SHIFTING FROM AUTOGRAPH TO SELFIE: CONSUMER SELF IN THE FACEBOOK WORLD

The recent revolution of information technology has diffused Facebook widely among consumers, and it has a considerable impact on consumer self and modern marketing communication. Purpose of this study is to investigate the acceptance and impact of dematerialized Facebook possessions on consumer extended self. The theoretical foundation of Technology Acceptance Model (TAM) and extended self were used to develop the hypotheses of the study. The current study is based on primary data collected through a self-administered questionnaire, among a sample of 327 Sri Lankan undergraduates. Partial Least Squares Structural Equation Modeling (PLSSEM) was used to estimate the path coefficients and test the hypotheses developed in the study. Findings reconfirmed the original TAM relationships, enabling the use of TAM in identifying the predictors for accepting dematerialized Facebook possessions. Moreover, findings revealed that dematerialized Facebook possession usage extended consumer self. Theoretical and practical implications of these study and directions for further research are discussed.


Introduction
The current trend of Facebook has become a part of consumers' everyday life. For instance, worldwide active Internet users are 4.338 billion (Kemp, 2019), and daily active Facebook users are 1.79 billion (Facebook, 2020). This wave creates ample ofopportunities for the business community.However, "digital technologies have not only created potent new social networks but also dramatically altered how culture works" (Holt ,2016, p. 42). Therefore, it is essential to understand the consumer culture in the realm of Facebook to get a conductive outcome from Facebook based marketing campaigns.
Consumers have different self-images such as actual self (how consumers see themselves), ideal self (how consumers would like to see themselves) and the extended self. Sivadas and Machleit (1994, p. 143) defined the extended self as the "contribution of possessions to individual identity". According to Belk (1988, p. 139) "knowingly or unknowingly, intentionally or unintentionally, we regard our possessions as parts of our selves". However, the current digital revolution is altering consumer behavior, and it has considerable implications on the development of the consumer extended self (Belk, 2013).
When Belk (1988) presented the extended self-concept, people were using personal computers, wherethere were no other digital products such as web pages, online games, search engines, social media etc. (Belk, 2013). Belk (2016) discussed some digital modifications of the extended selfand one such modification was dematerialization, which means possessions are no longer material. In the digital age, tangible things such as written communications, recorded music, photos, videos are disappearing in front of our eyes (Belk, 2013(Belk, , 2016, and it is a sensitive and complex issue (Magaudda, 2011).In the pre-digitalworld, consumers used material possessions to reflect their identities. Since possessions have become dematerialized in the Facebook world, it is a timely need to clarify whether these dematerialized possessions can extend consumer self.
Consumer intrinsic values differ in predicting general Facebook usage and specific feature usage (timeline, wall, number of friends etc.) (Wijesundara &Xixiang ,2017;Smock, Ellison, Lampe, & Wohn ,2011). Further, specific features are possessions to its users (Belk, 2013;Watkins & Molesworth, 2012). As such, thestudy considers specific Facebook feature usage as dematerialized possessions usage. Since Facebook is a new technology, acceptance of dematerialized Facebook possessions can be understood from the technology acceptance perspective. Thus, current study utilizes the theoretical perspective of Technology Acceptance Model (TAM) in finding answers to the first research question while extended self theory was used as the theoretical foundation to support the second research question.
Findings of this study contribute to both academia and industry by understanding the consumer self-concept in Facebook world. First, evidence from past studies has so far little highlighted extended self concept in the digital world (Belk, 2013;Lehdonvirta, 2012). As such, the current study contributes to filling the gap in the existing consumer behavior literature by identifying predictors to use dematerialized Facebook possessions and the impact of dematerialized Facebook possession usage on consumer extended self. Second, from the industry perspective, findings encourage both digital marketers and social networking sites companies to consider the importance of dematerialized possessions in creating consumer identities. Further, investors will be benefitted since the study provides useful insights into consumer behavior from the perspective of an emerging economy.
The remainder of the paper is organized as follows. Section two is devoted to hypotheses development by focusing on the rational relationships among constructs underpinned by the literature. Section three explains the methods used in the study, followed by results with answers to the research questions. Validity, reliability and structural model evaluation are discussed in this section. Section five articulates a discussion, and finally, section six elaborates conclusion.

Technology acceptance model (TAM)
TAM is the mostly used theoretical model in explaining the acceptance of new technologies (Venkatesh, 2000), and was developed by Davis (1986) based on the Theory of Reasoned Action (TRA). According to TAM, Perceived Ease of Use (PEOU) and Perceived Usefulness (PU) are the two basic antecedents of Attitudes toward Technology (A), Intention to Use Technology (IU) and finally, Actual Usage (AU) of Technology (Choi & Chung, 2013). Over time, the TAM model has been modified twice as TAM2  and TAM 3 (Venkatesh & Bala, 2008). Rauniar, Rawski, Yang, and Johnson ( 2014 )have developed the revised TAM for social media to explain the social media acceptance by incorporating new constructs to the model such as critical mass, capability, perceived playfulness and trustworthiness.

Acceptance of dematerialized Facebook possessions
The first three hypotheses were developed based on TAM in order to identify the antecedents to use dematerialize Facebook possessions. PEOU is the foremost factor in predicting PU in accepting personal computers (Igbaria, Zinatelli, Cragg, & Cavaye, 1997).Anumber of scholars have provided empirical evidence to support a positive relationship between PEOU and PU of technologies within different cultural contexts (Choi & Chung, 2013;Liu ,2010;Pinho &Soares, 2011;Qin, Kim, Hsu & Tan ,2011;Rauniar et al.,2014).
Based on the associations proved in the TAM, coupled with the above empirical evidence, this study develops the following three hypotheses for Dematerialized Facebook Possessions Usage (DFPU).

Extended self
William and James (1910), one of the first authors who wrote about the self (Epstein, 1973)mentioned that we were sum of our possessions. According to William and James (1910), individuals body, psychic power and other belongings such as family, clothes, land house are part of his self (as cited in Belk, 1988). After that, a number of scholars contributed to the body of knowledge (Dixon & Street, 1975;Rochberg-Halton, 1984). However, Belk (1988) applied extended self in the field of consumer behavior (Cushing, 2012). Individuals acquire possessions in order to support the fragile sense of self since they are mainly what they have and possess (Tuan, 1980). This indicates that possessions play a vital part in a consumer's life by creating identities and extending their selves. According to Solomon (2012), extended self considers external objects as a part of individuals and described four levels of extended self as individual level (you are what you wear), family level (includes your house and furniture), community level (includes your neighborhood and hometown) and group level (includes your religion, flag, sports team, etc.).

Dematerialized Facebook possessions usage and the extended self
Technological development significantly modifies the extended self concept, which was originally presented in 1988 by Belk and allows consumers to extend the self as science fiction writers imagined 25 years ago (Sheth & Solomon, 2014). One of those modifications is the dematerialization of possessions in the digital world. Collections, pictures, letters, music, and greeting cards are transforming into dematerialized goods in the digital world (Siddiqui & Turley, 2006). Digital goods are playing a substantial part in consumers' everyday lives, not just supplanting material equivalents (e.g., eBooks, digital music) yet in addition presenting new forms of possession (e.g., social networking profiles, virtual possessions within videogames) (Watkins & Molesworth, 2012). Lehdonvirta (2012)suggests that digital goods owners consider they are very real to them. Denegri-Knott, Watkins, and Wood (2012)highlighted that individuals ritually convert virtual goods into meaningful properties. Further, Lehdonvirta, Wilska, and Johnson (2009)pointed out that digital goods play social roles as same as its material counterpart.
Facebook is a collection of dematerialized possessions (profile, timeline, and friends etc.). These possessions have the features relating to the material self-extending possessions such as attachments, identity and fear of loss. Attachment means, "caring about, being fond of and being miserable if the object of our affect is absent" (Turner & Turner, 2013, p. 1).Regardless of immaterial nature and lack of legal ownership, individuals possess and form emotional attachments to digital goods (Watkins & Molesworth, 2012). A user becomes emotionally attached to the social media, and hence, it becomes a part of theuser's self-definition and representation in the digital world (Wang, Yeh, Yen, & Sandoya, 2016). Dalisay, Kushin, Yamamoto, Liu, and Buente (2016)mentioned that minority college students demonstrated moderate levels of emotional attachment to social media sites. Facebook group members request to put "likes" and "share" things happening in those groups, and some users send requests to join groups, which they are interested in. In addition, users comment and like their own photos on Facebook. All these behaviors represent the nature of the attachment to the dematerialized Facebook possessions.
In psychology, self-identity means cognitive construct of the self that answers the question of who am I (Hogg, 2001). Identity comprises the unique characteristics communicated by a specific individual's presence (Dennen & Burner, 2017, p. 1). Social categories such as groups, relationships, and personal characteristics act as a part of self-identity, which support individuals to define themselves (Carter & Grover, 2015). In order to develop an online identity, one needs to be an active online user. Individuals who maintain a profile on online platforms, and share words, images, and preferences through these profiles are forming online identities (Dennen & Burner, 2017).Communication of online identity through social media helps the user to create both social presence and emotional presence (Bozkurt & Tu, 2016). Undergraduates independently develop a social media presence and communicate their identity to interact with peers (Dennen & Burner, 2017). Individuals now move from "you are what you wear" to "you are what you post (Sheth & Solomon, 2014, p. 126). This indicates that dematerialize Facebook possessions have an ability to create identities to its users.
Fear of loss is also attached to dematerialized Facebook possessions. Mainly,Facebook users are using password protection method to protect Facebook account from unauthorized access. From the company side, Facebook uses various security methods to protect its users from hackers. These pieces of evidence suggest that dematerialized Facebook possessions have self-extending features.
Further, there is a tendency to treat people as possessions and extensions of self, similar to the way tools are used (Belk, 1988). In Facebook, users have thousands of friends; some friends are second-order friends (friends of friends) or more than that, sometimes they have never met or not have any idea to meet (Clemons, 2009). This indicates that people resource onFacebookhelps users to extend their self.Additionally, Facebook has replaced many facets of the materialized possessions with numerous features day by day. The trend of selfie provides strong evidence for this transition. A decade ago, when people meet favorite celebrities, they used autograph or any other material to get the signature of the particular celebrity and showed it to others. Now autographs are being gradually replaced by selfie with celebrities, which are posted and commented on Facebook with likes and reacts. Tangible photo albums are also being saved as intangible albums on Facebook. Creative people wrote short poems in pieces of papers to express themselves and feelings (sadness, happiness) and now they appear in Facebook status. Timeline of Facebook plays the role of diary for the users. All these evidence reveal that dematerialized Facebook possessions can extend consumer self positively.
Thus, this study proposes following hypothesis for dematerialized Facebook possessions usage and extended self.

Data collection and participants
The sample comprised of undergraduates from seven universities in Sri Lanka. Internet penetration was 47%, and social media penetration was 30% in January 2020 (Digital 2020: Sri Lanka). According to the Networked Readiness Index 2020, Sri Lanka is the top leading country in the South Asian region (rank 83), indicating as a feasible location for social media studies. Shaouf, Lü, and Li (2016)suggest thatuniversity students might be the most suitable sample for e-commerce related studies. In addition, the majority of the Sri Lankan Facebook users are in the age group of 18 to 34 (Amarasinghe, 2011), which can be represented by undergraduates. Convenience sampling technique was used to select the sample.
Self -administered questionnaire with closed ended questions, was used to collect the data. Content validity was ensured by sound literature review and expert opinion (academics, industry researchers). A pilot survey was conducted with 10 students to test the questionnaire, and all the items were suitable for further proceeding. Thefirst section of the questionnaire includes questions adapted from previous studiesto measure the main constructs, including PEOU (Rauniar et al., 2014), PU (Choi & Chung, 2013;Rauniar et al., 2014), IU (Choi & Chung, 2013;Rauniar et al., 2014), Facebook Usage (Smock et al., 2011) and Extended self (Sivadas & Machleit ,1994). Respondents had to make their level of agreement in the scale from strongly agree to strongly disagree.
Next section of the questionnaire consisted of demographic factors and information relating to Facebook usage. Total of 500 questionnaires were distributed among the students. From those questionnaires, 368 were received. However, 41 questionnaires were eliminated from the final analysis due to incomplete answers. Finally, 327 questionnaires were used in the analysis, indicating 65.4% response rate, which is at a satisfactory level (Baruch ,1999).

Sample description
The sample consisted of 45.25 % of male and 54.75% of female undergraduates. The majority (48.9%) of them have more than three year experience in Facebook (Table 1). Facebook usage is about six hours per week among most students (39.1%). Majority of the students (30.88%) have 251-500 Facebook friends.

Reliability and validity
PLS-SEM was used with Smart-PLS 3.2.7 software to test the hypotheses and estimate path models involving latent variables, which were observed through multiple indicators. PLS-SEM was selected since the goal of the study was to predicting key target constructs, and the study was an extension of an existing structural theory (Hair, Ringle, & Sarstedt ,2011).There was no critical issue in normality since all skewness and kurtosis values were between -1 and +1 (Sarstedt, Ringle, & Hair, 2017). As recommended by Hair , Hult, Ringle, and Sarstedt (2014), certain items were dropped from the analysis due to lower outer loadings. Table 2 illustrates the outer loadings and indicator reliability values for all selected items. Composite Reliability (CR) and Cronbach's alpha can be used to assess the internal consistency reliability.Both CR and Cronbach's alpha lie between 0 and 1, higher values represent higher levels of internal consistency reliability (Hair et al., 2014;Gliem & Gliem, 2003). All the reliability indicators are at a satisfactory level (Table  3).Convergent validity of a study can be ensured by average variance extracted (AVE) (Fornell & Larcker, 1981).AVE values greater than 0.5 can be considered as an adequate level of convergent validity (Bagozzi & Yi, 1988).There is no critical issue in convergent validity (Table 3).   Fornell-Larcker criterion (1981) "compares the square root of the AVE values with the latent variable correlations" (Hair et al., 2014, p. 105), and used to assess the discriminant validity. Table No 4illustrates that thediscriminant validity is met in this research since the square root of AVE of each latent variable is greater than the latent variable correlations.

Structural model evaluation and hypotheses testing
The first step of the structural model evaluation is to test the multicollinearity among the independent variables. Sarstedt et al. (2017)suggestVIF values above five indicate collinearity among the predictor variables. According to the Smart-PLS report, all VIF values were below 0.5, indicating that there is no multicollinearity issue in the current study. Next step is to assess the significance and relevance of structural model relationships. Table 5 illustrates path coefficients of the structural model and significance testing results. All paths were statistically significant, supporting all four hypotheses.  R2 value explains the variance explained of endogenous constructand lies between 0 to 1, and higher value point toward more predictive accuracy (Sarstedt et al., 2017). R ² values of 0.75, 0.50, or 0.25 for endogenous latent variables can be considered as substantial, moderate, or weak, respectively. However, R² results of 0.20 is considered high in some study disciplines, for instance in consumer behavior (Hair et al., 2011 p.145). Given that current study is about user behavior in Facebook context, the results indicate that model explains a substantial part of the variance in the endogenous variables PU, IU, ES with an average R2 of 0.244,0.139 and 0.170, respectively. However, for DFPU (R2= 0.139) was comparatively low (Table No 6). F20.02, 0.15 and 0.35, consider as small, medium, and large effects respectively (Sarstedt et al., 2017). Table No 6 illustrates that all F2 values are at a medium level. Q2 values above zero indicate the model's predictive accuracy (Sarstedt et al., 2017). As per Table No  6,Q2 values provide sufficient evidence to model's predictive accuracy.

Discussion
This study examines antecedents to use dematerialized Facebook possessions and the impact of dematerialized Facebook possessions on consumer extended self. First, the study assumed that PEOU of Facebook was positively related to PU of Facebook, which was accepted by empirical evidence. This finding is consistent with the previous studies (Liu, 2010;Qin et al., 2011;Rauniar et al., 2014). These cond Table 7 illustrates the summary of hypotheses testing results and Figure No 1depicts the final structural model.
hypothesis was that PU of Facebook is positively related to IU Facebook. Empirical evidence provides sufficient evidence to accept this relationship by supporting the previous literature (Choi & Chung, 2013;Davis, 1989;Liu, 2010;Rauniar et al., 2014). Then, the study assumed IU Facebook was positively related to DFPU, which was accepted at 95% confidence level. This finding is in agreement with the previous studies related to actual technology usage (Liu, 2010;Rauniar et al., 2014). Fourth hypothesis, the originality of the current study was accepted by sufficient evidence indicating DFPU can extend consumer self positively.
The current study contributes to social media and consumer behavior literature. This is the first study, which considers dematerialized Facebook possessions as a predictor for consumer extended self. Adapted indicators were validated through proper validity and reliability techniques, providing an opportunity to replicate them in future studies. This is a significant contribution to social media, consumer behavior theory and future studies.
First, three hypotheses were related to TAM and study reconfirmed the original TAM relationships, enabling to use TAM in Facebook context. With the digital revolution, most material possessions (greeting cards, gifts, and picture memories) are transforming into digital goods. However, the question aroused was whether consumers accept these dematerialized possessions as self extending goods. This study contributes to the existing literature by providing an answer to this question. Findings suggest that dematerialized possessions can extend consumer self positively. Digital marketers, SNS companies and investors will be benefited from the current findings. According to Cushing (2012), as compared to the old consumers, younger consumers consider digital possessions as a part of their extended self. Thecurrent study confirms that young consumers consider dematerialized Facebook possessions have self-extending features, with a sample of undergraduates. These young consumers will become mass consumers in future and coming generation will be even closer to digital culture. As such, digital platform is essential to communicate with customers in future. Digital marketers can use the findings of this study to develop marketing strategies for their companies. In the offline world, consumers need to purchase certain products to reflect their identities. However, in the digital world, they have many options to show their commitment to brands such as online brand communities, fan pages etc. Even though consumers cannot buy some brands (selfextendingbrands), they can show others that they have an association with brands by joining to these brand communities or simply putting like to a page. Further, marketers can arrange more online gathering forums using Social Networking Sites (SNS). These online forums create a good platform for two-way communication. Moreover, marketers can introduce online brand symbols (gold color symbol, silver color symbol) based on consumer loyalty. Therefore, users can use these symbols to show their identities in online forums. In Facebook, consumers can tag brands in their profile pictures. Since selfies are becoming a popular trend, brand ambassadors (celebrities) can upload selfies with tagged brands. Hence, their fans will see those brands frequently and they will follow the same trend.
Advertising income is one of the main revenue sources of many SNS companies and it depends on the number of members. Since dematerialized possessions can extend consumer self, SNS companies can introduce more attractive features (dematerialize possessions) in their SNS to attract more users.
Study was conducted in an emerging economy. Thus, findings will help investors to make their investment decisions in such economies. For instance, they can invest more in digital goods, which can be sold in emerging economies since consumers consider dematerialized possessions as self-extending goods.

Conclusion
This study was designed to achieve mainly two research objectives. First, to identify antecedents to use dematerialized Facebook possessions. The theoretical foundation of TAM was used to accomplish this objective and found PEOU, PU and IU were the predictors for DFPU. As such, the original TAM relationship was reconfirmed in Facebook context. Second, the study endeavored to examine how dematerialized Facebook possessions influenced extended self. Findings confirm that dematerialized Facebook possessions will extend consumer self positively. These findings will contribute to consumer behavior and social media literature and allow practitioners to see consumer behavior patterns in the Facebook world from a different perspective.
Although this study provides useful insights into consumer behavior and Facebook usage, the study is still subject to some limitations. Though Facebook is a global phenomenon, it is constrained by local conditions such as culture (Wijesundara, 2014). As such, future researchers can use this model to understand consumer extended self in other cultural settings. Further, there are many SNSs such as LinkedIn, Twitter, Wechat and Facebook. However, this study focused only on Facebook. In order to have a deeper understanding, it is better to study this model with other SNSs. In this study, only young adults were recruited as respondents, other demographic groups are also growing fast on Facebook. Future studies should take into account more age groups due to age can serve as an important factor in Facebook usage and extended self.