburg & Peter, 2013). Values thus can serve as moderators of media processes.

Constructs similar to values playing a part in media processes. Research on the role of

values in media effects processes is scarce. However, insights from other branches of research can be used to derive hypotheses. Reiss’ theory of 16 basic desires was suggested to be

implemented as predictors of media enjoyment (Reiss & Wiltz, 2004). The theory posits that individuals have 16 basic desires (e.g. power, independence, curiosity etc.). When the desire is fulfilled, one experiences a certain joy, connected to this desire, e.g. fun for the desire of social contact. According to the authors, core values are implied from these basic desires (Reiss & Wiltz, 2009). Another construct connected to values are personality traits. In a meta-analysis of 60 studies of the relationships between the Big Five personality traits (neuroticism, extraversion, openness to experience, agreeableness and conscientiousness) model and Schwartz personal values, it was found that the more cognitively based traits, e.g. openness to experience, show stronger relationships with values, while more emotionally based traits, e.g. neuroticism, show

(10)

little or no relationships with values (Parks-Leduc, Feldman, & Bardi, 2015). A meta-analysis conducted by Hoffner and Levine (2005) about the enjoyment of fright and violence in

entertainment media showed a mediation effect of personality traits: those higher in sensation seeking and lower in empathy were more likely to enjoy frightening and violent media. Also Facebook use is now being studied in light of personality traits as moderators of psychological outcomes (Simoncic, Kuhlman, Vargas, Houchins, & Lopez-Duran, 2014).

Values and negative social comparison. To the author’s knowledge, no research has yet

been done on the influence of individual values (as conceptualized by Schwartz) on the process of social comparison. Although some individual values are intuitively connected with the kind of content Facebook provides for social comparison, such as the values of Hedonism or

Achievement. However, studies have been carried out on the influence of personality traits on social comparisons. It has been found that those high in openness to experience are less eager to indulge in social comparison than those low in this trait (Buunk & Gibbons, 2007). As

personality traits, and especially the more cognitively based traits as openness to experience, are related to values (Parks-Leduc, Feldman, & Bardi, 2015), it is thus theoretically probable that values may have a meaningful relationship to social comparison. I expect that certain values will have a protective and others a detrimental effect on psychological outcomes and in this way either promote higher levels of negatives social comparison or, on the contrary, reduce its negative outcomes, depending on the effect of the specific value.

The four values as moderators. Schwartz describes ten values, which are believed to be

universal, according to the underlying motivation of each value (Schwartz, 2012). Out of all the ten values discussed by Schwartz I have reason to believe that there are four specific values that will affect the path between passive Facebook use and negative social comparison. These values are achievement, humility, hedonism and stimulation.

Achievement. The goal underlying the achievement value is defined as reaching towards

personal success by showing competence that is congruent with social standards (Schwartz, 2012). From this definition of achievement it is already clear that it is linked to the social dimension, thus making it relevant for a connection with social comparison. Empirical evidence suggests a link between achievement goals and social comparison (Darnon, Dompnier, Delmas, Pulfrey, & Butera, 2009; Régner, Escribe, & Dupeyrat, 2007; Van Yperen & Leander, 2014).

(11)

Van Yperen and Leander (2014, p. 677), argue for “the overpowering effect of social comparison information” when one makes self-evaluations of one’s performance. However, information for social comparison relevant to achievement can be presented on social networks, such as

Facebook, in a biased, overly positive manner (Dorethy, Fiebert, & Warren, 2014). In addition, when one is comparing oneself to someone similar but superior in a self-relevant domain, envy may ensue. Envy and social comparison are strongly linked to each other: Smith and Kim (2007) define envy as an unpleasant and painful mix of feelings caused by comparing oneself to

someone who has something one desires (thus indulging in negative social comparison). It has been shown that passive Facebook use leads to envy and that posts of Facebook friends’ about their successes in their jobs is mentioned as one of the envy-inducing incidents on Facebook (Krasnova et al., 2013). Therefore, for those high in achievement value, it may be more challenging to watch others post about their achievements without resulting to negative social comparison. I therefore hypothesize that the achievement value will moderate the path between passive Facebook use and negative social comparison in a positive way: those high in

achievement will result to more negative social comparison (H8a). Additionally, I hypothesize that the indirect negative relationship between passive Facebook use and self-perceptions through negative social comparison is heightened among young adults who report greater levels of

valuing achievement: for b) self-perceived social competence, c) self-perceived physical attractiveness and d) self-perceived career success (H8).

Humility. As defined by Schwartz, the humility value includes recognizing one’s

insignificance in the larger scheme of things. This value is found between the broader concepts of conformity on the one hand and universalism on the other hand (Cieciuch, Davidov, Vecchione, Beierlein, & Schwartz, 2014). While humility may also have negative associations, such as low self-esteem, true humility is considered a virtue that includes accurate self-assessment (Tangney, 2000). Humble people possess the skill to perceive themselves and others without distortions (Chancellor & Lyubomirsky, 2013; Peterson & Seligman, 2004) and can be genuinely happy for the successes of others (Chancellor & Lyubomirsky, 2013). Moreover, those high in self-esteem have been shown to have a positive view of humility (Exline & Geyer, 2004). Results of a correlational study showed that heavy Facebook users were lower in self-esteem and that this relationship was mediated by greater exposure to upward social comparisons (e.g. comparing oneself with someone who is doing better) on social media (Vogel, Rose, Roberts, & Eckles,

(12)

2014). Humble people have a secure and accepting self-concept (Chancellor & Lyubomirsky, 2013; Peterson & Seligman, 2004), which can also be called self-compassion (Chancellor & Lyubomirsky, 2013). In turn, self-compassion has a negative association with social comparison and self-rumination (e.g. being conscious of oneself in a negative way) (Neff & Vonk, 2009). It can thus be argued that humility will have a diminishing influence on negative social

comparison. True humility can be contrasted to narcissism on the one hand and low self-esteem on the other hand (Chancellor & Lyubomirsky, 2013). Narcissism, in contrast to humility, has been found to motivate competitiveness (Wallace & Baumeister, 2002). These outcomes lead us to believe that the value of humility may moderate the relationship between passive Facebook use and negative social comparison in a negative way: those higher in humility will be less subject to negative social comparison (H9a).Additionally, I hypothesize that the indirect negative

relationship between passive Facebook use and self-perceptions through negative social

comparison is reduced among young adults who report greater levels of valuing humility: for b) self-perceived social competence, c) self-perceived physical attractiveness and d) self-perceived career success (H9).

Hedonism. Schwartz describes the defining goal of the hedonism value as seeking

pleasure for oneself. Similarly, other theorists define hedonism as openness to pleasurable experience (Veenhoven, 2003). In a similar vein, casual leisure is understood to be hedonic, meaning free time spent in enjoying some activity that is perceived to be pleasurable (Stebbins, 2001).Viewing hedonism in the light of the process of experiential consumption, Jantzen and collegues (2012) posit that one’s success in life can now be measured in terms of experience. But that also implies that with centering on pleasurable experiences new fields of social comparison and competition are created. One of the critique points on hedonism is that it may impoverish social interaction between individuals when they resort to conspicuous consumption (Veenhoven, 2003). Conspicuous consumption is a term first introduced by the economist and sociologist Thorstein Veblen in 1899, which refers to buying expensive goods in order to display ones wealth and high social status. In his book, Veblen also discussed the role of leisure in social comparisons (Veblen, 1899). Veblen’s theory bears similarities to Festinger’s social comparison theory (Lambert, Bergman, & Prandy, 2005). According to Veblen, it is in the nature of people to compare themselves to others and match or surpass others in order to gain esteem. Leisure and consumption can be used to show one’s success and achievements to others (Lambert et al.,

(13)

2005). In addition, as was the case also with the achievement value, if one is comparing oneself to someone similar but superior in a self-relevant domain, in this case the domain of pleasurable experiences, one may become envious. It has been shown that passive Facebook use leads to envy and that, interestingly, information about travel and leisure of one’s Facebook friends seems to be the most frequently mentioned cause of envy (Krasnova et al., 2013). Therefore, for those high in hedonism, it may be more challenging to see others on Facebook having fun, eating out and going on holidays (and this is exactly the information you typically encounter on social media) without resulting to negative social comparison. I therefore hypothesize that the hedonism value will moderate the path between passive Facebook use and negative social comparison in a positive way: those high in hedonism will result to more negative social comparison (H10a). Additionally, I hypothesize that the indirect negative relationship between passive Facebook use and self-perceptions through negative social comparison is heightened among young adults who report greater levels of valuing hedonism: for b) perceived social competence, c) self-perceived physical attractiveness and d) self-self-perceived career success (H10).

Stimulation. The stimulation value, according to Schwartz, is defined by the goals of

excitement, novelty and challenge in life (Schwartz, 2012). In a review of the literature on the topic by Roberti (2004), the terms stimulation and sensation seeking are used interchangeably: individuals engaged in sensation seeking tend to favor behaviors that increase the amount of stimulation they experience. One can satisfy a preference for stimulation by engaging in travel or sports or more deviant behavior such as taking drugs or alcohol (Roberti, 2004). Research shows that the ones seeking stimulation travel more as well as to less known places and are able to enjoy unusual and even unpleasant art forms (Rawlings, 2003; Roberti, 2004; Zuckerman, 1994). Interestingly, those interested in higher stimulation do not view the environment around them as threatening (Roberti, 2004) and prefer to be in the company of other stimulation seekers, which also influences their perception of threat in a negative way (Roberti, 2004). As high sensation seekers are found to experience less anxiety and stress (Roberti, 2004), it may be that they are less prone to negative social comparison in general. As social comparison has been shown to correlate with anxiety (Gibbons & Buunk, 1999; Butzer & Kuiper, 2006 i), it may be that those high in sensation seeking, on the contrary, are not engaging eagerly in negative social

comparison. Moreover, the stimulation value has been shown to strongly correlate with the trait of openness to experience (Parks-Leduc et al., 2015). In turn, it has been found that those high in

(14)

openness to experience are less prone to social comparison than those low in this trait (Buunk & Gibbons, 2007). Thus, I hypothesize that the stimulation value will moderate the path from passive Facebook use to negative social comparison negatively: those high in stimulation will be less likely to fall prey to negative social comparison (H11a). Additionally, I hypothesize that the indirect negative relationship between passive Facebook use and self-perceptions through negative social comparison is reduced among young adults who report greater levels of valuing stimulation: for b) self-perceived social competence, c) self-perceived physical attractiveness and d) self-perceived career success (H11).

Overall, I thus suppose that passive Facebook use will positively influence negative social comparison, which in turn will negatively influence self-perceptions. The negative indirect effect of passive Facebook use will be mediated by negative social comparison. The path from passive Facebook use to negative social comparison will be moderated by life satisfaction and personal values. Active Facebook use will positively influence self-perceptions. The conceptual model can be seen in Figure 1.

(15)

Method

Sample and procedure

An online survey was conducted in December 2015. The survey was provided in Dutch, Italian and Russian and was conducted in these three countries. The participants were recruited through the online social networks of the author. The recruitment yielded 264 respondents who completed the survey. 32 participants were excluded because they did not meet the inclusion criteria for age (30). The total sample then comprised 232 participants (76 % female) aged 18-30 (M = 26.1, SD = 3.0). The participants differed in their nationalities. The most common nationalities were Russian (48%), Italian (17%) and Dutch (16%). The remaining 19% included individuals with Ukrainian (6%), Belorussian (2%) and other former USSR nationalities, as well as some EU nationalities. The majority of the respondents (46%) was employed for wages, 29% were students, with the remaining 25% consisting of unemployed (10%), homemakers (7%), self-employed (5%) and other (3%).

Ethical approval was obtained from the University of Amsterdam. As one of the measures in this questionnaire (the values) required to be presented in the respondents native language, the initial English questionnaire was translated into identical questionnaires in Dutch, Italian and Russian. The online survey started with the first question in English, requesting the respondents to select their native language (Dutch, Italian or Russian). If the option “none of the above” was selected, the survey stopped for that respondent. The participants were provided with written explanation about the goal of the study in their own languages and asked for their consent. After this they responded to the measures in the following order: socio-demographical questions, self-perception measures (social competence, physical appearance, and career success), values, life satisfaction, Facebook use and negative social comparison. All the measures were randomized within the topics. The English version of the questionnaire can be found in Appendix A.

Pretest

The pretest was employed to make sure that the translations into Russian, Italian and Dutch were understood correctly. The participants for the pretest were similar to the target population (Hunt, Sparkman, & Wilcox, 1982) in age (30 or younger) and the fact that they used Facebook. They were also all highly educated, as highly educated pretesters can offer more

(16)

useful and articulated critique than less educated participants (Foddy, 1998). Two testers were employed for each language: a sociology/communications professional and a “regular”

participant. During the pretest respondents answered the online questionnaire, discussed the questions that seemed strange or awkward (both in meaning and wording) and commented on their general perception of the survey (including its overall appeal, esthetics and duration). The communication was facilitated mainly via online Facebook-chat. The feedback obtained from the respondents was overall positive and resulted in a couple of helpful specifications of translations of the measures into Dutch and Italian, which were successfully modified.

Measures

Facebook use. Facebook use was assessed using an adapted version of the Facebook use

scale by Frison and Eggermont (2015), which distinguished between passive and active Facebook uses. To asses passive Facebook use two additional items were added to the two original ones, amounting to four items in total. The original items asked about the frequency of visiting a friends’ Facebook profile and a profile of someone who does not belong to one’s Facebook friends list. The authors suggested (Frison & Eggermont, 2015) adding a passive Facebook use measure assessing news feed reading. Previous research supported the notion that news feed reading (Burke et al., 2010; Pempek, Yermolayeva, & Calvert, 2009) as well as the viewing of photos of others (Pempek et al., 2009) were highly popular passive Facebook use activities. Thus, the two new items “How often do you read/scroll through your Facebook feed?” and “How often do you look at photos of your friends on Facebook?” were added. The items that assessed

different types of Facebook use were rated on a 5-point scale. The scale ranged from “once a month or less” (=1) to “several times a day” (=5).

In the original research the items loaded on three scales: public active FB use, private active FB use and passive Facebook use. In this study three factors with an eigenvalue greater than 1 emerged, accounting for the cumulative variance of 55.9%. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the factor analysis, KMO = .73, which is well above the acceptable limit of .5 (Field, 2009). Principal axes analysis with oblique rotation showed that there were indeed three factors. The outcome of the rotation suggested by the pattern matrix showed a grouping of items (range .65 to .88 ) for passive Facebook use (visiting profiles of friends and strangers, looking at photos of friends and, to a lesser extent, watching the Facebook

(17)

feed), active public Facebook use (range .69 to .79 ) (posting messages, photos and pictures on one’s own timeline), and negative factor for active private Facebook use (using the chat (-. 90) and sending messages (-.90)).ii_{As the both the active Facebook use dimensions are expected to} have the same effect and as summarizing them into one measure of active Facebook use will increase the parsimony of the analyses, the decision was taken to make a common active Facebook use measure out of factors two and three, with five items in total. Thus two subscales were created, for passive (M = 2.7, SD = .9) and active (M = 2.1, SD = .8) Facebook use. The scores ranged between 1 and 5, with higher scores reflecting more intense Facebook use. The active Facebook use subscale thus included the questions about posting information on own timeline and engaging in personal communication. Cronbach’s alpha for this measure was .74. For the passive Facebook use, featuring the questions about watching other’s profiles, photos and the Facebook feed, the Cronbach’s alpha was .72.

Negative social comparison on Facebook. The negative social comparison measurement

(two items) was taken from de Vries & Kühne (2015). Following Krasnova et al. (2013) three items were added to the scale. The study of Krasnova et al. showed that travel and leisure account for most of the envy on Facebook, while posts about successes are among the top three most envied content on Facebook. All the items started with “When I read news feeds (or see photos of others) I often think that…”. The two original items read : “…others have better lives than me” and “…others are doing better than me”. The three new added items read as follows: “…others are more successful than me”, “…others are having more fun than me” and “…others are leading more exciting lives than me”. The resulting five items were averaged and formed into a scale. The scale ranged from 1 to 5 (M = 2.5, SD = .9). Cronbach's alpha was .93. Factor analysis revealed that all the five items loaded on one factor that explained 73.0 % of the variance. The items loadings ranged from .80 to .88. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the factor analysis, KMO = .87.

Life satisfaction. This construct was measured with the Satisfaction with Life Scale

(Diener, Emmons, Larsen, & Griffin, 1985). This scale has been widely used and has been shown to be valid and reliable (Pavot & Diener, 2008). The scale contains five items, such as “In

general, I have the feeling that I am satisfied with my life” rated on a 5-point scale, ranging from totally disagree (=1) to totally agree (=5). The average of the items was taken to construct a scale,

(18)

where higher scores indicate higher life satisfaction. Scores ranged from 1 to 5 (M = 3.5, SD = .8). Cronbach's alpha was .85. Factor analysis revealed that all the five items loaded (range: .64 to .84) on one factor that explained 55.3 % of the variance. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the factor analysis, KMO = .86

Values. The four values of Stimulation, Hedonism, Humility and Achievement were

assessed using the latest Schwartz Personal Values Questionnaire (PVQ-RR) (Cieciuch et al., 2014). The scale assesses 19 values and contains 57 questions in total. Each value is assessed by three questions. Although the current study required only the information about the four values (achievement, hedonism humility and stimulation), all 19 values were measured, as it was stressed by the author of the scale that the values formed a circular continuum and could only be correctly measured in relationship to each other (Cieciuch et al., 2014). The values were

examined by asking the respondents “How much like you is this person?”. When the respondent states that he or she is similar to the described person, it indicates that he or she shares this value. The answers ranged on a scale of 1–6: from “Not like me at all” to “Very much like me”.

Stimulation was measured with three items (e.g., “It is important to him to have all sorts of new experiences”). These three items were averaged to create a total scale (M = 4.4, SD = .8). Cronbach's alpha was .60. A factor analysis revealed that the three items loaded (range: .43 to .73) on one factor that explained 38.0 % of the variance. The Kaiser–Meyer–Olkin measure suggested that the sampling was adequate for the factor analysis, KMO = .62, which is above the acceptable limit of .5 (Field, 2009).

To measure the hedonism value such items as “It is important to her to have a good time” were used. The three items were averaged to create a total scale. Scores ranged from 2.7 to 6 (M = 4.7, SD=.8). Cronbach's alpha was .63. Factor analysis revealed that all three items loaded (range: .58 to .66) on one factor that explained 37.8% of the variance. The Kaiser–Meyer–Olkin measure suggested that the sampling was adequate for the factor analysis, KMO = .65.

The value of humility was measured with such items as “It is important to him to be humble”. The three items were averaged to create a total scale. Scores ranged from 1 to 6

(M=4.6, SD=0.9). Cronbach’s alpha was .69. Factor analysis revealed that all three items loaded (range: .40 to .71) on one factor that explained 35.5% of the variance. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the factor analysis, KMO = .61.

(19)

Achievement value was measured with such items as “It is important to her to have ambitions in life”. The three items were averaged to create a total scale. Scores ranged from 1 to 6 (M=4.6, SD=0.9). Cronbach’s alpha was .69. Factor analysis revealed that all three items loaded (range: .43 to .93) on one factor that explained 48.7% of the variance. The Kaiser–Meyer– Olkin measure verified the sampling adequacy for the factor analysis, KMO = .59.

Self-perception. Following De Vries and Kühne (2015), self-perceived social

competence and physical appearances were assessed using an adapted version (Valkenburg et al., 2006) of Harter's Self-Perception Profile for Adolescents (Harter, 1988). Participants were asked to indicate how they felt about themselves. Self-perceived social competence was measured using five items (e.g., “I am accepted by others”). Self-perceived physical attractiveness was measured with four items (e.g., “I am satisfied with my appearance.”).The third component of

self-perception, career success, was assessed with the Career Satisfaction Scale (Greenhaus,

Parasuraman & Wormley, 1990). An example item of the scale: “I am satisfied with the success I have achieved in my career”. The scale contained five items. The answer scale ranged from 1 (totally disagree) to 5 (totally agree). Items reflecting the three subscales were recoded where necessary (where there were reverse-scored items) and the average of the items was taken to create a score per subscale. Higher scores, in this way, reflected more positive self-perception.

A principal axes analysis with direct oblimin rotation was conducted on the 14 items of the three self-perception measures. The analysis supported the existence of the three dimensions of self-perception. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the factor analysis, KMO = .81, which is well above the acceptable limit of .5 (Field, 2009). The items about career satisfaction loaded on the first factor, which explained 29% of variance (loadings ranged from .44 to .93). The items about self-perceived physical appearance loaded on the second factor, which explained 12% of variance (loadings ranged from .66 to .91). The third factor explained 7% of variance and contained the items about self-perceived social competence (loadings ranged from .41 to .61). According to Stevens’ guidelines for a sample of 200

respondents an item loading should be greater than .36 (Field, 2009), therefore all the items were retained. The tables containing the scree plot and the outcomes for the pattern and structure matrix can be found in the Appendices D, E and F respectively. Mean indices for self-perceived

(20)

social competence (M = 3.5, SD = 0.6; α = .68), self- perceived physical attractiveness (M = 3.4, SD = .9; α = .83), and self-perceived career success (M = 3.3, SD = .8; α = .89) were computed.

Control variables. Because Facebook use, values, self-perception and life satisfaction

may depend on age, gender, culture (nationality) and type of employment, these variables were controlled for in later analyses.

Results

First bivariate correlations were checked between the key variables. iii Then the

hypothesized moderated mediation models were tested using the Process macro (Hayes, 2013) which estimates all paths using OLS estimation. In total twelve models were estimated. Each model included passive Facebook use as the independent variable, negative social comparison as mediator, two moderators of the relationship between passive Facebook use and negative social comparison and one of the self-perception measures (social competence, physical attractiveness or career success) as dependent variablesiv.

Life satisfaction was included in every model as the first moderator of the path from passive Facebook use to negative social comparison, as this role of life satisfaction was confirmed in a previous study (De Vries & Kühne, 2015). In each model, one of the relevant values –

achievement, hedonism, humility, or stimulation – was entered as the second moderator of the path from passive Facebook use to negative social comparison. The independent variable (passive Facebook use) and the moderators were mean-centered prior to the analysis so that the effects of the independent variable and the moderators could be interpreted as average effects (Hayes, 2013). Finally, age, gender, the three nationality dummies (comparison category: Russian), the three type of employment dummies (comparison category: working) and active Facebook use were included as covariates. Tests of conditional indirect effects, which are reported below, are based on bootstrapped and bias-corrected confidence intervals (5000 bootstrap samples).

H1 suggested that passive Facebook use positively predicts negative social comparison. Indeed, as can be seen in the upper part of Table 1, H1 is supported by the analyses (b = .21, p <.01), thus for every unit increase in passive Facebook use, negative social comparison grows by .21 on its 1 to 5 scale. v

(21)

H2 stated that the effect of active Facebook use on negative social comparison is weaker than the effect of passive Facebook use. As can be seen in the first part of Table 1, not only is the effect of active Facebook use on negative social comparison weaker, but it is also negative and

Table 1. Moderated mediation models: ordinary least squares regression model coefficients for models with passive Facebook use as predictor, life satisfaction and achievement value as moderators, negative social comparison as mediator and self-perceptions as outcome

(standard errors in parentheses).

Model 1 Model 2 Model 3

DV: Self-perceived social competence DV: Self-perceived physical appearance DV: Self-perceived career success

Predictors for the mediator Coef. b Coef. b Coef. b

Passive Facebook use .21 (.08)**

Life satisfaction -.36 (.07)***

Interaction:PFU x LS .16 (.08)*

Achievement -.03 (.07)

Interaction:PFU x ACH -.01 (.08)

Active Facebook use -.02 (.09)

Age -.03 (.02) Gender .21 (.14) Other nationality -.31 (.16) Dutch nationality -.21 (.18) Italian nationality .00 (.17) Student -.11 (.16) Not working .06 (.16)

Other type of employment .09 (.33)

Model Summary R2 = .23, F(14,217)=4.57, p<.001 Predictors for the

dependent variable

Negative social comparison -.21 (.04)*** -.28 (.05)*** -.25 (.06)***

Passive Facebook use -.03 (.05) .04 (.06) .06 (.07)

Active Facebook use .15 (.06)* .02 (.07) .18 (.08)*

Age .02 (.02) .00 (.02) -.02 (.02) Gender -.06 (.10) .08 (.12) .14 (.13) Other nationality .06 (.11) .06 (.13) .06 (.14) Dutch nationality .16 (.12) .03 (.14) .27 (.16) Italian nationality .04 (.12) -.11 (.14) .14 (.15) Student .08 (.11) -.12 (.13) -.37 (.14)* Not working .14 (.11) -.24 (.13) -.63 (.15)***

Other type of employment -.15 (.23) -.16 (.28) .29 (.31) Model Summary R2 _{= .16,} F(11,220)=3.88, p<.001 R2 _{= .15,} F(11,220)=3.47, p<.001 R2 _{= .21,} F(11,220)=5.30, p<.001 Note. N=232. PFU represents passive Facebook use, LS stands for life satisfaction, ACH represents the achievement value. Age, gender, nationality, type of employment and active Facebook use were entered in the models as covariates. All coefficients are unstandardized.* p <.05, **p<.01, ***p <.001 (two-tailed).

(22)

non-significant (b= -.02, SE= .09, p = 839) while the effect of passive Facebook use on negative social comparison is positive and significant (b=.21, SE= .08, p <.01). H2 is therefore supported. H3 suggested that negative social comparison will negatively predict self-perception. In the second part of Table 1 the outcomes of the three models of the three self-perception domains are presented. Model 1 with self-perceived social competence as outcome shows that negative social comparison negatively predicts social competence (b = -.21, p <.001), H3a is thus supported. This means that for every unit increase in negative social comparison on its 1 to 5 scale, there will be a .21 unit decrease in self-perceived social competence on its 1 to 5 scale. Model 2 with self-perceived physical appearance as outcome shows that self-perceived physical appearance is also negatively predicted by negative social comparison (b = -.28, p <.001). Therefore, H3b is supported. For every unit increase in negative social comparison on its 1 to 5 scale, there will be a .28 unit decrease in self-perceived physical appearance on its 1 to 5 scale. Finally, Model 3 shows that negative social comparison negatively predicts self-perceived career success (b = -.25, p <.001), H3c is thus supported as well. For every unit increase in negative social comparison on its 1 to 5 scale, there will be a .25 unit decrease in self-perceived career success on its 1 to 5 scale. vi

H4a-c suggests that there’s a negative indirect effect of passive Facebook use on self-perceptions through negative social comparison. The significant effect of passive Facebook use on negative social comparison (H1) and the significant effect of negative social comparison on self-perceptions (H3) indicate that a mediation process is at work here. The indirect effects were tested using the bootstrap procedure to estimate 95% bias-corrected confidence intervals (95% bc CI)vii_{. For the effect to be significant, the confidence intervals should not cross zero.There is a} negative indirect effect of passive Facebook use on self-perceived social competence through negative social comparison that accounts for -.04 (SE=.02; 95% bc CI: -0.10 to -0.01). In predicting self-perceived physical attractiveness the mediation effect of passive Facebook use through negative social comparison accounts for -.06 (SE=.03; 95% bc CI:-0.12 to -0.02 ). In self-perceived career success the mediation accounts for -.05 (SE=.03, 95% bc CI:-0.12 to -0.01) viii_{. In this way, the results are significant and the hypotheses H4a-c are therefore supported.}

(23)

Finally, the (residual) direct effect of passive Facebook use on self-perception was inspected, the outcomes can be seen in Table 1. None of the three self-perception domains was directly

significantly affected by passive Facebook use, as can be seen in Models 1-3 respectively: self-Table 2. Moderated mediation models: ordinary least squares regression model

coefficients for models with passive Facebook use as predictor, life satisfaction and humility value as moderators, negative social comparison as mediator and self-perceptions as outcome (standard errors in parentheses).

Interaction: PFU x LS .18 (.08)*

Humility -.02 (.06)

Interaction: PFU x HU -.08 (.06)

Age -.03 (.02) Gender .21 (.14) Other nationality -.32 (.15)* Dutch nationality -.18 (.17) Italian nationality .05 (.18) Student -.08 (.16) Not working .09 (.16)

dependent variable

Other type of employment -.15 (.23) -.16 (.28) .29 (.31) Model Summary R2 = .16, F(11,220)=3.88, p<.001 R2 = .15, F(11,220)=3.47, p<.001 R2 = .21, F(11,220)=5.30 , p<.001 Note. N=232. . PFU represents passive Facebook use, LS stands for life satisfaction, HU represents the humility value. Age, gender, nationality, type of employment and active Facebook use were entered in the models as covariates .All coefficients are unstandardized.* p <.05, **p<.01, ***p <.001 (two-tailed).

(24)

perceived social competence (b = -.03, p=.570), self-perceived physical attractiveness (b = .04, p=.499) and self-perceived career success (b = .06, p=.374) are not significantly predicted by passive Facebook use.

H5a-c suggested that active Facebook use will positively predict self-perceptions of social competence, physical appearence and career success. Table 1 shows, that while that is true for self-perceived social competence (b = .15, p <.05) in Model 1 and self-perceived career success (b = .18, p <.05) in Model 3, the effect does not hold for self-perceived physical appearance (b = .02, p =.839). With every one unit increase in active Facebook use (on its 1 to 5 scale), self-perceived social competence will grow by .15 on its scale and self-self-perceived career success will grow by .18. But, no matter how intensively one will actively use Facebook, it will not have an influence on their self-perceived physical appearance. ix Thus, H4a and H4c are supported, while H4b is rejected.

H6 suggests that the positive relationship between passive Facebook use and negative social comparison on Facebook is moderated by life satisfaction. The weakening of this relationship was expected for emerging adults who are more satisfied with their life. As can be seen in Table 1, the interaction of passive Facebook use with life satisfaction is significant (b = .16, p <.05). The direction, however, is the opposite of the hypothesized: it seems that life

satisfaction is making the effect of passive Facebook use on negative social comparison stronger, rather than weakening it. Therefore H6 is rejected. This unexpected finding will be addressed in the discussion.

H7 suggested that the indirect negative relationship between passive Facebook use and self-perceptions through negative social comparison is reduced among emerging adults who report greater satisfaction with life. As life satisfaction showed a significant interaction with passive Facebook use on negative social comparison (H6), and negative social comparison has been shown to be a significant mediator of passive Facebook use (H5), the moderated mediation effects could be tested. The test for moderated mediation shows that when the values of life satisfaction are one standard deviation below the mean (e.g. at low levels of life satisfaction), the effect of negative social comparison on self-perception is not significant.x This is true for social competence (b = -.02, SE=.03; 95% bc CI:-0.07 to 0.03), physical appearance (b = -.02, SE=.03; 95% bc CI:-0.09 to 0.04) and career success (b = -.02, SE=.03; 95% bc CI:-0.09 to 0.04).

(25)

However, when life satisfaction is at its mean or one standard deviation above the mean (e.g. at average and high levels of life satisfaction), negative social comparison has a significant negative effect on self-perceptions and this effect grows, as life satisfaction grows.xi This effect

Table 3. Moderated mediation models: ordinary least squares regression model coefficients for models with passive Facebook use as predictor, life satisfaction and hedonism value as moderators, negative social comparison as mediator and self-perceptions as outcome (standard errors in parentheses).

Hedonism -.08 (.07)

Interaction: PFU x HE -.08 (.08)

dependent variable

Other type of employment -.15 (.23) -.16 (.28) .29 (.31) Model Summary R2 = .16, F(11,220)=3.88, p<.001 R2 = .15, F(11,220)=3.47, p<.001 R2 = .21, F(11,220)=5.30 , p<.001 Note. N=232. . PFU represents passive Facebook use, LS stands for life satisfaction, HE represents the hedonism value. Age, gender, nationality, type of employment and active Facebook use were entered in the models as covariates. All coefficients are unstandardized.* p <.05, **p<.01, ***p <.001 (two-tailed).

(26)

holds for social competence when life satisfaction is at its mean (b = -.04, SE=.02; 95% bc CI:-0.10 to -0.01) and when life satisfaction is one standard deviation above the mean (b = -.07, SE=.04; 95% bc CI:-0.16 to -0.02). It is also the case for physical appearance when life

satisfaction is at its mean (b = -.06, SE=.03; 95% bc CI:-0.12 to -0.02) and when life satisfaction is one standard deviation above the mean (b = -.10, SE=.04; 95% bc CI:-0.20 to -0.03). In this way, the data showed that life satisfaction moderates the path from passive Facebook use through negative social comparison to self-perceptions, but not in the expected direction: the mediation effect of negative social comparison on self-perceptions becomes stronger with higher life satisfaction, instead of getting weaker. H7 is thus rejected.

H8a suggested that the relationship between passive Facebook use and negative social comparison on Facebook is moderated by the value of achievement and that the relationship between passive Facebook use and negative social comparison would be stronger for those high in achievement value. Table 1 shows that the interaction of achievement value with passive Facebook use is non-significant (b = -.01, p =.927). H8a is therefore rejected. This precludes testing the moderating role of the achievement value in the mediation model (H8b-c).

H9a suggested that the relationship between passive Facebook use and negative social comparison would be weaker for those who value humility more. Table 2 shows that the interaction of humility value with passive Facebook use is non-significant (b = -.08, p =.226). H9a is therefore rejected. This precludes testing the moderating role of the humility value in the mediation model (H9b-c).

H10a suggested that the relationship between passive Facebook use and negative social comparison would be stronger for those high in hedonism value. Table 3 shows that the

interaction of hedonism value with passive Facebook use is non-significant (b = -.08, p =.294). H9a is therefore rejected. This precludes testing the moderating role of the hedonism value in the mediation model (H9b-c).

H11a suggested that the relationship between passive Facebook use and negative social comparison would be weaker for those high in stimulation value. Table 4 shows that the

interaction of stimulation value with passive Facebook use is non-significant (b = -.01, p =.924). H11a is therefore rejected. This precludes testing the moderating role of the stimulation value in the mediation model (H11b-c). Personal values do not moderate the process between passive

(27)

Facebook use and negative social comparison. This precludes testing their moderating role in the mediation model.

Table 4. Moderated mediation models: ordinary least squares regression model coefficients for models with passive Facebook use as predictor, life satisfaction and stimulation value as moderators, negative social comparison as mediator and self-perceptions as outcome (standard errors in parentheses).

Stimulation -.07 (.06)

Interaction: PFU x STI -.01 (.07)

dependent variable

Other type of employment -.15 (.23) -.16 (.28) .29 (.31) Model Summary R2 = .16, F(11,220)=3.88, p<.001 R2 = .15, F(11,220)=3.47, p<.001 R2 = .21, F(11,220)=5.3 0, p<.001 Note. N=232. PFU represents passive Facebook use, LS stands for life satisfaction, STI represents the stimulation value. Age, gender, nationality, type of employment and active Facebook use were entered in the models as covariates All coefficients are unstandardized.* p <.05, **p<.01, ***p <.001 (two-tailed).

(28)

Some of the control variables showed an influence on the examined outcomes. As can be seen in Table 2, the “other nationality” category, as compared to Russian nationality had (b= -.32, p <.05), meaning that the “other nationality” group experienced .32 less negative social

comparison on its 1-5 scale than the Russian group. Additionally, not surprisingly, self-perceived career successxii_{was influenced by the employment situation of the individual. Compared to the} “working” category, students overall rated their career success lower (b= -.37, p <.05) by .37 points on the 1 to 5 scale, while those not working were even more negative (b= -.63, p <.05) and perceived their career success as .63 points lower on the 1 to 5 scale. These outcomes can be seen in Tables 1 to 4.

Discussion

This study demonstrated that the negative influence of passive Facebook use on self-perceived social competence, physical attractiveness and career success was mediated by negative social comparison. It also showed that active Facebook use had a direct significant positive effect on self-perceived social competence and career success, but not on self-perceived physical appearance. Life satisfaction was found to have an unexpected negative moderating effect on the positive relationship between passive Facebook use and negative social comparison. Personal values of achievement, hedonism, humility and stimulation did not show any significant moderating influence on this relationship.

The results of this study can offer at least three important contributions that can be used to guide future research. First, this study demonstrated that passive and active Facebook use have differential outcomes regarding self-perceptions. A positive association was detected between passive Facebook use and negative outcomes: a direct positive association with negative social comparison and an indirect one with more negative self-perceptions. This is in line with previous research linking passive Facebook use and negative psychological outcomes (e.g. Burke et al., 2010; De Vries & Kühne, 2015; Frison & Eggermont, 2015; Krasnova et al., 2013; Verduyn et al., 2015). These findings are especially meaningful today, since recent reports suggest that passive Facebook use is on the rise globally while active Facebook use is declining (Olson, 2015). This situation thus poses a danger to the users’ wellbeing. However, there is room for hope and possibly a way out. Recent studies on the negative effects of passive Facebook use (including this one) have shown that negative psychological responses mediate the relationship

(29)

between passive Facebook use and negative outcomes (e.g. envy in Krasnova et al., 2013 and Verduyn et al., 2015 and negative social comparison in the present study). Research should further investigate the ways in which these negative psychological effects could be counteracted by employing psychological strategies. For instance, one strategy is simply becoming aware of the negative emotions one experiences – by doing this one can already regulate those emotions (Herwig, Kaffenberger, Jäncke, & Brühl, 2010). In terms of practice this could mean that Facebook could post useful tips in the user’s feed. Additionally, as both negative social comparison (De Vries & Kühne; this study) and envy (Krasnova et al., 2013; Verduyn et al., 2015) have been shown to mediate the relationship from Facebook use to negative outcomes and since the both constructs are interconnected (Appel, Gerlach, & Crusius, 2016; Smith 2004; Smith & Kim, 2007) it would be interesting to examine how they compare to each other in strength of mediation.

This study also shows that active Facebook use is not associated with negative social comparison, but is a direct significant positive predictor of self-perceived social competence and self-perceived career success. Active Facebook use had no influence on self-perceived physical appearance. The finding of the beneficial effect of active Facebook use on psychological

outcomes is in line with previous research (e.g. Burke et al., 2010; Koroleva et al., 2011; Deters & Mehl, 2013). One explanation for the positive role of active Facebook use in the influence on self-perceptions may be just the sole act of disclosing information about the self (e.g. sharing ones thoughts and feelings through updating ones status) which has been found to have a positive effect on well-being (Ko & Kuo, 2009). Another explanation may be the positive feedback that the active users are getting on the posted content. Empirical data suggests that updating about ones achievements, social activities and everyday life predicted getting more likes (Marshall, Lefringhausen, & Ferenczi, 2015). In turn, not surprisingly, positive feedback on social media enhances self-esteem (Valkenburg, Peter, & Schouten, 2006). An alternative explanation for the positive association of active Facebook use on self-perception would be that narcissists tend to actively use Facebook (update their Facebook status) more frequently (Ong et al., 2011) while at the same time they are known for positively distorting their self-perceptions, especially in the domains of likability, attractiveness and leadership (Grijalva & Zhang, 2015). Future research might benefit from further delineating and measuring the distinct active Facebook use activities and linking them to psychological outcomes, as well as investigating which kind of content