Understanding how upward social comparison stimulates impulse buying on image‑sharing social commerce platforms: A moderated mediation model of benign envy and self‑esteem / Trần Văn Đạt

Understanding how upward social comparison stimulates impulse buying on image-sharing social commerce platforms: A moderated mediation model of benign envy and self-esteem Article in Cur

Understanding how upward social comparison stimulates impulse buying on image-sharing social commerce platforms: A moderated mediation model of benign envy and self-esteem

Article  in   Current Psychology · April 2022

DOI: 10.1007/s12144-022-03042-w

CITATIONS

7

READS

1,147

3 authors:

Van-Dat Tran

Banking University of Ho Chi Minh City, Vietnam

20PUBLICATIONS    362CITATIONS    

SEE PROFILE

Trieu Nguyen

National Kaohsiung University of Science and Technology (NKUST)

4PUBLICATIONS    19CITATIONS    

SEE PROFILE

Dung Minh Nguyen

National Kaohsiung University of Science and Technology (NKUST)

8PUBLICATIONS    101CITATIONS    

SEE PROFILE

Understanding how upward social comparison stimulates impulse

buying on image‑sharing social commerce platforms: A moderated

mediation model of benign envy and self‑esteem

Dat Van Tran 1  · Trieu Nguyen 2  · Dung Minh Nguyen 3

Accepted: 23 March 2022

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022

Abstract

The pervasiveness of image-sharing social commerce platforms has led to renewed interest in the topic of online impulse buying These platforms possess unique characteristics that facilitate both social comparison and impulse buying behavior However, little research has explored the relationship between upward social comparison and impulse buying in the context

of image-oriented social commerce Moreover, little is known about the underlying mechanism mediating and moderating this association Drawing on social comparison theory, this study developed and examined a moderated mediation model which integrates upward social comparison, impulse buying, benign envy, and self-esteem to address this gap The data was collected from 318 university students in Vietnam and analyzed using structural equation modeling The results reveal that upward social comparison directly stimulates adolescents’ impulse buying behavior, and benign envy mediates this rela-tionship Interestingly, self-esteem moderates the effects of upward social comparison and benign envy on impulse buying, respectively This study provides researchers with a new approach to explain online impulse buying on social commerce platforms In addition, our findings can serve as a reference for businesses and marketers to implement concerted strategies that encourage consumers to purchase impulsively on image-sharing social commerce platforms

Keywords Upward social comparison · Benign envy · Impulse buying · Self-esteem · Image-sharing social commerce platforms · Adolescents

Introduction

In recent years, the pervasiveness of social networking sites (SNSs), including Facebook, Instagram, Twitter, What-sApp, Pinterest, has made a significant change in the way people live, purchase, and communicate every day The growing popularity of SNSs has ushered in a new era of electronic commerce, known as social commerce, which is changing our thinking about online shopping (Xiang et al., 2016) Basically, social commerce is a combination of social media and e-commerce that conducts various types of social media-supported commercial activities (Liang & Turban, 2011; Xiang et al., 2016) Both consumers and retailers can benefit from social commerce It offers consumers a con-venient, seamless, and social shopping experience Social commerce also allows retailers to develop high relationship quality with consumers, resulting in increased sales and cus-tomer engagement (Oladapo, 2021) Thus, social commerce

is predicted as a potential research topic in marketing and information systems in the upcoming decades (Hajli, 2014)

* Dung Minh Nguyen

I108123114@nkust.edu.tw

Dat Van Tran

dattv@buh.edu.vn

Trieu Nguyen

F110121116@nkust.edu.tw

1 Faculty of Business Administration, Ho Chi Minh University

of Banking, Ho Chi Minh City, Vietnam

2 International Master of Business Administration, College

of Management, National Kaohsiung University of Science

and Technology, No 1, University Road, Yanchao District,

Kaohsiung City 82445, Taiwan, Republic of China

3 College of Management, National Kaohsiung University

of Science and Technology, No 1, University Road, Yanchao

District, Kaohsiung City 82445, Taiwan, Republic of China

Emerging from Web 2.0, social commerce platforms (SCPs)

enable social interaction to assist online selling and buying

activities and allow consumers to review, recommend, share,

and purchase products or services (Hajli, 2015) Since 2006,

there has been a new emerging form of SCPs concentrating on

image orientation (i.e., image-sharing SCPs) The prominent

examples of image-sharing SCPs include Pinterest, Instagram

(in the US), and Mogujie (in China) Distinguishable from other

forms of SCP (e.g., text-oriented and video-oriented SCPs),

image-sharing SCP users share pictorial information rather than

textual information on SCPs, such as pictures about products

and lifestyles (Kim et al., 2017) However, just like with most

Web 2.0 applications, image-sharing SCPs focus on a particular

segment of consumers (e.g., females) or a particular product

(i.e., lifestyle and fashion) (Liang & Turban, 2011)

In recent years, image-sharing SCPs have deeply

influ-enced the users’ consumption behavior, especially in the

fash-ion segment Represented by Instagram, it has become the

biggest fashion source where the young are inspired by

vari-ous fashion styles from celebrities and influencers According

to a survey by FacebookIQ (2017), 56% of the surveyed social

media users consider Instagram as a major online platform

for fashion shopping, significantly higher than 42% of those

who cited Facebook Similarly, a recent report shows that

Pinterest users spend 80% more in retail than non-Pinterest

users (Burnie, 2020) Meanwhile, a recent report indicated

that customers tend to buy goods impulsively as 84% of

consumers have purchased products without any particular

plan, and approximately two-fifths of online consumption is

purchased impulsively (Saleh, 2017) In addition, for

image-sharing SCPs, impulse buying behavior is even more easily

simulated because “a picture is worth a thousand words”,

and the human brain can cruise through and process visual

information up to sixty thousand times faster than the textual

one (Tanda, 2017) The visualizing power of images has been

also proven to trigger viewers’ emotions effectively

(Messa-ris, 1997) These evidence suggests that image-sharing SCP

users are more susceptible to impulse purchases than users of

other social commerce platforms (Xiang et al., 2016) It is also

worth noting that among different consumer groups,

adoles-cents are prone to impulse buying (Liu et al., 2019) because

adolescents’ skills for controlling impulse sentiments are not

much dexterous, enabling them to try novel things and take

more risks (Pechmann et al., 2005) Thus, when surfing on

image-sharing SCPs (e.g., Instagram), the young are exposed

to a volume of visual information from vendors, news, friends,

celebrities, or experts, making it difficult for them to resist

the temptation of impulse buying Furthermore, adolescents

also represent a sizable online market segment owing to their

remarkable purchasing power (Statista Research Department,

2017) and heavy Internet usage (Liu et al., 2019)

Given the prevalence of image-oriented SCPs and the

importance of online impulse buying, understanding of

determinants of impulse buying on image-sharing SCPs has become essential (Liu et al., 2019) Extant literature has explored various factors that stimulate consumers to pur-chase impulsively on SCPs They can be categorized into four main types: social factors (e.g., social presence, similar-ity), website-related factors (e.g., social interaction, informa-tion quality), consumer characteristics (e.g., impulsiveness, usage intensity), and marketing-related factors (e.g., product availability, aesthetic appeal) (Abdelsalam et al., 2020) How-ever, to the best of our knowledge, understanding of factors influencing consumer impulse buying behavior in the context

of image-sharing SCPs has remained limited Xiang et al (2016) is a pioneer in studying impulse buying behavior on image-oriented SCPs that emphasized the role of parasocial interaction Most recently, Djafarova and Bowes (2021) iden-tified three antecedents of impulse purchases on Instagram: advertisement, micro-celebrity, and user-generated content

In addition to these variables, we argue that other factors still exist that could well explain impulse buying behavior

on image-sharing SCPs, yet have not been explored before

To address the above mentioned gap, this study focuses

on social comparison, which is a common social phenom-enon on SCPs, and explores its possible effects on impulse buying behavior The association between social compari-son and impulse buying is expected to be more prominent

on image-sharing SCPs rather than text-based or video-based SCPs due to its unique characteristics Besides that, the underlying mechanism of relationship between social comparison on image-sharing SCPs as well as impulse buy-ing and the correspondbuy-ing mediatbuy-ing mechanism (e.g., how this relationship occurs) and moderating mechanism (e.g., when the effect is stronger) remains inadequately explored Identifying these mechanisms will provide a comprehen-sive understanding of how upward social comparison on image-sharing SCPs stimulates impulse buying behavior Thus, this study attempts to enrich the existing literature

by examining the mediating role of benign envy and the moderating role of self-esteem in the effect of upward social comparison on SCPs on impulse buying

The contribution of this study is threefold First, the study is grounded on social comparison theory to examine

a model of impulse buying behavior in the context of image-sharing SCPs that had been rarely explored before Second, while benign envy was considered an outcome of upward social comparison, this research is a pioneer in examin-ing it as a mediator between upward social comparison and impulse buying Further, our study also sheds light on the moderating role of self-esteem on the effect of upward social comparisons on impulse buying Third, the findings from this study can be helpful for companies and market-ers to reconsider the positive aspect of impulse buying on image-sharing SCPs and thus formulate more effective strat-egies to induce impulse buying among online consumers

Theoretical Background

Impulse Buying

Impulse buying was defined by Stern (1962) as an

unplanned, captivating, and hedonically complicated

pur-chasing behavior exhibited by consumers This buying

behavior usually occurs when the consumer has a sudden,

solid, and irresistible desire to purchase a particular product

immediately without considering the consequences (Park

et al., 2012; Chen & Wang, 2016), and the purchase

deci-sion is made in a short amount of time (Kim & Johnson,

2016) Notably, impulse buying possesses unique

charac-teristics that distinguish it from other purchasing behavior

types: unintended, unplanned, quick decision, inactive, and

thoughtless (Abdelsalam et al., 2020) The literature has

identified four distinct types of impulse purchases: pure,

reminder, suggestive, and planned impulse buying (Stern,

1962) Consistent with Stern’s definition, most users’

buy-ing behavior on SCPs can be considered impulse buybuy-ing

(Xiang et al., 2016; Abdelsalam et al., 2020) For example,

consumers browse images of clothes on SCPs purposeless

and then decide to buy a blouse or T-shirt could be viewed

as a purely impulsive purchase In contrast, planned impulse

buying occurs when consumers browse product images on

the SCPs with a prepared shopping list but only purchase a

particular product based on the offered discounts or

promo-tions A reminder impulse buying occurs when a user sees

an image of a cosmetic product and realizes that his or her

home stock is running low Meanwhile, after seeing a new

dress recommended by a SCP and purchasing it, it can be

considered suggestive impulse buying

So far, researchers have adopted many social and

psy-chological related theories to explain consumers’ online

impulse buying behavior in social commerce from different

angles The theoretical foundations for existing online impulse-buying studies are presented in Table 1 Accord-ingly, the stimulus-organism-response (i.e., S-O-R) frame-work are the most prominent theoretical approach over the past decade Impulse buying behavior in social commerce has also been explained with other theoretical approaches, such as flow theory (Hsu et al., 2012; Huang, 2016; Wu

Zafar et al., 2020a; Chen et al., 2016), social capital theory (Huang, 2016), social influence theory (Yang et al., 2021), uses and gratification theory (Zafar et al., 2020b) These diverse angles have offered a more in-depth and compre-hensive understanding of consumers’ online impulse buying

on social commerce platforms However, since the focus of this study is on image-sharing SCPs, which easily stimulate social interaction and comparison between users, we attempt

to explain impulse buying under the perspective of social comparison using the social comparison theory

Social Comparison Theory

Initially introduced by Festinger (1954), social comparison theory is a well-known psychological theory in explaining people’s tendency to compare themselves to others The theory suggests that individuals have an innate drive to accurately evaluate their opinions and abilities by comparing themselves to others in the social world (Festinger, 1954; Liu

et al., 2019) Based on social comparison theory, plenty of empirical studies have been conducted to offer insights into three aspects of social comparison, including comparison target selection (i.e., whom individuals choose to compare themselves to), comparison motivation (i.e., why individuals engage in social comparison), and comparison reaction (i.e., how individuals react to social comparison) (Gerber et al., 2018; Nakayama et al., 2018)

Table 1 Theoretical foundation

of studies on impulse buying on

social commerce

Theoretical foundation Study Cognitive emotion theory Verhagen and Van Dolen ( 2011 ) Flow theory Wu et al ( 2020 ); Huang ( 2016 ); Hsu et al ( 2012 );

Latent state-trait theory Zafar et al ( 2021 ); Chen et al ( 2016 ); Wells et al ( 2011 ) Parasocial interaction theory Vazquez et al ( 2020 )

Reflective–impulsive model Shen and Khalifa ( 2012 ) S–O–R framework Ming et al ( 2021 ); Zafar et al ( 2020a ); Zheng et al ( 2019 ); Chen and

Yao ( 2018 ); Leong et al ( 2018 ); Xiang et al ( 2016 ); Zhang et al ( 2014 ); Floh and Madlberger ( 2013 ); Shen and Khalifa ( 2012 ) Social capital theory Huang ( 2016 )

Social influence theory Yang et al ( 2021 ) Social impact theory Chuang et al ( 2015 ) Social network paradigm Husnain et al ( 2016 ) Uses and gratification theory Zafar et al ( 2020b )

Regarding comparison target selection, individuals tend

to compare themselves with those who have similarities

(Buunk & Gibbons, 2007; Li, 2018) In general,

schol-ars have identified two main types of social comparison:

upward social comparison and downward social comparison

Upward social comparison occurs when individuals compare

themselves with others who are superior or better than they

are In contrast, downward social comparison occurs when

people compare themselves with others who are worse than

themselves (Liu et al., 2019) A recent meta-analysis study

on the social comparison (Gerber et al., 2018) reports that

individuals generally prefer making upward comparisons to

downward comparisons While upward social comparison

can lower self-regard (Tesser et al., 1988), Collins (2000)

argued that this was not always the case In many situations,

individuals compare themselves to others who are

bet-ter than themselves to obtain more useful information for

evaluating themselves (Buunk & Gibbons, 2007; Festinger,

1954), developing a positive self-image (Park & Park, 2017),

or creating a sense of belonging to a superior group and

enhance their perception of individual worth (Collins, 2000)

With the recent emergence of image-sharing SCPs, social

comparison has become more ubiquitous We find several

features on image-sharing SCPs that facilitate upward

social comparison First, celebrities and influencers often

share relevant images of sponsored products and services

onto image-sharing SCPs for advertising purposes Users,

thereby, can quickly come across these perfect and beautiful

images that stimulate them to make comparisons Second,

image-sharing SCPs operate on SNSs with many

commer-cial activities while SNSs are a fertile ground for upward

social comparison (Taylor & Strutton, 2016) Third, with

increasing time for using SNSs, users stand a higher chance

of approaching comparison-making stimulus on

image-shar-ing SCPs through a glimpse at advertisements (Liu et al.,

2019) Therefore, upward social comparison on

image-shar-ing SCPs may happen when users look at captivatimage-shar-ing images

of online celebrities in terms of their superior possessions

and appearance

Research Model and Hypotheses Development

We build on the social comparison theory to propose the research model (see Fig. 1) consisting of four constructs The model explains how impulse buying behavior on image-sharing SCPs is influenced

by upward social comparison, benign envy, and self-esteem

Upward Social Comparison, Benign Envy, and Impulse Buying

Existing literature on social comparison has identified two opposing types of consequences from upward social com-parison On the one hand, when individuals compare them-selves with others who are better than themthem-selves, they will feel inferior about themselves (de Vries & Kühne, 2015), leading to some negative emotions or outcomes, such as depression (Li, 2018), mental health damage (Jang et al., 2016) On the other hand, upward social comparison can also bring positive benefits, such as enhancing job per-formance (Cadsby et al., 2019) or eliciting inspiration on SNSs (Meier & Schäfer, 2018) Notably, some scholars have recently paid much attention to a common post-comparison emotion on SNSs, namely envy (Ahn et al., 2021; de Vries

& Kühne, 2015), which is defined as a painful experience when facing the good fortune of others (Tai et al., 2012)

In general, there are two types of envy: malicious envy and benign envy (Smith & Kim, 2007) Malicious envy views envy from a negative perspective that may carry harmful consequences, such as burnout (Liu & Ma, 2018) or fear of missing out (Yin et al., 2019), while benign envy considers envy from a positive perspective Although malicious envy

is the traditional and dominant approach, benign envy has gained more attention from scholars in recent studies due to its positive nature The latter approach emphasizes how envy can motivate individuals to excel and achieve desired things

by improving themselves, rather than bringing the envied targets down (Li, 2018; Crusius & Lange, 2014)

Fig 1 The research model

The logic mentioned above can be applied in the current study

When making social comparison, people realize that they lack

and desire something others have, such as possessions, superior

qualities, or achievements, leading to feeling of envy (Li, 2018)

Prior studies have concurred that envy often comes from the

upward social comparison (Ahn et al., 2021; Lange & Crusius,

2015; Li, 2018; Zheng et al., 2018) Suls et al (2002) found that

making a social comparison is a way to fulfill their need for

self-enhancement In the context of image-sharing SCPs, the

appeal-ing images of products and appearance of other users, especially

celebrities and influencers, are the key to intuitively arousing

upward social comparison After making upward social

com-parison with images of deemed superior celebrities, consumers

may form a sense of benign envy that spurs them to take action

in self-improvement to obtain the desired thing similar to that

possessed by the compared one (Crusius & Mussweiler, 2012;

to get caught up in thoughtless behaviors (i.e., impulse buying)

to get the desired items quickly The effect of benign envy on a

desire to buy luxury fashion items has been validated in a recent

study by Loureiro et al (2020) Accordingly, we hypothesize that:

H1 Upward social comparison on image-sharing SCPs

is positively associated with benign envy.

H2 Benign envy is positively associated with impulse

buying on image-sharing SCPs.

The social comparison theory indicates that individuals tend

to evaluate themselves by comparing with others about owned

assets or consumption behaviors (Festinger, 1954) When

brows-ing SCPs, users are immersed in a colossal amount of information

and pictures from other users, especially from celebrities and

influ-encers on online advertisements (Xiang et al., 2016; Abdelsalam

com-parison with the advertising pictures of celebrities or influencers

on image-sharing SCPs would provoke cravings for desired

prod-ucts Hence, we argue that users are easily stimulated by the visual

appeal to make social comparisons and then purchase products

impulsively The association between upward social comparison

and impulse buying has been validated in some contexts, such as

SNSs (Liu et al., 2019) and superior goods (Crusius & Mussweiler,

2012) Accordingly, the following hypothesis is proposed:

H3 Upward social comparison on image-sharing SCPs

is positively associated with impulse buying

The Moderating Role of Self‑esteem

Although upward social comparison and benign envy are

expected to impact impulse buying behavior positively, the

impacts may not be the same among adolescents Some

individ-ual characteristics may moderate these effects, for example,

self-esteem Self-esteem refers to an individual’s subjective evaluation

of their worth and denotes how individuals feel about themselves (Rosenberg, 1965; Wang et al., 2017) Self-esteem is an attrac-tive psychological construct since it is used to predict or explain certain outcomes, such as counterproductive working behaviors (Whelpley & McDaniel, 2016), fashion clothing involvement (Khare et al., 2012), and depression (Zhou et al., 2020)

In addition, past research has been attentive to the importance

of self-esteem degree (i.e., low self-esteem and high self-esteem) and empirically indicated that self-esteem exerts a moderating role between some psychological variables (e.g., Ahmad & Begum, 2020; Lee et al., 2018; Wang et al., 2018a, ) For example, Wang

was stronger for SNSs users with low self-esteem than those with high self-esteem Some scholars posited that the lower self-esteem users seem to feel satisfied in life and benefit more from their use of Facebook than the higher self-esteem users (Chen & Bello, 2017; Lee et al., 2018) Regarding image-sharing SCPs, little research has explored the moderating role of self-esteem on the indirect relations between upward social comparison and users’ impulse buying behav-ior Verplanken et al (2005) posited that low self-esteem individuals find impulse buying difficult to avoid while Dhandra (2020) sug-gested that low self-esteem individuals are fall victim to impulsivity Therefore, we predict that when making social comparisons with the superior on image-sharing SCPs, low self-esteem consumers are more susceptible to impulse buying, compared to high self-esteem ones Additionally, when the sense of benign envy occurs after mak-ing the upward social comparison, lower self-esteem individuals are more likely to purchase goods impulsively to quell the gap with the compared people Accordingly, we propose the following hypotheses:

H4a Self-esteem negatively moderates the relationship

between upward social comparison and impulse buying on image-sharing SCPs

H4b Self-esteem negatively moderates the relationship

between benign envy and impulse buying on image-sharing SCPs

Research Methodology

Sample and Data Collection

The research model was tested using data collected from stu-dents in the three most prominent universities in the south of Vietnam through convenience sampling method The partici-pants were asked to complete questionnaires provided directly by the authors The college students were chosen as respondents for two main reasons First, prior scholars argued that adolescents are more impulsive than adults and children (Pechmann et al., 2005) Second, teenagers tend to spend more time on social media compared to children and adults (Ballard, 2019) There-fore, the image or information presented on SCPs may offer ado-lescents ample opportunities to make social comparisons More-over, many prior studies have suggested that it is appropriate to use college student samples in online impulse buying research

because young people are the predominant online consumers

(e.g., Liu et al., 2019; Chan et al., 2017; Abdelsalam et al., 2020)

Before proceeding to the official survey, the pilot study was

con-ducted with 50 students to determine the feasibility of this research

The Cronbach’s alpha values of all constructs in the pilot test were

higher than 0.7 (e.g., 0.916 for upward social comparison, 0.907 for

impulse buying, 0.921 for benign envy, and 0.930 for self-esteem)

The results demonstrate that the questionnaire is reliable The official

survey was conducted in ten weeks, from March to May 2020 A total

of 361 students completed the survey questionnaire After

remov-ing 43 invalid responses, the final sample included 318 respondents

Among them, 141 (44.3% are male) and 177 (55.7%) are female

Respondents’ ages range between 18 and 24 Besides, 67

respond-ents (21.1%) are freshman, 112 (35.2%) are second-year studrespond-ents, 95

(29.8%) are third-year students and 44 (13.9%) are senior More than

half of surveyed students (51.9%) spent from 1 to 3 h per day, and

only 18.9% of respondents spent less than 1 h per day using SCPs

Measurement

The present research model consisted of four latent constructs,

in which each construct was measured with multiple items The

measurement items used in this research were adapted from

rel-evant studies discussed in the foregoing section Accordingly, the

scale for upward social comparison on image-sharing SCPs was

measured with six items adapted from Bai et al (2013) Benign

envy scale was assessed with five items adapted from Lange and

Crusius (2015) Self-esteem scale was adopted from Rosenberg

(1965) with ten items Impulse buying was measured with eight

items adapted from Badgaiyan et al (2016) and Verplanken and

Herabadi (2001) All measurement items were rated on a

seven-point Likert scale (i.e., 1 = strongly disagree; 7 = strongly agree)

and provided in Appendix A, Table 6

Control Variables

Prior studies have suggested age, gender, and monthly living

expenditure were related to students’ impulse buying behaviors

(Vohs & Faber, 2007) Additionally, Arampatzi et al (2018)

dem-onstrated that the amount of time on SNSs significantly influenced

students’ well-being Thus, this study included gender and amount

of time on SCPs per day as two control variables in the research

model to limit their confounding effects on the dependent variable

Results

We utilized structural equation modeling (SEM) as the main

data analysis method We followed the two-step approach

(Anderson & Gerbing, 1988) to test the measurement model

and structural model

Measurement Model

Confirmatory factor analysis (CFA) was performed to evaluate the measurement model’s overall fit, reliabil-ity, convergent validreliabil-ity, and discriminant validity using AMOS software First, the goodness of fit is evaluated via six indexes suggested by Bagozzi and Yi (1988), including

index (GFI > 0.90), adjusted goodness of fit index (AGFI

>0.90); normed fit index (NFI > 0.90), comparative fit index (CFI > 0.90), and root mean square error of approximation (RMSEA <0.08) As shown in Table 2, all goodness-of-fit indices exceed the recommended cut-off values, indicating that the measurement model is relevant to the collected data Second, Cronbach’s alpha (CA) and composite reliability (CR) were employed to examine the reliability of all vari-ables As illustrated in Table 2, the CA and CR values are higher than the recommended threshold of 0.70 (Hair et al., 2010) Thus, the reliability of all variable is satisfied Third, the convergent validity was evaluated by two criteria, including (1) average variance extracted (AVE) of variables and (2) factor loadings index Table 2 shows that the AVE for all variables is greater than 0.50 (Fornell & Larcker, 1981), and all factor loadings are significantly greater than 0.60 (Hair et al., 2010), supporting convergent validity of the instrument Fourth, the discriminant validity is evaluated by compar-ing the AVE with the square of the correlation parameters

As shown in Table 3, the AVE values of each construct are higher than the square of correlations between that and any other construct, demonstrating a solid discriminant validity (Fornell & Larcker, 1981) In addition, the confidence inter-val of correlation coefficients does not include the inter-value of

1, confirming sufficient discriminant validity

Because this study used the self-reported survey to collect data, the likelihood of common method variance (CMV) is a concern Following suggestions of Podsakoff

et al (2012), we took more efforts to minimize the effects

of CMV by guaranteeing the respondent anonymity and pre-testing the scale items adapted from previous stud-ies In addition, we performed Harman’s one-factor test

to examine the severity of CMV As a result, the explana-tory variance for the first factor is 33.87%, lower than the threshold of 50% (Podsakoff et al., 2012) Therefore, the problem of CMV is not serious in this study

Structural Model

df (150) = 2.521, RMSEA = 0.069, NFI = 0.917, RFI = 0.905, GFI = 0.900, IFI = 0.948, TLI = 0.941 and CFI = 0.948 These indices are acceptable (Bagozzi & Yi, 1988), indicating that the structural model achieves overall fit

The results of the structural model are provided in Table 4

and Fig. 2 Accordingly, upward social comparison on SCPs

has significant and positive effect on benign envy

(coeffi-cient path = 0.727, p < 0.001) and impulse buying (coeffi(coeffi-cient

path = 0.386, p < 0.001), supporting H1 and H3 In addition, the

influence of benign envy on impulse buying is significant and

positive (coefficient path = 0.532, p < 0.001), confirming H2

Figure 2 also shows the values of explained variances for benign

demonstrate that the model achieves good explanatory power Besides that, to identify whether the inclusion of the control variables (i.e., gender, usage time on SCPs per day) influences interpretation of the results of the proposed model (Spector & Brannick, 2011), we conducted a SEM analysis using bootstrap 2.000 re-samples The two control variables were inserted between upward social comparison on SCPs and impulse buying to inves-tigate whether these relationships can still be supported with the effect of the control variables As a result, all hypotheses are still supported when adding the two control variables Thus, the control variables do not have a bias towards the current outcomes

Mediating Effect of Benign Envy

To examine the mediating role of benign envy in the asso-ciation between upward social comparison on SCPs and impulse buying, we used model 4 of PROCESS macro (Hayes, 2013; Hayes et al., 2017) and bootstrapping method

Table 2 Reliability and

convergent validity of the

measurement model

CR = Composite reliability; AVE = Average variance extracted, ***: p < 0.001

Goodness-of- fit indexes: χ 2 /df (371) = 1.220, RMSEA = 0.026, CFI = 0.988, NFI = 0.938, GFI = 0.911, AGFI = 0.896

Upward social comparison on

USC2 0.802 ***

USC3 0.715 ***

USC4 0.755 ***

USC5 0.777 ***

USC6 0.861 ***

EV2 0.810 ***

EV3 0.779 ***

EV4 0.797 ***

EV5 0.851 ***

IB2 0.807 ***

IB3 0.788 ***

IB4 0.756 ***

IB5 0.758 ***

IB6 0.763 ***

IB7 0.841 ***

IB8 0.851 ***

SE2 0.832 ***

SE3 0.783 ***

SE4 0.746 ***

SE5 0.779 ***

SE6 0.777 ***

SE7 0.753 ***

SE8 0.728 ***

SE9 0.792 ***

SE10 0.851 ***

Table 3 The correlation matrix and discriminant validity

N = 318 Average variances extracted (AVEs) are in bold

The correlation coefficients are in italics

Constructs USC EV IB SE

(with bootstrap resamples N = 5000) Table 5 shows that the

95% bias-corrected confidence interval for the indirect effect

of benign envy on impulse buying is between 0.144 and

0.317, which excludes the value of zero The mediating path

is also significant (β = 0.226; p < 0.001) The results confirm

that benign envy mediates the effect of upward social

com-parison on SCPs on impulse buying

Moderating Effect of Self‑esteem

To examine moderating role of self-esteem, we used model 59

of PROCESS macro (Hayes, 2013; Hayes et al., 2017) First, the

direct effect of self-esteem on impulse buying was significant

(p < 0.01) Next, we examined whether interaction term

(self-esteem x upward social comparison) has a significant effect

on impulse buying The result shows a significant interaction

effect (β = −0.09, p = 0.024 < 0.05), indicating that self-esteem

negatively moderates the relationship between upward social

comparison on SCPs and impulse buying Thus, the H4a is

sup-ported In addition, the effect of interaction term (self-esteem

x benign envy) on impulse buying is also significant (β = 0.12,

p = 0.006 < 0.01) This result suggests that self-esteem moderates

the relationship benign envy and impulse buying, but this

mod-erating effect is positive Thus, the H4b is partially supported

self-esteem reports a higher level of impulse buying when making comparisons on SCPs It means that the positive effect of upward social comparison on SCPs on impulse buy-ing is more pronounced when self-esteem is low rather than when it is high However, the positive effect of benign envy

on impulse buying is strengthened for consumers with high level of self-esteem (see Fig. 4)

Discussion

This study set out to examine the effect of upward social comparison on impulse buying in the context of image-shar-ing SCPs through a moderated mediation model, in which benign envy is a mediator and self-esteem is a moderator Overall, the empirical results lend support to all proposed hypotheses In the following, the key findings of this study are discussed in more detail

First, the results confirm a significant and positive effect of upward social comparison on benign envy The result deepens our understanding of how benign envy was elicited by upward social comparisons occurring on image-sharing SCPs This finding is consistent with extant literature (e.g., Latif et al., 2021; Appel et al., 2016) which indicates that social compari-son on Facebook elicits benign envy Additionally, the finding is

Table 4 Path coefficients and significances

USC = Upward social comparison on SCPs,

EV = Benign envy, IB = Impulse buying, *** p < 0.001

Hypothesis Path Standardized path

coefficient Result H1 USC ➔ EV 0.727 *** Support

H2 EV ➔ IB 0.532 *** Support

H3 USC ➔ IB 0.386 *** Support

Fig 2 The path coefficients of

research model USC = Upward

social comparison on SCPs, BE

= Benign envy; SE =

Self-esteem *** p < 0.001; ** p <

0.01; * p < 0.05

Table 5 Mediating effect testing result

USC = Upward social comparison on SCPs, EV = Benign envy,

IB = Impulse buying, *: p < 0.001

Path Indirect effect 95% Confidence

interval Lower Upper USC ➔ EV ➔ IB 0.226 *** 0.144 0.317

also consistent with the study of Li (2018) that adolescents are

still developing their own self-concept, which is conducive to

self-evaluating and self-improving Nevertheless, due to a lack

of objective standards for self-evaluation, they tend to take the

better people as a yardstick to evaluate themselves, thus

eas-ily exposing themselves to envy Furthermore, image-sharing

SCPs users are more likely to be visually appealed by product

images featured by online celebrities or influencers who have a

vast number of media followers (Djafarova & Rushworth, 2017;

Pittman & Reich, 2016) As a result, image-sharing SCP users

are prone to make upward social comparisons with these digital celebrities, prompting benign envy

Second, upward social comparison on SCPs affects impulse buying directly and indirectly On the one hand, the direct effect of upward social comparison on impulse buying suggests that users are susceptible to impulse purchases after upwardly comparing with digital celebrities on Instagram or Pinterest whose captivating pictures are advertised for specific products

It also consolidates the findings from prior studies in various contexts, such as social networking sites (Liu et al., 2019) and

Fig 3 Plotted interaction of

upward social comparison

on SCPs and self-esteem on

impulse buying

Fig 4 Plotted interaction of

benign envy and self-esteem on

impulse buying