Unit 3 Project-Is Facebook Bad for College Students

Negative Social Comparison on Facebook and Depressive Symptoms: Rumination as a Mechanism

Brian A. Feinstein, Rachel Hershenberg, Vickie Bhatia, Jessica A. Latack, Nathalie Meuwly, and Joanne Davila

Stony Brook University

Recent research demonstrates that it is the quality rather than the frequency of social networking experiences that places individuals at risk for negative mental health outcomes. However, the mechanisms that account for this association have yet to be examined. Accordingly, this study examined whether the tendency to negatively compare oneself with others while using Facebook leads to increases in depressive symptoms, and whether this association is mediated by increases in rumination. A sample of 268 college-age young adults completed an initial online survey and a 3-week follow-up. Path analysis was used to test the hypothesized model, wherein negative social comparison on Facebook was predicted to be associated with increases in rumination, which, in turn, was predicted to be associated with depressive symptoms. The model controlled for general social comparison to test the specific effect of social comparison on Facebook over and above the tendency to engage in social comparison in general. Results indicated that the hypothesized mediation effect was significant. In sum, in the context of social networking, negatively comparing oneself with others may place individuals at risk for rumination and, in turn, depressive symptoms. Findings increase understanding of the mechanisms that link social networking use to negative mental health outcomes and suggest a continued emphasis on examining the specific processes that take place in the context of social networking that may be pathogenic.

Keywords: Facebook, social comparison, social networking, depression, rumination

As social networking sites such as Face- book have become more popular, researchers have become increasingly interested in under- standing the potential consequences of their use. Initially, it was suggested that Internet use was associated with negative mental health outcomes such as depressive symptoms. However, as data accumulated, the results became increasingly mixed; some studies supported the associations between Internet use and mental health problems (Kraut et al., 1998;

Selfhout, Branje, Delsing, ter Bogt, & Meeus, 2009; van den Eijnden, Meerkerk, Vermulst, Spijkerman, & Engels, 2008), whereas others demonstrated positive effects of Internet use (Bessière, Kiesler, Kraut, & Boneva, 2008; Morgan & Cotten, 2003; Valkenburg & Peter, 2007). To reconcile discrepant findings, Da- vila et al. (2012) examined the amount of time spent engaging in social networking activities as well as the quality of interactions people had while using social networking mediums. Consistent with the larger literature on mood and anxiety disorders and interpersonal func- tioning, the authors hypothesized that poorer quality interactions, rather than use alone, would be associated with negative mental health out- comes. As predicted, in cross-sectional and 3-week prospective analyses, more negative and less positive self-reported social networking inter- actions were associated with depressive symp- toms, whereas time spent engaging in social net- working was not (Davila et al., 2012).

Brian A. Feinstein, Rachel Hershenberg, Vickie Bhatia, Jessica A. Latack, Nathalie Meuwly, and Joanne Davila, Department of Psychology, Stony Brook University.

This research is supported, in part, by a National Science Foundation Graduate Research Fellowship awarded to Brian A. Feinstein.

Correspondence concerning this article should be ad- dressed to Brian A. Feinstein, Department of Psychology, Stony Brook University, Stony Brook, NY 11794-2500. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Psychology of Popular Media Culture © 2013 American Psychological Association 2013, Vol. 2, No. 3, 161–170 2160-4134/13/$12.00 DOI: 10.1037/a0033111

161

As Facebook use becomes virtually ubiqui- tous, it is important to continue to identify the specific behaviors and processes that may be “risky.” Previous research on Facebook use has garnered a great deal of media attention, partic- ularly a clinical report claiming that researchers had documented a phenomenon called “Face- book depression,” or depression that results from spending too much time on Facebook (O’Keeffe & Clarke-Pearson, 2011). In fact, no research supports this claim, and, although scholars have attempted to clarify this (Davila, 2011; Magid, 2011), such false claims empha- size the importance of testing hypothesis-driven research questions that shed light on specific mechanisms that may lead to poorer well-being in the context of social networking. As noted, Davila et al. (2012) suggest that it is the quality rather than the frequency of social networking experiences that predicts negative mental health outcomes, but it remains unclear what specifi- cally takes place on social networking sites be- yond poor quality interactions that may be pathogenic.

Social networking sites provide venues for people to engage in a variety of behaviors, such as actively interacting with others (e.g., instant messaging), passively interacting with others (e.g., posting a message on someone’s profile), and obtaining information about others (e.g., looking at someone’s profiles). Similar to tradi- tional social activities, these online activities provide individuals with ample opportunity to com- pare themselves with others on numerous charac- teristics such as appearance, popularity, and success. When people are presented with infor- mation about others, they tend to relate that information to themselves (Mussweiler, Ruter, & Epstude, 2006), and this social comparison provides them with self-evaluative information that can be used to make positive or negative self-judgments (Festinger, 1954). For instance, if a person sees that many of his or her friends are getting jobs and he or she is unemployed and having a difficult time getting a job, then he or she might feel inadequate in that domain. In contrast, if a person sees that many of his or her friends are unemployed and he or she has just gotten a job, then he or she might feel espe- cially adequate in that domain. Importantly, social comparison is a pervasive and auto- matic feature of relating to others on an indi- vidual and group level (Pratto, Sidanius,

Stallworth, & Malle, 1994; Sidanius, Pratto, & Bobo, 1994; Wood, 1989), making it nearly impossible to circumvent.

Social comparison is not in itself problem- atic, and indeed, even negative1 social compar- isons (i.e., comparisons with others who are perceived as superior) can have positive effects (e.g., self-improvement and self-enhancement; Wood, 1989). Still, negative social comparison can also maintain or exacerbate negative self- appraisals (for reviews, see Ahrens & Alloy, 1997; Suls & Wheeler, 2000; Swallow & Kuiper, 1988; Wood & Lockwood, 1999). No- tably, individuals report increases in negative affect subsequent to negative social compari- sons (Ahrens & Alloy, 1997; Antony, Rowa, Liss, Swallow, & Swinson, 2005; Giordano, Wood, & Michela, 2000; Wheeler & Miyake, 1992). Despite evidence that negative social comparison may lead to poorer well-being in general, little research exists on this process in the context of social networking. As an excep- tion, a qualitative study found that users of MySpace, another social networking site, re- ported engaging in social comparison on the site and reported negative self-views subsequent to such comparison (Manago, Graham, Greenfield, & Salimkhan, 2008). Chou and Edge (2012) also found that individuals who spent more time on Facebook were more likely to agree that others were “happier” and “had better lives.” Finally, Haferkamp and Krämer (2011) found that individuals who looked at profile pictures of attractive members of the same-sex reported less positive affect than those who looked at nonattractive members of the same-sex. This emerging body of research provides preliminary support for a social networking site such as Facebook to provide a context for engaging in negative social comparison, which may be as- sociated with negative consequences.

In addition to examining the relationship be- tween specific processes taking place in the

1 Several different terms have been used throughout the literature to describe the process wherein an individual compares himself or herself with others who are perceived as superior. Although most studies refer to this process as upward social comparison, others have referred to it as downward social comparison or negative social compari- son. We chose to use the term negative social comparison throughout this article, as it seems to be the least ambigu- ous.

162 FEINSTEIN ET AL.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

context of social networking and mental health outcomes, it is important to understand the mechanisms that may account for these associ- ations. One possible factor that may account for the association between negative social compar- ison on Facebook and depressive symptoms is the tendency to engage in maladaptive emotion regulation strategies, such as rumination. Rumi- nation refers to repetitively focusing on one’s distress, including its potential causes and con- sequences (Nolen-Hoeksema, Wisco, & Ly- ubomirsky, 2008), and has been consistently linked to negative mental health outcomes, in- cluding depression (for a meta-analytic review, see Aldao, Nolen-Hoeksema, & Schweizer, 2010). Although there has been little research on social comparison and emotion regulation, one study found that social comparison was positively associated with rumination (Cheung, Gilbert, & Irons, 2004). Building on these cross-sectional findings, the present study tests the hypothesis that negative social comparison on Facebook will lead to increases in the use of rumination as an emotion regulation strategy. Given that negative social comparison may be associated with rumination, and that rumination has consistently been linked to depressive symptoms, it is likely that rumination may act as a mechanism through which negative social comparison increases depressive symptoms. Notably, Locatelli, Kluwe, and Bryant (2012) found that rumination mediated the association between negative status updates on Facebook and depressive symptoms. Although their study provides preliminary support for our meditation hypothesis, the current study extends this to social comparison and uses a prospective design to test changes over time.

In sum, to refine our understanding of the processes that render individuals at risk when using social networking sites, the current study used a 3-week prospective design to examine a mediation model, wherein it was hypothesized that negative social comparison on Facebook would be associated with increases in depres- sive symptoms 3 weeks later through its asso- ciation with increases in rumination. We con- trolled for general social comparison tendencies (i.e., social comparison not specific to Face- book) to examine the specific effects of Face- book social comparison. Additionally, given that there are some data to suggest bidirectional associations between rumination and depressive

symptoms (Nolen-Hoeksema, Larson, & Gray- son, 1999; Nolen-Hoeksema, Stice, Wade, & Bohon, 2007; for an exception, see McLaugh- lin, Hatzenbuehler, Mennin, & Nolen-Hoek- sema, 2011), we examined an alternate media- tion model with depressive symptoms as the mediator and rumination as the outcome. By examining both models, we hoped to gain a better sense of the direction of the proposed mediation effect. The current study extended previous research in several ways, including examining a specific process that may put indi- viduals at risk for depressive symptoms in the context of social networking, using path analy- sis to test mediation models to better understand mechanisms of action, and controlling for gen- eral social comparison tendencies to provide a more stringent test.

Method

Participants

Eligible participants were at least 18 years old and enrolled in a psychology course for which they earned research credits for study participation. The present study was posted on the online system used by the psychology department to advertise research projects and track student credit. Participants included 105 male and 181 female students from Stony Brook University. Nine participants did not complete the follow-up assessment (retention rate � 96%), and nine additional participants did not complete one or more of the measures. These 18 participants were excluded from analyses; they did not significantly differ from the rest of the sample on gender, race/ ethnicity, age, Facebook social comparison, general social comparison, rumination, or de- pressive symptoms (ps range from .07 to .68). The final sample included 268 individuals (62% female), with an average age of 19.66 years (SD � 2.29) and a racial/ethnic distri- bution including Caucasian (40%), Asian (42%), Latino/a (5%), African American (4%), Middle-Eastern (3%), and other (6%).

Procedure

Participants completed an online survey (Time 1; T1) consisting of questionnaires as- sessing social comparison (in general and spe-

163NEGATIVE SOCIAL COMPARISON

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

cific to Facebook), rumination, and depressive symptoms. To assess change over time, a fol- low-up online survey was conducted 3 weeks later (Time 2; T2). Given that the surveys were administered online, respondents were able to participate from any location that had Internet access. Participants received course credit for their participation. This research was approved by the Stony Brook University Committee on Research Involving Human Subjects.

Measures

Social comparison on Facebook. The ten- dency to engage in social comparison when using Facebook was assessed with the Social Comparison Rating Scale (SCR; Allan & Gil- bert, 1995), an 11-item self-report measure that presents respondents with an incomplete sen- tence followed by a series of bipolar constructs. We modified the instructions so that partici- pants endorsed items based on social compari- son while using Facebook. Specifically, whereas the original scale began using the stem, “In relationship to others I generally feel. . .” we used the stem, “When I compare myself to others on Facebook, I feel. . .” Participants se- lected a number from 1 to 10 that best described their perceived position between two poles (e.g., inferior/superior, incompetent/more com- petent, unlikeable/more likable, undesirable/ more desirable). We reverse coded and summed all responses to compute a total score in which higher values indicated more negative self- perceptions compared with others. Possible to- tal scores could range from 11 to 110, and in our sample, ranged from 13 to 110 at T1. Allan and Gilbert (1995) reported internal consistency (al- pha) of .88 and significant correlations with measures of psychopathology, including de- pression. In the current sample, the alpha was .94 at T1.

General social comparison. The general tendency to engage in social comparison was assessed with the Iowa-Netherlands Compari- son Orientation Measure (INCOM; Gibbons & Buunk, 1999), which is an 11-item self-report measure. Participants were instructed to rate their degree of agreement with each statement, ranging from 1 (I disagree strongly) to 5 (I agree strongly). Sample items include, “I often com- pare how my loved ones are doing with how others are doing,” “If I want to learn more about

something, I try to find out what others think about it,” and “If I want to find out how well I have done something, I compare what I have done with how others have done.” Responses to all items were summed to create a total score, where higher scores reflected a greater tendency to compare oneself with others. Possible total scores could range from 11 to 55, and our sample had a large distribution (15–54 at T1). Gibbons and Buunk (1999) demonstrated inter- nal consistencies (alphas) ranging from .78 to .84, 3 to 4 week test–retest reliability of .71, and significant convergent and divergent validity across numerous samples. The alpha was .85 at T1 in the current sample.

Rumination. Rumination was assessed with the Ruminative Responses Scale (RRS; Nolen-Hoeksema & Morrow, 1991). The RRS is a 22-item self-report measure that assesses how frequently individuals experience or en- gage in various thoughts, feelings, and actions during a depressed mood. Sample items include “Think about how sad I feel” and “Analyze recent events to try to understand why I am depressed.” Each item is rated on a 1 (almost never) to 4 (almost always) scale, and responses are summed to compute a total score, which could range from 22 to 88, with higher scores representing higher levels of rumination. At both time points in our sample, we had a full range of scores. Excellent internal consistency and a significant correlation with depressive symptom severity have been reported (Nolen- Hoeksema & Morrow, 1991). In this sample, the alpha was .94 at T1 and .96 at T2.

Depressive symptoms. Depressive symp- toms were assessed with the Center for Epide- miological Studies-Depression Scale (CES-D; Radloff, 1977). The CES-D was specifically designed for use with community samples and included 20 items assessing past week experi- ence of depressive symptoms. Responses ranged from 0 (rarely or none of the time) to 3 (most or all of the time), with greater scores indicating greater depressive symptoms. Sam- ple items included, “I was happy” (reverse scored), “I felt that I could not shake off the blues even with help from my family and friends,” and “I felt hopeful about the future” (reverse scored). Construct validity, internal re- liability, and other psychometric strengths of the CES-D have been widely supported (Rad- loff, 1977). Possible total scores could range

164 FEINSTEIN ET AL.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

from 0 – 60. In this sample, scores ranged from 1 to 44 at T1 and 0 to 51 at T2. The alpha was .88 at T1 and .89 at T2.

Results

Table 1 presents the means, standard devia- tions, and zero-order correlations for all of the variables. As hypothesized, Facebook social comparison was significantly and positively as- sociated with general social comparison, and both were significantly and positively associ- ated with rumination and depressive symptoms at both time points. Rumination was also sig- nificantly and positively associated with depres- sive symptoms at both time points. Notably, the correlation between Facebook social com- parison and general social comparison was modest (r � .18), suggesting that they are distinct constructs.

Path analysis with measured variables was conducted using IBM SPSS Amos Version 20 to examine the primary hypothesized mediation model (see Figure 1). Path analysis has the advantage of being able to simultaneously test the associations among multiple predictor and outcome variables. The hypothesized model proposed that Facebook social comparison at T1 would lead to increases in depressive symp- toms at T2 through increases in rumination at T2. Given the prospective design, we controlled for rumination and depressive symptoms at T1. Additionally, general social comparison was controlled for to test the specific effect of Fa- cebook social comparison. To test the signifi- cance of the hypothesized indirect effect of Facebook social comparison on depressive symptoms mediated through rumination, we conducted bootstrapping analyses to estimate bias-corrected confidence intervals (cf. Mackin- non, Lockwood, & Williams, 2004). Less than 1% of the data from the final sample was miss- ing, and it was handled by imputing a partici- pant’s mean score on a measure in place of a missing value on one of the measure’s items. All paths included in the model were estimated freely, and two paths were not included, thus they were set to 0 (the path from T1 rumination to T2 depressive symptoms and the path from T1 depressive symptoms to T2 rumination). Model fit was assessed by the comparative fit index (CFI), the Tucker–Lewis index (TLI), and the root-mean-square error of approximation

(RMSEA), with acceptable model fit indicated by a CFI and TLI � .90 and an RMSEA � .06 (Hu & Bentler, 1999; Kline, 2005).

Results indicated that the hypothesized model fit the data very well, �2(2, N � 268) � 1.78, p � .41, CFI � 1.00, TLI � 1.00, and RMSEA � .00 (90% CI � [.00 –.12]). The model demonstrated that Facebook social compari- son was significantly associated with increases in rumination, which, in turn, were significantly associated with depressive symptoms2 (see Ta- ble 2). Bootstrapping analyses supported the significance of the indirect effect of Facebook social comparison on increases in depressive symptoms through increases in rumination, � � .05, bias-corrected 90% CI � [.02–.09], SE � .02. The direct effect of Facebook social com- parison on increases in depressive symptoms was not significant in the model, but it was significant when the mediator (rumination) was not included in the model (standardized path coefficient � .11, p � .03). Together, these findings provide support for the hypothesized mediation effect. Further, the significance of the indirect effect of general social comparison on depressive symptoms through rumination was not significant, � � .004, bias-corrected 90% CI � [�.02 to .03], SE � .01, suggesting that this effect may be specific to negative social comparison on Facebook.

Given that the mediator (rumination) and the outcome (depressive symptoms) were both measured at the T2 assessment, we also exam- ined an alternate model with depressive symp- toms as the mediator and rumination as the outcome. Results indicated that the alternate model did not fit the data well, �2(2, N � 268) � 9.36, p � .01, CFI � .99, TLI � .90, and RMSEA � .12 (90% CI � [.05–.20]). Although bootstrapping analyses indicated that the indi- rect effect of Facebook social comparison on

2 Given that rumination and depressive symptoms were both measured at the T2 assessment, it is worth noting that, in a separate regression analysis, rumination at T1 was associated with depressive symptoms at T2, controlling for depressive symptoms at T1 (� � .14, p � .01). Thus, even though the model included a path from rumination at T2 to depressive symptoms at T2, it is likely that rumination does, in fact, predict increases in depressive symptoms over time (as demonstrated by our data and previous research).

165NEGATIVE SOCIAL COMPARISON

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

increases in rumination through increases in depressive symptoms was significant, � � .03, bias-corrected 90% CI � [.01–.06], SE � .02, this should be interpreted with caution be- cause of the poor model fit. In sum, results provide stronger support for our primary model than our alternate model, suggesting that rumination mediates the association be- tween negative social comparison on Face- book and depressive symptoms.

Discussion

As the use of Facebook becomes increasingly commonplace, it is important to identify the behaviors and processes that may place users at risk for negative consequences. Given the in- herently social nature of Facebook, we tested the hypothesis that negatively comparing one-

self with others on Facebook leads to increases in depressive symptoms. Further, to better un- derstand the mechanism underlying this associ- ation, we tested the hypothesis that negative social comparison would lead to increases in rumination, and this passive and repetitive focus on distress would in turn be associated with depressive symptoms. We review key findings below.

First, Facebook social comparison, rumina- tion, and depressive symptoms were all posi- tively and significantly associated with one an- other. As noted, the correlation between general social comparison and Facebook social compar- ison was modest at T1 and nonsignificant at T2, suggesting that these are distinct constructs. When we included all of the variables in the proposed path model, we found a strong fit to the data. Specifically, controlling for the general

Table 1 Means, Standard Deviations, and Zero-Order Correlations

Variable 1 2 3 4 5 6 Mean (SD)

1. T1 Facebook social comparison — — — — — — 57.38 (16.95) 2. T1 General social comparison .18� — — — — — 35.68 (7.21) 3. T1 Rumination .42�� .37�� — — — — 49.63 (14.82) 4. T1 Depressive symptoms .38�� .19� .57�� — — — 12.89 (8.50) 5. T2 Rumination .42�� .26�� .66�� .43�� — — 48.07 (15.23) 6. T2 Depressive symptoms .35�� .16� .49�� .69�� .54�� — 13.07 (8.60)

� p � .01. �� p � .001.

Rumination at T1

Depressive

symptoms at T1

Facebook social

comparison at T1

General social

comparison at T1

Rumination at T2

Depressive

symptoms at T2

Figure 1. Path model depicting the significant indirect effect of Facebook social comparison on increases in depressive symptoms through increases in rumination. Note. Correlations among all T1 variables are included in the model, but not depicted for parsimony; paths representing the mediation component of the model are emphasized in bold.

166 FEINSTEIN ET AL.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

tendency to engage in social comparison, neg- atively comparing oneself with others while us- ing Facebook predicted increases in rumination, which in turn was associated with increases in depressive symptoms. The significant indirect effect of Facebook social comparison on de- pressive symptoms through rumination coupled with the nonsignificant direct effect with the mediator in the model provides preliminary support for the possibility that rumination does indeed mediate the association in question. Fur- ther, the nonsignificant indirect effect of general social comparison on depressive symptoms through rumination suggests that this effect may be specific to engaging in negative social com- parison on Facebook rather than the general tendency to compare oneself with others. Nota- bly, when we tested an alternate model with depressive symptoms as the mediator and rumi- nation as the outcome, the model fit was poor. This provides additional support for the pro- posed direction of the mediation effect in our primary model, and it is consistent with previ- ous research that failed to find a significant association between depressive symptoms and increases in emotion dysregulation over time (McLaughlin et al., 2011). However, other stud- ies have demonstrated that depressive symp- toms are associated with increases in rumination over time (Nolen-Hoeksema et al., 1999, 2007). Thus, although findings suggest that rumination may be a mechanism underlying the association between negative social comparison on Face- book and depressive symptoms, it will be im- portant for future research to collect at least three time points and include the prospective association between the mediator and outcome.

Findings are consistent with and expand pre- vious research in several meaningful ways. Consistent with findings that link negative so- cial comparison with increases in negative af- fect (Ahrens & Alloy, 1997; Antony et al., 2005; Giordano et al., 2000; Wheeler & Mi- yake, 1992), we found support for a significant association between Facebook social compari- son and depressive symptoms in the zero-order correlations. This suggests that social network- ing sites can provide novel opportunities for individuals to compare themselves with others, and these comparisons can have negative influ- ences on well-being. Further, this provides in- sight into what specifically may be happening on social networking sites that has the potential to be pathogenic—namely, comparing oneself with others who are perceived as superior. We also found that negative social comparison on Facebook predicted increases in rumination at a 3-week follow-up, which extends previous cross-sectional findings (Cheung et al., 2004). Why might this be the case? Some evidence suggests that individuals tend to self-disclose more positive information about themselves on Facebook compared with “real life” (Qiu, Lin, Leung, & Tov, 2012), and individuals who spend more time on Facebook are more likely to agree that others are “happier” and have “better lives” (Chou & Edge, 2012). As such, given that rumination involves passive and repetitive fo- cus on one’s distress, social comparison may provide ample opportunity to mull over causes and consequences of perceived inferiority.

Finally, consistent with the larger literature on risk factors for depression, we found that rumination was associated with depressive

Table 2 Standardized Path Coefficients, Unstandardized Path Coefficients, and Standard Errors

Path Standardized path

coefficient Unstandardized path

coefficient Standard error

T1 Facebook social comparison ¡ T2 Rumination .17� .15� .05 T1 Facebook social comparison ¡ T2 Depressive symptoms .03 .01 .02 T2 Rumination ¡ T2 Depressive symptoms .29� .16� .03 T1 General social comparison ¡ T2 Rumination .01 .03 .10 T1 General social comparison ¡ T2 Depressive symptoms �.02 �.03 .05 T1 Rumination ¡ T2 Rumination .58� .60� .05 T1 Depressive symptoms ¡ T2 Depressive symptoms .56� .57� .05 T2 Depressive symptoms error term — 33.40� 2.89 T2 Rumination error term — 125.97� 10.90

Note. ¡ � predicting. � p � .001.

167NEGATIVE SOCIAL COMPARISON

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

symptoms. If an individual ruminates on his or her perceived inferiority subsequent to nega- tively comparing oneself with others on Facebook, he or she is engaging in an emotion regulation strat- egy known to maintain and exacerbate distress. Nolen-Hoeksema et al. (2008) speculated that rumination may prolong and increase depres- sion, in part, because it is associated with re- duced interpersonal problem solving, less will- ingness to engage in pleasant activities to lift mood when given the chance, and more pessi- mistic views about positive events in the future. Thus, rather than problem solving (i.e., chang- ing the situation that led to the negative social comparison) or switching the focus of attention to more positive or rewarding aspects of their environment, these individuals may continue to passively focus on distress (perhaps via in- creased time spent on the site) and/or seek out others in potentially problematic ways (e.g., corumination, excessive reassurance seek- ing). The significant indirect effect of Face- book social comparison on depressive symp- toms through rumination is also consistent with a recent study that found that rumination mediated the association between negative status updates on Facebook and depressive symptoms (Locatelli et al., 2012). Together, these findings lend confidence to the notion that rumination may play a mechanistic role in the associations between negative social networking experiences and depressive symp- toms, as two studies have now demonstrated similar meditational effects despite different predictor variables and methodological fea- tures.

It is useful to note that research on Facebook use and mental health has important public health implications. First, given the bidirec- tional and transactional nature of depression and problematic interpersonal functioning (see Joiner & Timmons, 2009), in order to continue to understand the bounds of their association it is incumbent upon researchers to examine in- terpersonal processes in the contexts in which they currently occur, which includes social net- working sites. Next, the media has taken a par- ticular interest in research on social networking, and has oft made large-scale claims about the “dangers” of Facebook use (e.g., “Facebook depression”). As scientists developing a pro- grammatic line of research to test a priori hy- potheses regarding Facebook use and mental

health, we hope to shed light on the processes that may render individuals more vulnerable to negative effects of Facebook use.

The current study has several strengths, in- cluding a large and racially/ethnically diverse sample, a prospective design, and the use of path analysis to test the proposed mediation model. Nevertheless, it is important to keep in mind the following limitations that underscore the importance of replicating and extending these findings. We relied on a one-time self- report assessment of social comparison tenden- cies, and findings would be strengthened by a laboratory manipulation of this process. Addi- tionally, our sample was composed of nonclini- cal emerging adults and may not generalize to older or younger adults or a more clinically depressed sample. That said, that this effect was so robust in a nonclinical sample that com- monly uses social networking sites is quite no- table and suggests that this process may be especially relevant for those with elevated symptomatology. Consistent with current inter- ventions designed to enhance adaptive emotion regulation (see Kring & Sloan, 2010), findings suggest that targeting rumination in a clinical context may minimize the negative conse- quences of engaging in negative social compar- ison. It will be useful for future research to test more complex models that account for the various types of maladaptive emotion regula- tion strategies that individuals may engage in to regulate their mood subsequent to negative social comparison on Facebook, as well as additional types of mental health outcomes that may be affected by these processes. Fu- ture research could also benefit from assess- ing specific behaviors related to rumination that may take place in the context of social networking (e.g., rereading posts on Face- book that contribute to one’s dysphoria). Fi- nally, our study focused on only one set of variables that may contribute to associations between Facebook use and depressive symp- toms. Ongoing development of theory about the function of social networking, the con- texts in which it occurs, and the ways in which it is related to self-concept and identity may provide new avenues of research that will allow for further understanding of its effect on people of all ages.

168 FEINSTEIN ET AL.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

References

Ahrens, A. H., & Alloy, L. B. (1997). Social com- parison processes in depression. In B. P. Buunk & F. X. Gibbons (Eds.), Health, coping, and well- being: Perspectives from social comparison theory (pp. 389 – 410). Mahwah, NJ: Erlbaum Publishers.

Aldao, A., Nolen-Hoeksema, S., & Schweizer, S. (2010). Emotion-regulation strategies across psy- chopathology: A meta-analytic review. Clinical Psychology Review, 30, 217–237. doi:10.1016/j .cpr.2009.11.004

Allan, S., & Gilbert, P. (1995). A social comparison scale: Psychometric properties and relationship to psychopathology. Personality and Individual Dif- ferences, 19, 293–299. doi:10.1016/0191- 8869(95)00086-L

Antony, M. M., Rowa, K., Liss, A., Swallow, S. R., & Swinson, R. P. (2005). Social comparison pro- cesses in social phobia. Behavior Therapy, 36, 65–75. doi:10.1016/S0005-7894(05)80055-3

Bessière, K., Kiesler, S., Kraut, R., & Boneva, B. S. (2008). Effects of Internet use and social resources on changes in depression. Information, Communi- cation, & Society, 11, 47–70. doi:10.1080/ 13691180701858851

Cheung, M. S., Gilbert, P., & Irons, C. (2004). An exploration of shame, rank, and rumination in re- lation to depression. Personality and Individual Differences, 36, 1143–1153. doi:10.1016/S0191- 8869(03)00206-X

Chou, H., & Edge, N. (2012). ‘They are happier and having better lives than I am’: The impact of using Facebook on perceptions of others’ lives. Cyber- psychology, Behavior, and Social Networking, 15, 117–121. doi:10.1089/cyber.2011.0324

Davila, J. (2011). The “Facebook Depression” con- troversy. Retrieved from http://www.psychology .sunysb.edu/jdavila-/webpage/facebook%20 depression%20controversy.htm

Davila, J., Hershenberg, R., Feinstein, B. A., Gor- man, K., Bhatia, V., & Starr, L. (2012). Frequency and quality of social networking experiences: As- sociations with depressive symptoms, rumination, and co-rumination. Psychology of Popular Media Culture, 1, 72– 86. doi:10.1037/a0027512

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117–140. doi: 10.1177/001872675400700202

Gibbons, F. X., & Buunk, B. P. (1999). Individual differences in social comparison: Development of a scale of social comparison orientation. Journal of Personality and Social Psychology, 76, 129 –142. doi:10.1037/0022-3514.76.1.129

Giordano, C., Wood, J. V., & Michela, J. L. (2000). Depressive personality styles, dysphoria, and so- cial comparisons in everyday life. Journal of Per-

sonality and Social Psychology, 79, 438 – 451. doi: 10.1037/0022-3514.79.3.438

Haferkamp, N., & Krämer, N. C. (2011). Social com- parison 2.0: Examining the effects of online pro- files on social-networking sites. Cyberpsychology, Behavior, and Social Networking, 14, 309 –314. doi:10.1089/cyber.2010.0120

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conven- tional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. doi:10.1080/ 10705519909540118

Joiner, T. E., & Timmons, K. A. (2009). Depression in its interpersonal context. In I. H. Gotlib & C. L. Hammen (Eds.), Handbook of depression (2nd ed., pp. 322–339). New York, NY: Guilford Press.

Kline, R. B. (2005). Principles and practice of struc- tural equation modeling (2nd ed.). New York, NY: Guilford Press.

Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., & Scherlis, W. (1998). Internet paradox: A social technology that reduces social involvement and psychological well-being? Amer- ican Psychologist, 53, 1017–1031. doi:10.1037/ 0003-066X.53.9.1017

Kring, A. M., & Sloan, D. M. (2010). Emotion reg- ulation and psychopathology: A transdiagnostic approach to etiology and treatment. New York, NY: Guilford Press.

Locatelli, S. M., Kluwe, K., & Bryant, F. B. (2012). Facebook use and the tendency to ruminate among college students: Testing meditational hypotheses. Journal of Educational Computing Research, 46, 377–394. doi:10.2190/EC.46.4.d

Mackinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the indirect effect: Distribution of the product and resampling meth- ods. Multivariate Behavioral Research, 39, 99 – 128. doi:10.1207/s15327906mbr3901_4

Magid, L. (2011). “Facebook Depression”: A nonex- istent condition. The Huffington Post. Retrieved from http://www.huffingtonpost.com/larry-magid/ facebook-depression-nonexistent_b_842733.html

Manago, A. M., Graham, M. B., Greenfield, P. M., & Salimkhan, G. (2008). Self-presentation and gen- der on MySpace. Journal of Applied Developmen- tal Psychology, 29, 446 – 458. doi:10.1016/j .appdev.2008.07.001

McLaughlin, K. A., Hatzenbuehler, M. L., Mennin, D. S., & Nolen-Hoeksema, S. (2011). Emotion dysregulation and adolescent psychopathology: A prospective study. Behaviour Research and Ther- apy, 49, 544 –554. doi:10.1016/j.brat.2011.06.003

Morgan, C., & Cotten, S. R. (2003). The relationship between Internet activities and depressive symp- toms in a sample of college freshmen. CyberPsy- chology & Behavior, 6, 133–142. doi:10.1089/ 109493103321640329

169NEGATIVE SOCIAL COMPARISON

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Mussweiler, T., Ruter, K., & Epstude, K. (2006). The why, who and how of social comparison: A social- cognition perspective. In S. Guimond (Ed.), Social comparison processes and social psychology (pp. 33–54). Cambridge, England: Cambridge Univer- sity Press.

Nolen-Hoeksema, S., Larson, J., & Grayson, C. (1999). Explaining the gender difference in depres- sive symptoms. Journal of Personality and Social Psychology, 77, 1061–1072. doi:10.1037/0022- 3514.77.5.1061

Nolen-Hoeksema, S., & Morrow, J. (1991). A pro- spective study of depression and distress following a natural disaster: The 1989 Loma Prieta earth- quake. Journal of Personality and Social Psychol- ogy, 61, 115–121. doi:10.1037/0022-3514.61.1 .115

Nolen-Hoeksema, S., Stice, E., Wade, E., & Bohon, C. (2007). Reciprocal relations between rumina- tion and bulimic, substance abuse, and depressive symptoms in female adolescents. Journal of Ab- normal Psychology, 116, 198 –207. doi:10.1037/ 0021-843X.116.1.198

Nolen-Hoeksema, S., Wisco, B. E., & Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on Psychological Science, 3, 400 – 424.

O’Keeffe, G. S., & Clarke-Pearson, K. (2011). The impact of social media on children, adolescents, and families. Pediatrics, 127, 800 – 804.

Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. (1994). Social dominance orientation: A personal- ity variable predicting social and political atti- tudes. Journal of Personality and Social Psychol- ogy, 67, 741–763. doi:10.1037/0022-3514.67.4 .741

Qiu, L., Lin, H., Leung, A. K.-y., & Tov, W. (2012). Putting their best foot forward: emotional disclo- sure on facebook. Cyberpsychology, Behavior, and Social Networking, 15, 569 –572. doi:10.1089/ cyber.2012.0200

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general popu- lation. Applied Psychological Measures, 1, 385– 401. doi:10.1177/014662167700100306

Selfhout, M. H., Branje, S. J., Delsing, M., ter Bogt, T. F., & Meeus, W. H. (2009). Different types of Internet use, depression, and social anxiety: The

role of perceived friendship quality. Journal of Adolescence, 32, 819 – 833. doi:10.1016/j .adolescence.2008.10.011

Sidanius, J., Pratto, F., & Bobo, L. (1994). Social dominance orientation and the political psychol- ogy of gender: A case of invariance? Journal of Personality and Social Psychology, 67, 998 –1011. doi:10.1037/0022-3514.67.6.998

Suls, J., & Wheeler, L. (2000). A selective history of classic and neo-social comparison theory. Hand- book of Social Comparison. New York: Kluwer Academic/Plenum Press Publishers. doi:10.1007/ 978-1-4615-4237-7

Swallow, S. R., & Kuiper, N. A. (1988). Social com- parison and negative self-evaluations: An application to depression. Clinical Psychology Review, 8, 55–76. doi:10.1016/0272-7358(88)90049-9

Valkenburg, P. M., & Peter, J. (2007). Preadoles- cents’ and adolescents’ online communication and their closeness to friends. Developmental Psychol- ogy, 43, 267–277. doi:10.1037/0012-1649.43.2 .267

van den Eijnden, R. M., Meerkerk, G., Vermulst, A. A., Spijkerman, R., & Engels, R. E. (2008). Online communication, compulsive Internet use, and psychosocial well-being among adolescents: A longitudinal study. Developmental Psychology, 44, 655– 665. doi:10.1037/0012-1649.44.3.655

Wheeler, L., & Miyake, K. (1992). Social compari- sons in everyday life. Journal of Personality and Social Psychology, 62, 760 –773. doi:10.1037/ 0022-3514.62.5.760

Wood, J. V. (1989). Theory and research concerning social comparisons of personal attributes. Psycho- logical Bulletin, 106, 231–248. doi:10.1037/0033- 2909.106.2.231

Wood, J. V., & Lockwood, P. (1999). Social com- parisons in dysphoric and low self-esteem people. In R. Kowalski & M. Leary (Eds.), The social psychology of emotional and behavioral problems: Interfaces of social and clinical psychology (pp. 97–135), Washington, DC: American Psychologi- cal Association. doi:10.1037/10320-004

Received September 28, 2012 Revision received February 21, 2013

Accepted February 25, 2013 �

170 FEINSTEIN ET AL.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.