If I can make it, you are a feminist
The effects of the popularity of Instagram users posting female empowering
content on young adolescent’s attitude towards feminism, moderated by
credibility, homophily and wishful identification
Master’s Thesis J.H. Pikaar (11850620) Supervisor: Dr. J.M.F. Van Oosten University of Amsterdam Graduate School of Communication
Entertainment Communication 6375 words June 29th, 2018
2
This study aimed to check for effects of the popularity of Instagram users within the female empowering community on young female adolescent’s attitude towards feminism, and whether this effect is mediated by wishful identification, credibility and/or relatability. In addition, this experiment checked if this effect is moderated by a pre-existing positive attitude towards feminism. Ninety-one participants took part in the study, equally divided over two conditions. An Instagram profile from a popular user (condition 1) or a unpopular user (condition 2) was shown to the participants for a minimum of 15 seconds, where after participants filled in a questionnaire, measuring their attitude towards feminism, wishful identification with the user, credibility of the user and relatability with the user. The analyses showed no significant effects, inconsistent with previous research. There were no significant effects for popularity on attitude towards feminism, no significant effects for popularity on wishful identification, credibility and relatability and pre-existing attitudes towards feminism played no significant role. Hence, there was concluded that the popularity (amount of followers and follower engagement) of a social influencer on Instagram is of no importance when transferring societal attitudes, particularly in the case of (popular) feminism.
3
1. Introduction
2. Theoretical framework
2.1 The effectives of popular social influencers 2.2 Underlying mechanisms
2.2.1 Wishful identification 2.2.2 Relatability
2.2.3 Credibility
2.3 Pre-existing feminist attitudes
3. Research design
3.1 Research design and participants 3.2 Procedure 3.3 Measures 3.4 Randomization check 3.5 Manipulation check 3.6 Analytic approach 4. Results 5. Discussion 5.1 Research objectives
5.2 Limitations and future research 5.3 Final conclusion References Appendix 4 7 7 9 9 10 11 12 13 13 14 15 17 18 18 18 21 21 23 25 26 30
4
1 Introduction
Online photo sharing communities have grown at an impressive pace. With 500 million daily active users, the photo sharing app Instagram has grown into one of the most popular social platforms in the world, especially among women between 18 and 30 years old (Aslam, 2018; Sheldon & Brant, 2016). Aiming to make the world feel more connected through mobile photo (and video) capturing and sharing, it provides users an instantaneous way to capture and share their life moments with friends through a series of (filter manipulated) pictures and videos.
This increasing popularity of social media and online photo sharing communities has led to the development of social media not only being used to share personal stories and fun life events, but likewise to share political and societal statements and encouraging people to take action. A 2016 study by Instagram Business indicated that 75% of Instagram users indeed take action after being inspired by a post. These actions include visiting a website, going (online) shopping, or sharing content with friends (Instagram Business, 2016). The combination of these factors makes Instagram a very influential platform - not only for fashion or lifestyle advertisements, but also to draw attention to the before mentioned political or societal issues.
Over the past years, there has been an increase of these societal statements in the form of female empowerment. However, mostly in the form of commodity feminism, which implies that female empowerment is mainly used to make feminism attractive and effective for gaining followers (Goldman, 1993; Gill, 2008). Banet-Weiser and Portwood-Stacer (2017) found that especially on blogs and Instagram, this form of feminism is used. They indicated that since feminism has become a prevalent theme in popular culture, Instagram users use feminism to stay on-trend and, as a consequence, gain more popularity.
5
Due to this trend of female empowerment, it is questionable whether Instagram users really believe in the importance of gender equality, or whether they are just responding to current trends. To be precise, social media influencers, herein defined as “people who possess greater than average potential to influence others due to such attributes as frequency of communication, personal persuasiveness or size of — and centrality to — a social network” (Audrezet, De Kerviler & Moulard, 2017, p.170), may be ‘re-encoding’ feminism to make it work for them: to gain more attention and followers (Pike, 2011).
Celebrity endorsement, a form of advertising that involves a well-known person using their fame to help promote a product or idea (Business Dictionary, 2018), has been frequently used to effectively transfer messages and meanings to audiences (Becker, 2013). However, the continuing enhancement of social media popularity has also led to an increase of social media celebrities or so called social media influencers. This in turn has made a change in the definition of a ‘celebrity’. Hence, it is possible that these same processes take place in endorsements using social influencers, but the academic field currently lacks this information.
While regular celebrities gain their followers because their audience enjoys the movies they play in or the music they make, social influencers gain followers based on a particular interest or lifestyle. They gain followers who share that interest and want to learn more about that subject. Therefore, when they are promoting a certain product that aligns with their interest, it’s more likely to resonate with their audience (Lerario, 2018).
The popularity of these social media influencers hugely affects the way the posted photo is perceived by the audience. At first, a study by Sherman, Greenfield, Hernandez and Dapretto (2018) found that female adolescents were tended to like photos on Instagram that already had many likes than the same photos with less likes. In line with this, Sherman, Payton, Hernandez, Greenfield and Dapretto (2016) investigated ‘the power of the like’ in
6
adolescence. They found that there was greater activity in the hub of reward circuitry in the brain when participants looked at a neutral photo with many likes, compared to when they looked at a neutral photo with few likes. This underlines that regardless of the type or kind of content, popularity of the Instagram posters strongly matters.
Besides, Instagram users with more followers on their accounts were perceived to be more credible and trustworthy. Research by Djafarova and Rushworth (2017) confirmed that people saw the amount of followers of an Instagram user as an indication of how credible they were. A higher amount of followers also led to greater wishful identification: especially women tended to aspire to act and/or behave in the same way as popular Instagram user (Hoffner & Buchanan, 2005). Likewise, Chatterjee (2011) found that profile viewers saw the number of friends or followers users have on a social network site as a signal or cue of quality. On the other hand, too much followers led to a negative evaluation of the user, since participants could not relate with too popular social network users (Tong, Van Der Heide, Langwell & Walther, 2008).
The current study will investigate to what extent popular social influencers are persuasive in changing social attitudes. Hereby, this study focuses on the topic of popular feminism, because of the increasing popularity of female empowerment and the use of social media to express societal relevant topics. This research therefore aims to check for effects of the popularity of Instagram users within the female empowering community on young female adolescent’s attitude towards feminism, and whether this effect is mediated by wishful identification, credibility and/or relatability. In addition, this experiment will check if this effect is moderated by a pre-existing positive attitude towards feminism – since earlier research found that message receivers tended to agree more with a statement when their personal viewpoints were already similar in orientation (Jackson & Darrow, 2005). This leads to the following question: what is the impact of popularity of an Instagram user posting
7
female empowering photos on young female adolescent’s attitude towards feminism, and is this mediated by wishful identification, credibility and/or relatability?
2 Theoretical framework
2.1 The effectiveness of popular social influencers
Since today’s society is said to be obsessed with celebrities, it is not surprising that celebrity endorsement is frequently used as a marketing strategy among many marketers worldwide (Becker, 2013). Not without reason, since Friedman and Friedman (1979) found that celebrities are the most effective endorsers of all endorser types, such as ‘professional experts’ or ‘the typical consumer’.
A celebrity endorser is defined as ‘any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement’ (McCracken, 1989, p. 310). The process of celebrity involvement is described as a psychological course whereby media users react to, think and feel about celebrities to whom they are exposed to (Brown & De Matviuk, 2010). The effectiveness of using celebrity endorsements can be explained by McCracken’s (1989) meaning transfer model. McCracken indicates that celebrities give their personal symbolic meanings to the process of endorsement. This way, the celebrity’s meanings move from the celebrity to the message, and from the message through the consumer (Erdogan, 1999; McCracken, 1989).
There is a high chance that these same processes take place in endorsements using social influencers, but the academic field currently lacks this information. Well-known media users such as celebrities are seen as role models for many people. Because the audience feels as if they do actually know the celebrities, they are considered highly trustworthy, believable, persuasive and likable (Silvera & Austad, 2004). Since social media influencers are found to be even more relatable than traditional celebrities, and are often considered as online friends
8
and/or peers, they may greatly function as successful promoters of products or societal issues and beliefs (Freberg, Graham, McGaughey & Freberg, 2011).
The success of endorsement can be partially explained by Bandura’s (1986) social cognitive theory. Bandura states that individuals can learn vicariously through observation, which may subsequently lead to behavioural adoptation. In case of a famous and/or popular person, identification may play a substantial role in behavioural conformity. When the audience thus relates to the celebrity, they are more likely to copy the celebrity’s behaviour.
Next to identification and trustworthiness, research showed that social media posters with a high amount of followers have more influence on attitudes and behaviour of people who see these posts. For example, Sherman, Greenfield, Hernandez and Dapretto (2018) found that young adolescents were far more likely to like Instagram photographs that already had many likes. And, interestingly, media users with more followers on their accounts were also perceived to be more credibly and trustworthy, especially among young adults.
This can be explained by Bandura’s social cognitive theory. According to this theory, people tend to copy behaviour that is rewarded. When a popular Instagram user thus posts anything related to the topic of feminism and is rewarded for that in the form of many likes and comments, viewers of that particular post are more likely to adopt these views and behaviours.
In addition, most participants in a study by Djafarova and Rushworth (2017) held the common view that ‘[..] number of followers was an indication of a person’s credibility and defined them as being Instafamous’ (p. 11). This is in line with the before mentioned Source Credibility Theory (Hovland, Janis & Kelly, 1953), as media users perceive senders with a high amount of followers or subscribers as more trustworthy and attractive. Hence, when a
9
photo already is popular, people will thus consider the poster as more credible and thus more likeable.
Hence, this led to the following hypotheses: the more popular the Instagram poster behind the female empowering post is, the more the viewer will get influenced by his or her message and the more positive their attitude towards feminism will become (H1) and the more popular the Instagram user behind the female empowering post is, the higher the wishful identification (H2a), credibility (H2b) and relatability (H2c) with the Instagram poster.
Before mentioned research thus indicates that wishful identification, credibility and reliability may function as important underlying mechanisms considering the effectivity of popular social influencers. Former research showed that these three factors predominantly played important roles when examining the persuasiveness of traditional celebrities in the media and on Instagram (Djafarova and Rushworth, 2017; Ohanian, 1990; Silvera & Austad, 2004), when looking at popular Twitter and/or Facebook users (Donath & Boyd, 2004; Jin & Phua, 2014; Tong, Van Der Heide, Langwell & Walther, 2008) and when studying the effects of popular Instagram users on influencing brand and product attitudes (De Veirman, Cauberghe & Hudders, 2017; Evans, Phua, Lim & Jun, 2017; Kudeshia & Kumar, 2017). These mechanisms will therefore be further explained in the following section.
2.2 Underlying mechanisms 2.2.1 Wishful identification
An important mediator may be the identification with the sender. As mentioned before, identification is an important factor in copying observed behaviour: when receivers identify more with a sender, there is a greater chance that the message of the endorser achieves lasting attitudinal or behavioural change (Basil, 1996). However, wishful identification – an
10
individual’s desire to look or act like a specific person (Hoffner & Buchanan, 2005) – may be an even stronger mediator, since it focuses more on the desire of direct imitation of the perceived characteristics and behaviour (Hoffner, 1996). Identification and wishful identification mainly differ in the fact that identification “represents an emotional and cognitive process, whereby the viewer takes on the role of the character she or he is viewing, rather than a desire to make himself or herself more like the character” (Moyer-Gusé, 2008, p. 410).
Hoffner and Buchanan additionally found that participants showed greater wishful identification with same gender characters – and with characters that already had attitudes that were somehow alike to the viewer’s attitude (see paragraph 2.5). Besides, their study also pointed out that women tend to have greater wishful identification with characters that are positioned as successful and popular. Since a high amount of followers and follower engagement indicates a media user to be successful and popular, greater wishful identification is expected among high popular Instagram users. This greater wishful identification, in turn, may explain the influence of popularity of Instagram influencers on attitudes.
This led to the following hypothesis: the higher Instagram users’ level of wishful identification with the Instagram poster, the more they will get influenced by his or her message and the more positive their attitude towards feminism will become (H3).
2.2.2 Relatability
A recent study by Djafarova and Rushworth (2017) showed that non-traditional celebrities (or micro-celebrities) were even more influential than traditional celebrities, because these persons were considered to be more relatable. Non-traditional celebrities are groups as vloggers, bloggers and ‘Instafamous’ personalities, that became famous through their own
11
online media usage. Identification with these non-traditional celebrities is based on admiration and aspiration, but for the most part on recognition and relatability (Kutthakaphan & Chokesamritpol, 2013). The non-traditional celebrity shares a connection with their consumer audience by giving personal accounts and experiences that cause a sympathetic relationship with their peers.
De Vries, Gensler and Leeflang (2012) found that women between 16 and 30 were mainly attracted by messages that were posted by their surroundings or acquaintances. Since they could relate more to these persons, they indicated them as more credible and authentic – which emphasizes the importance of relatability in persuasion.
A higher amount of followers also led to greater wishful identification: especially women tended to aspire to act and/or behave in the same way as popular Instagram user (Hoffner & Buchanan, 2005). Likewise, Chatterjee (2011) found that profile viewers see the number of friends or followers users have on a social network site as a signal or cue of quality. However, too much followers led to a negative evaluation of the user, since participants could not relate with too popular social network users (Tong, Van Der Heide, Langwell & Walther, 2008).
However, the literature on this topic is not completely unambiguous. Whereas there is data proving the effectiveness of a message via relatability (Djafarova and Rushworth, 2017; Kutthakaphan & Chokesamritpol, 2013), a study by Tong, Van Der Heide, Langwell and Walther (2008) implied that too much followers led to a negative evaluation of the user, since participants could not relate with too popular social network users.
This therefore led to the following, non-directional, hypothesis: there is a relationship between Instagram users’ amount of relatability with the Instagram poster, which in turn will affect their attitude towards feminism (H4).
12
2.2.3 Credibility
The Source Credibility Theory by Hovland, Janis and Kelly (1953) indicates that consumers are more likely to be persuaded when the source presents itself as credible. This credibility is divided into three main components measuring the consumer’s perception of the information source: attractiveness, trustworthiness and knowledge in the area of the endorsed product (Ohanian, 1990).
Past literature on celebrity endorsements stated that endorsements are successfully able to persuade consumers and generate engagement for the brand when the source is a credible individual, perceived by consumers as trustworthy and an expert (Erdogan, 1990; Uzunoğlu & Kip, 2014). Research by Djafarova and Rushworth (2017) indicated that in general, consumers do consider non-traditional media users as a trustworthy source of online information. In fact, non-traditional celebrities are regarded as being even more trustworthy, when they have a high amount of followers (Hovland, Janis & Kelly, 1953; Jin & Phua, 2014). Similarly, a study by Westerman, Spence and Van Der Heide (2012) stressed that the perceived credibility of an information source increased with number of followers on social media.
This led to the following hypothesis: the more Instagram users believe in the credibility of the user behind an Instagram post, the more they will get influenced by his or her message and the more positive their attitude towards feminism will become (H5).
2.3 Pre-existing feminist attitudes as a moderator
There are different media effect theories arguing that the strengths of media effects may depend on the receiver’s already existing opinion(s). For example, Valkenburg and Peter’s (2013) differential susceptibility to media effects model (DSMM) states that media effects are conditional, and depend on three types of differential-susceptibility variables:
13
developmental, social and dispositional variables. This latter is defined as “all person dimensions that predispose the [..] responsiveness to media, including gender, temperament, personality, cognitions, values, attitudes, beliefs, motivations, and moods” (Valkenburg & Peter, 2013, p. 226). This model predicts that if your pre-existing values and beliefs match those of the perceived message, you are more likely to be influenced by that particular message.
Hence, if you agree with the actual feminist ideas, popularity may be less useful for triggering wishful identification, relatability or credibility – since these factors will probably already be high. But for young women who are not that into feminism before, a popular post may be influential (while a non-popular post is not). An important moderator for the effect of popularity on relatability, credibility and wishful identification, and finally the total attitude towards feminism, is the participants’ pre-existing attitude towards feminism. To illustrate, research by Brown, Basil and Bocarnea (2003) indicated that celebrities, especially extremely popular figures, have the power to reinforce pre-existing political opinions and more generalized attitudes and orientations toward society.
This led to the following hypothesis: popularity will trigger wishful identification, relatability and credibility when the pre-existing feminist attitude is low or non-existent (H6).
3 Research Design
3.1 Research Design and Participants
This study employed a one-factorial (popularity: high vs low) between-subjects design to test the hypotheses, using an online experiment.
One hundred thirty-two participants were recruited. Of the 132 participants, 31 participants dropped out of the study or did not complete all survey questions. Their data were further excluded from the analyses, just as the 8 men and 2 gender-neutral participants
14
that did not fit the target group of women. Only women were taken into analysis, since Instagram is mostly used by women (Aslam, 2018) and research showed the essential measured factors relatability and wishful identification are more likely to appear when the portrayed character and viewer are of the same sex (Baker & Raney, 2007; Hoffman, 1996; Knobloch-Westerwick & Hastall, 2006). Hence, in total, the final sample consisted of 91 participants (N = 91). Of this sample, 50 participants were in the first condition (low popularity) and 41 in the second condition (high popularity).
The average age of the sample of the participants was 23.22 (SD = 3.67), ages ranged from 17 to 38. In addition, the majority of the sample obtained an academic degree (76.9%). Other obtained degrees were HAVO (1.1%), MBO (3.3%), VWO (5.5%) and HBO (13.2%). Participation was voluntary and no reimbursement was offered. This study obtained ethical approval from the Ethics Committee of the University of Amsterdam (UvA) and all participants gave written informed consent.
3.2 Procedure
Participants were recruited through an online open survey link, which was distributed through Facebook, LinkedIn, Instagram and WhatsApp. The participants were first asked to sign an informed consent. They were randomly assigned to one of the two conditions (low versus high popularity of the Instagram user).
Based on the experimental design the study by Jin and Phua (2014) used, two identical social media influencer profiles were created using stock photos. To maximize internal validity, the Instagram user’s personas, characteristics and (motivational) expressions were based on the profiles of various actual, real-life, Instagram influencers. However, the two Instagram profiles differed in their number of followers, comments and likes to use them as the stimuli (see appendix). The user on the popular profile had 25K followers and between
15
24-68K likes per post, whereas the user on the non-popular profile had 281 followers and between 64-73 likes per post. These numbers were based on the numbers on actual profiles from popular social influencers and ‘average’ similar girls with both the same lifestyle, heritage and characteristics.
Participants were asked to pay attention to an Instagram post that they were about to be exposed to. To ensure exposure to the stimuli, the stimulus was shown to participants for a minimum of 15 seconds and participants were only able to proceed with the questionnaire after this time had passed. After the exposure, participants continued with the questionnaire. This included the following measures: wishful identification, credibility, relatability, attitudes towards feminism and descriptive statistics: gender, age, country of origin and level of education. Finally, the participants were debriefed and thanked for their participation. Participants were recruited within a three week time frame between April and May 2018.
3.3 Measures
Pre-existing attitudes towards feminism The variable pre-existing attitudes towards
feminism was measured using a 5-point Likert scale, measuring items such as ‘Men and women should be treated equally’ (Fassinger, 1994; Morgan, 1996). The average score on this scale was 4.46 (SD = 0.76). A factor analysis was performed with principal axis factoring, which revealed 1 factor with an eigenvalue above 1 (2.88), and an explained variance of 58%. Reliability analysis showed the scale to be reliable (α= .81).
Wishful identification Wishful identification was measured using a 5-point Likert scale and
is defined an individual’s desire to look or act like a specific person (Hoffner & Buchanan, 2005). This is measured using items such as ‘I wish I could be like this influencer’, ‘I wish I could do what these influencers do’, ‘I wish I could look like this influencer’ and ‘I wish I
16
could have as much fun as these influencers have’ (Weintraub Austin, Chen & Grube, 2006; Van Looy, Courtois, De Vocht, De Marez, 2012; Hoffman, Weintraub Austin, Pinkleton & Austin, 2017). The average score on this scale was 2.49 (SD = 0.79). A factor analysis was performed with principal axis factoring, which revealed 1 factor with an eigenvalue of 4.81 and an explained variance of 69%. Reliability analysis showed the scale to be reliable (α= .91).
Credibility Credibility was measured using a 7-point Likert scale and is set as the extent to
which the receiver sees the source as a valid source of accurate and helpful information (Hovland, Janis & Kelly, 1953). This concept is divided into two main components: expertise – the extent to which a source was perceived to have the knowledge about what he or she is posting about, and trustworthiness – the extent to which the source was perceived to be making assertions which he or she personally perceived to be valid (McLaughlin, 2016). Credibility is measured asking participants whether they found the Instagram poster to be for example reliable or trustworthy. Examples of items measuring credibility were ‘To what extent do you think the person behind this profile is reliable’ and ‘To what extent do you think the person behind this profile is trustworthy’ (MacKenzie & Lutz, 1989; Choi & Rifon, 2002). The average score on this scale was 3.80 (SD = 1.31)
A factor analysis was performed with principal axis factoring, which revealed only 1 factor with an eigenvalue above 1. This factor explained 58,4% of the variance. Besides, looking at the factor loadings, there was one item that did not load on this factor (factor loading below 0.40: ‘To what extent do you think this person is repetitious’). Therefore, it was decided to leave out this item, which then ensured a reliable scale (α= .85).
17
Relatability Relatability, in this study, is defined as the degree to which Instagram post
viewers connect with the person behind the Instagram profile (Borum Chattoo & Feldman, 2017; Hoffman, Weintraub Austin, Pinkleton & Austin, 2017; Kleemans, Daalmans, Carbaat & Anschütz, 2018). This variable is measured using using a 7-point Likert scale with statements such as ‘She has the same thoughts and ideas as I have’ or ‘We have a lot in common’ (McCroskey, McCroskey & Richmond, 2006). The average score on this scale was 2.26 (SD = 0.64). A factor analysis was performed with principal axis factoring, which revealed 1 factor with an eigenvalue of 4.34 and an explained variance of 62%. Reliability analysis showed the scale to be reliable (α= .89).
Attitude towards feminism Attitude towards feminism is measured using a 7 point Likert
scale with statements as ‘Women are treated unequally in this society’, ‘Gender equality is an important goal’ and ‘Feminists are too progressive, it does not work in the real world’ (Fassinger, 1994; Morgan, 1996). A factor analysis was performed with principal axis factoring, which revealed 3 factors with an eigenvalue above 1. The scree plot showed a nod right before the 2, so there was assumed to be only one factor. This is therefore a factor with an eigenvalue of 3.70 and a explained variance of 37%. Looking at the factor loadings, there are two items that do not load on this factor (factor loadings below 0.40: ‘Gender stereotyping is a bad thing’ and ‘Feminist ideas may be extreme, but the idea behind it is good’). That is why these two items were taken out. Reliability analysis then showed the scale to be reliable (α= .80). The average score on this scale was 5.49 (SD = 0.78).
18
A randomization check was conducted to detect between-group differences. Independent samples t-tests were used to test age and pre-existing attitude towards feminism, a Chi-square test was used to test gender and educational level.
The experimental groups showed no significant between-group differences with respect to pre-existing attitude towards feminism (t(89) = 0.06; p = 0.95; 95% CI = [-0.30, 0.32]) and educational level (χ2 (4) = 9.12; p = 0.06). There was however a really small between-group difference in age (t(89) = 2.04; p = 0.05; 95% CI = [0.02, 2.89]), since the average age in the low popularity group was slightly higher (M = 3.98, SD = 0.56) than in the high popularity group (M = 3.08, SD = 0.48).
3.5 Manipulation check
In this experiment was attempted to manipulate the popularity of the Instagram user in the sense of amount of followers and follower engagement. After the participants were exposed to the stimulus material and the questions measuring the scores on the mediators, they were asked to estimate the amount of followers of the profile they just viewed, and the amount of likes the photos had on average.
In the low popularity condition the mean estimated follower score was significantly lower (M = 4437.60, SD = 20636.20) than in the high popularity condition (M = 25002.63,
SD = 24271.12) (t(83) = -4.22; p < 0.001; 95% CI = [-30254.05, -10876.02]). In addition, in
the low popularity condition the mean estimated amount of likes was significantly lower (M = 306.22, SD = 970.74) than in the high popularity condition (M = 12014.81, SD = 18508.08) (t(35.15) = -3.70; p = 0.001; 95% CI = [-7976.49, -5440.47]). Therefore, it can be concluded
19
that two different popularity groups were successfully created in this experiment and thus that the manipulation had succeeded.
3.6 Analytic approach
Hypothesis 1, measuring a direct effect of Instagram user popularity on attitude towards feminism, was tested using an independent samples t-test. The other hypotheses were tested using model 4 and model 7 of the PROCESS macro developed by Hayes (2013).
4. Results
To test hypothesis 1, namely that the amount of popularity of the Instagram poster positively predicts the viewer’s attitude towards feminism, an independent samples t-test was conducted with popularity as the independent variable and attitude towards feminism as the dependent variable. The t-test showed no significant direct effect from popularity on attitude towards feminism (t(89) = 0.93, p = 0.62, 95% CI [-.25, 0.41];Mlow= 5.42, SDlow = 0.72; Mhigh= 5.29,
SDhigh= 0.59).
To test hypothesis 2, predicting the indirect effects through the three mediators, Hayes’ (2013) model 4 was used to calculate the effect of popularity on feminism via the three mediators credibility, relatability and wishful identification. The regression models for the prediction by the independent variable, popularity of the social media influencer, was not significant for credibility (F(3, 87) = 1.13, p = 0.34, R2 = 0.037), relatability (F(3, 87) = 0.89,
p = 0.45, R2 = 0.029) or wishful identification (F(3, 87) = 0.79, p = 0.45, R2 = 0.027). In addition, the regression model predicting the outcome variable feminism, precluding any indirect relationships (F(4, 86) = 0.8, p = 0.53, R2 = 0.036) (see Table 1).
20
Regression coefficients for the indirect relationship between popularity and feminism via credibility, relatability and wishful identification
B SE p 95% CI Predictor on mediators Popularity on credibility 0.005 0.28 0.99 -0.55/0.56 Popularity on relatability -0.03 0.14 0.82 -0.30/0.24 Popularity on wishful identification 0.07 0.17 0.681 -0.26/0.403
Mediators on outcome variable Wishful identification on attitude towards feminism
-0.03 0.15 0.84 -0.32/0.26
Credibility on attitude towards feminism
-0.01 -0.07 0.87 -0.16/0.13
Relatability on attitude towards feminism
0.25 0.18 0.16 -0.09/0.61
To test the final hypothesis, namely that the indirect relationship between popularity and feminism via the three mediators would be moderated by pre-existing attitudes towards feminism, model 7 of the PROCESS macro developed by Hayes (2013) was used. This way, conditional indirect effects of popularity on attitude towards feminism through the mediators was measured, with the pre-existing attitudes towards feminism moderating the relationship between popularity and the mediators. A marginally significant interaction was found between popularity and preexisting attitude towards feminism on credibility (b = 0.68, t(1.02) = 2.36, p = 0.08). No signification interactions were found on relatability (b = -0.22, t(0.19) = -1.18, p = 0.24) or wishful identification (b = -0.34, t(0.23) = -1.47, p = 0.15).
21
Table 2 shows the effects for popularity on attitude towards feminism via credibility, relatability and wishful identification, by pre-existing attitude towards feminism. By filling in low, medium and high (based on 16%, 50% and 84% quartiles) values for the moderator, the confidence intervals show there are no significant effects.
Table 2
Indirect effects for popularity on attitude towards feminism via the mediators credibility, relatability and wishful identification, with different values for the moderator of pre-existing attitude towards feminism
Values of moderator Bootstrapped coefficient of indirect effect 95% CI Mediators Credibility Relatability Wishful identification Low -0.006 -0.10/0.06 Medium -0.005 -0.04/0.04 High 0.007 -0.07/0.13 Low 0.033 -0.08/0.17 Medium -0.004 -0.1/0.08 High -0.06 -0.24/0.06 Low -0.01 -0.14/0.09 Medium -0.003 -0.07/0.05 High 0.008 -0.09/0.14 5. Discussion 5.1 Research objectives
22
The objective of this research was to gain insights into the growing importance and presence of social media influencers. This study therefore examined the influence of (perceived) popularity of social influencers making feminist statements by wishful identification, relatability and credibility. While much research had already been conducted around the topic of using traditional celebrities in evoking change in the viewers’ attitude, there was found to be less empirical research on the effect of popularity of social influencers in the form of likes and follower engagement. Research on celebrity endorsement has shown that credibility, relatability and wishful identification were important factors explaining the effects on attitude change. These factors were expected to be even more relevant with regard to social media influencers, who are often described as ‘online friends’ by their audiences and engage with their audience by regularly updating them with the latest information (Liu, Jin, Briones & Kuch, 2012). And since social media influencers are starting to play a huge role in marketing strategies, it is interesting to see what popularity can mean for the viewer’s attitudes towards products or societal discussions (Jarvis, 2006).
Results of this research suggest that the popularity of an Instagram user does not affect the persuasiveness of feminist Instagram messages, which is largely inconsistent with previous research.
The lack of popularity effects may be explained by Tong, Van Der Heide, Langwell and Walther’s (2008) curvilinear U reversed shape relationships between online popularity and social attractiveness. They indicated that the relation between popularity and attractiveness was not valid anymore when it came to computer-mediated communication and social media platforms. This was explained using Brunswik’s Lens Approach, suggesting that “observers interpret artifacts as clues to the behaviors one likely committed, from which personality assessments are inferred” (Tong et al., 2008, p. 542). So when a profile has too many followers, viewers might think that the poster on Instagram spends too much time on
23
the platform, and/or is only successful and communicative in a computer-mediated environment. A study by Donath and Boyd (2004) likewise found that Facebook users with a too high number of friends were called friendster whores by other Facebook users.
While Tong et al.’s study focused on Facebook, this curvilinear relationship could also be applied to this study regarding the relationship between popularity and wishful identification, credibility and relatability. This may explain the non-significant results for the relationship between popularity and these three mediators: too many followers and too much follower engagement may lead to a decrease of social attractiveness. This may in turn lead to a decrease of credibility, relatability and wishful identification, since attractiveness is one of the main components of credibility, identification is mostly based on admiration and attractiveness, and greater wishful identification is experienced with successful and attractive characters (Hoffner & Buchanan, 2005; Kutthakaphan & Chokesamritpol, 2013; Ohanian, 1990). Moreover, Tong et al.’s before mentioned curvilinear u-shape did not effectively show up in this study, since there were only two conditions. To more directly explore the relationship between online popularity and social attractiveness on Instagram, future research should involve more conditions with more and different degrees of popularity, in the sense of amount of followers and follower engagement.
An effect was also expected for popularity on wishful identification, relatability and credibility in the case of a low or non-existent pre-existing feminist attitude. However, this effect was not significantly found. This may be due to the fact that the feminism cues in the stimuli were not explicit enough. Research by Ault et al. (2017) examined important cues that web users pay attention to in order to decide whether an internet page is perceived credible, and found that different cues of credibility within the stimuli had different effects on the final attitude of the website. Hence, it could be that different (more, stronger or more explicit) feminist cues in this study’s stimuli would have caused different outcomes.
24
Moreover, in the case of a low or non-existent pre-existing attitude, the transfer of strong societal or political ideas mostly does not occur within one exposure to a message (Blee, 2002; Liu, White & Dumais, 2010). The participants in this study without a strong pre-existing attitude towards feminism were only exposed to several photos of Instagram profiles with female empowering content, which may could not have affected the final attitude towards feminism. Longitudinal research in which participants without a strong pre-existing attitude towards feminism are longer and more extensively exposed to female empowering content may affect their final attitude.
5.2 Limitations and future research directions
A first limitation is that the experiment in this study was carried out through the online software service Qualtrics. In an online setting, perfect control over respondents’ environment is hardly possible. Especially when it comes to sensitive and personal subjects such as social attitudes and feminism, the distracting and/or social environment the participants where in during the survey should be taken into account. When participants are taking part in a research while being in their own safe environment, they may take the experiment less seriously. In turn, they might feel a greater urge to finish the experiment faster, which might trigger the participants to give emotional and instinctive answers than rational ones – leading to the possibility that they will give more socially common or acceptable answers. Besides, during an online experiment, there is no knowledge of substantial distractions that may have occurred during the experiment, which can affect this study’s validity.
In addition, the study relied on a small and convenient student sample. Of the 132 participants who initially took part in this study, 31 participants dropped out of the study or did not complete all survey questions (24%). Besides, almost all (77%) participants obtained
25
an academic degree and had a Western European background, which is not representative for the total population of women between 18 and 30 years old. Future research should therefore aim to test comparable hypotheses on a larger and more diverse sample of participants.
An additional limitation of this study was the fact that no separate pre-tests were done on how female empowering the stimuli were. The shown content was identical for both the low popularity as the high popularity condition, but the content was not tested beforehand. Future research should therefore aim to pre-test the valid working of the stimuli by asking a separate group of participants about the expected amount or activity of feminism.
Another limitation of the study is that the stimuli were self-constructed. Although the screenshots were securely photoshopped, the Instagram page and photos did not contain the exact features and interaction that would be there on the actual app. There were only 9 photos and 3 zoomed-in, while on Instagram you can scroll endlessly through a feed. The situation and experience of using Instagram was thus not the same as a normal Instagram visit would be. When participants can interactively use Instagram, scroll up and down themselves and have the options to click, zoom in and go back to certain pictures, they will most probably experience the Instagram profile and the user behind it differently, which may affect the study’s results.
One more limitation of the stimuli was that it only involved one (type of) person. Participants’ feelings and social preferences for this specific type of girl and corresponding feed – Lizzy van der Velden is a typical Dutch girl with blonde hair and she has a mainly pink feed – affect their final feminist attitude. Therefore, in future research, different versions (next to difference in amount of followers and follower engagement) of Instagram users posting female empowering content should be used. This can prevent effects for certain participants who do not feel attracted to the outer or/and inner characteristics of the influencer made up in this study.
26
5.3 Conclusion
The present research focuses on one of the most popular and booming marketing forms, that is to say using non-traditional celebrities – also known as social media influencers – to transmit social norms and attitudes to their following audience. The analyses showed no significant effects of popularity on attitude towards feminism, nor any for indirect effects of popularity on attitude towards feminism via relatability, wishful identification and credibility.
Hence, the popularity (amount of followers and follower engagement) of a social influencer on Instagram is of no importance when transferring societal attitudes, particularly in the case of (popular) feminism.
References
Aslam, S. (2018). Instagram by the numbers: stats demographics and fun facts. Retrieved on April 19th, 2018 via https://www.omnicoreagency.com/instagram-statistics/
Audrezet, A., de Kerviler, G., & Moulard, J. G. (2017). Authenticity under threat: when social media influencers need to go beyond passion. In 2017 Global Fashion
Management Conference at Vienna, 170-172.
Ault, M. K., Ness, A. M., Taylor, W. D., Johnson, G., Connelly, S., Jensen, M. L., & Dunbar, N. E. (2017). Ideological lens matters: Credibility heuristics, pre-existing attitudes, and reactions to messages on ideological websites. Computers in Human
Behavior, 68, 315-325.
Austin, E. W., Chen, M. J., & Grube, J. W. (2006). How does alcohol advertising influence underage drinking? The role of desirability, identification and skepticism. Journal of
adolescent health, 38(4), 376-384.
Baker, K., & Raney, A. A. (2007). Equally super?: Gender-role stereotyping of superheroes in children's animated programs. Mass Communication & Society, 10(1), 25-41. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ, US: Prentice-Hall, Inc.
27
the commentary and criticism on popular feminism. Feminist Media Studies, 17(5), 884-888.
Basil, M.D. (1996). Identification as a Mediator of Celebrity Effects. Journal of Broadcasting
& Electronic Media, 40(4), 478-495.
Becker, A. B. (2013). Star power? Advocacy, receptivity, and viewpoints on celebrity involvement in issue politics. Atlantic Journal of Communication, 21(1), 1-16.
Blee, K. (2002). Women of the Ku Klux Klan: Racism and gender in the 1920s. University of California Press, Berkeley, CA.
Borum Chattoo, C., & Feldman, L. (2017). Storytelling for social change: Leveraging documentary and comedy for public engagement in global poverty. Journal of
Communication, 67(5), 678-701.
Brown, W. J., Basil, M. D., & Bocarnea, M. C. (2003). The influence of famous athletes on health beliefs and practices: Mark McGwire, child abuse prevention, and
androstenedione. Journal of Health Communication, 8(1), 41-57.
Brown, W. J., & De Matviuk, M. A. C. (2010). Sports celebrities and public health: Diego Maradona's influence on drug use prevention. Journal of Health
Communication, 15(4), 358-373.
Business Dictionary (2018). Celebrity Endorsement definition. Retrieved via
http://www.businessdictionary.com/definition/celebrity-endorsement.html on June 25th, 2018.
Chatterjee, P. (2011). Drivers of new product recommending and referral behaviour on social network sites. International Journal Of Advertising, 30(1), 77-101.
Choi, S. M., & Rifon, N. J. (2002). Antecedents and consequences of web advertising credibility: A study of consumer response to banner ads. Journal of Interactive
Advertising, 3(1), 12-24.
Djafarova, E., & Rushworth, C. (2017). Exploring the credibility of online celebrities' Instagram profiles in influencing the purchase decisions of young female users. Computers in Human Behavior, 68, 1-7.
Donath, J. and Boyd, D. (2004). Public Displays of Connection. BT Technology Journal,
22(4), 71-82.
De Veirman, M., Cauberghe, V., & Hudders, L. (2017). Marketing through Instagram influencers: the impact of number of followers and product divergence on brand attitude. International Journal of Advertising, 36(5), 798-828.
De Vries, L., Gensler, S., & Leeflang, P. S. (2012). Popularity of brand posts on brand fan pages: An investigation of the effects of social media marketing. Journal of
interactive marketing, 26(2), 83-91.
Erdogan, B.Z. (1999) Celebrity Endorsement: A Literature Review. Journal of Marketing
Management, 15(4), 291-314.
Evans, N. J., Phua, J., Lim, J., & Jun, H. (2017). Disclosing Instagram influencer advertising: The effects of disclosure language on advertising recognition, attitudes, and
behavioral intent. Journal of Interactive Advertising, 1-12.
Fassinger, R. E. (1994). Development and testing of the Attitudes Toward Feminism and the Women's Movement (FWM) Scale. Psychology of Women Quarterly, 18(3), 389-402. Freberg, K., Graham, K., McGaughey, K., & Freberg, L. A. (2011). Who are the social media
influencers? A study of public perceptions of personality. Public Relations
Review, 37(1), 90-92.
Friedman, H.H. & Friedman, L. (1979). Endorser effectiveness by product type. Journal of
Advertising Research, 19(1), 63-71.
Gill, R. (2008). Empowerment/Sexism: Figuring female sexual agency in contemporary advertising. Feminism & Psychology, 18(1), 35-60.
28
Goldman, R. (1992) Reading Ads Socially. London and New York: Routledge
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.
Hoffman, E. W., Austin, E. W., Pinkleton, B. E., & Austin, B. W. (2017). An exploration of the associations of alcohol-related social media use and message interpretation outcomes to problem drinking among college students. Health Communication, 32(7), 864-871.
Hoffner, C. (1996). Children’s wishful identification and parasocial interaction with favorite television characters. Journal of Broadcasting & Electronic Media, 40(3), 389-402. Hoffner, C., & Buchanan, M. (2005). Young adults' wishful identification with television
characters: The role of perceived similarity and character attributes. Media
Psychology, 7(4), 325-351.
Instagram Business (2016). 500.000 adverteerders. Retrieved on April 19th, 2018 via https://business.instagram.com/blog/500000-advertisers
Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion; psychological studies of opinion change.
Jackson, D. J., & Darrow, T. I. (2005). The influence of celebrity endorsements on young adults’ political opinions. Harvard international journal of press/politics, 10(3), 80-98.
Jarvis, J. (2006). Power couple: How bloggers and advertisers can finally hook up.
Advertising Age, 77(19), 42.
Jin, S. & Phua, J. (2014). Following Celebrities’ Tweets About Brands: The Impact of Twitter Based Electronic Word-of-Mouth on Consumers’ Source Credibility Perception, Buying Intention, and Social Identification With Celebrities. Journal of
Advertising, 43(2), 181-195.
Kleemans, M., Daalmans, S., Carbaat, I., & Anschütz, D. (2018). Picture perfect: The direct effect of manipulated Instagram photos on body image in adolescent girls. Media
Psychology, 21(1), 93-110.
Knobloch-Westerwick, S., & Hastall, M. R. (2006). Social comparisons with news personae: Selective exposure to news portrayals of same-sex and same-age
characters. Communication Research, 33(4), 262-284.
Kudeshia, C., & Kumar, A. (2017). Social eWOM: does it affect the brand attitude and purchase intention of brands?. Management Research Review, 40(3), 310-330.
Kutthakaphan, R., & Chokesamritpol, W. (2013). The use of Celebrity Endorsement with the help of electronic communication channel Instagram (Unpublished Master’s Thesis). Mälardalen University School of Business, Society and Engineering, Västerås, Sweden.
Lerario, M. (2018). Influencer marketing: social influencers vs. celebrity influencers. Retrieved on April 23d, 2018, via https://www.mni.com/influencer-marketing-social-vs-celebrity-trends.html
Liu, B. F., Jin, Y., Briones, R., and Kuch, B. (2012). Managing turbulence in the blogosphere: Evaluating the blog-mediated crisis communication model with the American red-cross. Journal of Public Relations Research, 24(4), 353– 370. Liu, C., White, R.W., Dumais, S. (2010). Understanding web browsing behaviors trough
Weibull analysis of dwell time. Paper presented at the SIGIR ’10 Proceedings of the 33rd international ACM SIGER conference on research and development in I
nformation retrieval, New York, NY.
MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural
antecedents of attitude toward the ad in an advertising pretesting context. The Journal
29
McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the endorsement process. Journal of Consumer Research, 16(3), 310-321.
McCroskey, L. L., McCroskey, J. C., & Richmond, V. P. (2006). Analysis and improvement of the measurement of interpersonal attraction and homophily. Communication
Quarterly, 54(1), 1–31.
McLaughlin, C. (2016). Source Credibility and Consumers' Responses to Marketer Involvement in Facebook Brand Communities: What Causes Consumers to Engage?. Journal of Interactive Advertising, 16(2), 101-116.
Morgan, B. L. (1996). Putting the feminism into feminism scales: Introduction of a Liberal Feminist Attitude and Ideology Scale (LFAIS). Sex roles, 34(5-6), 359-390
Moyer‐Gusé, E. (2008). Toward a theory of entertainment persuasion: Explaining the persuasive effects of entertainment‐education messages. Communication
Theory, 18(3), 407-425.
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers' perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39-52.
Pike, K. (2011). ‘The New Activists’: Girls and Discourses of Citizenship, Liberation, and Femininity in Seventeen, 1968–1977. In Mediated Girlhoods: New Explorations of
Girls’ Media Culture (pp. 55–73). New York.
Sheldon, P., & Bryant, K. (2016). Instagram: Motives for its use and relationship to narcissism and contextual age. Computers in Human Behavior, 58, 89-97.
Sherman, L. E., Greenfield, P. M., Hernandez, L. M., & Dapretto, M. (2018). Peer influence via Instagram: Effects on brain and behavior in adolescence and young
adulthood. Child development, 89(1), 37-47.
Sherman, L. E., Payton, A. A., Hernandez, L. M., Greenfield, P. M., & Dapretto, M. (2016). The power of the like in adolescence: effects of peer influence on neural and
behavioral responses to social media. Psychological science, 27(7), 1027-1035. Silvera, D. H., & Austad, B. (2004). Factors predicting the effectiveness of celebrity
endorsement advertisements. European Journal of marketing, 38(11/12), 1509-1526.
Tong, S. T., Van Der Heide, B., Langwell, L., & Walther, J. B. (2008). Too much of a good thing? The relationship between number of friends and interpersonal impressions on Facebook. Journal of Computer‐Mediated Communication, 13(3), 531-549.
Uzunoğlu, E., & Kip, S. M. (2014). Brand communication through digital influencers: Leveraging blogger engagement. International Journal of Information
Management, 34(5), 592-602.
Valkenburg, P. M., & Peter, J. (2013). The differential susceptibility to media effects model. Journal of Communication, 63(2), 221-243.
Van Looy, J., Courtois, C., De Vocht, M., & De Marez, L. (2012). Player identification in online games: Validation of a scale for measuring identification in MMOGs. Media
Psychology, 15(2), 197-221.
Westerman, D., Spence, P. R., & Van Der Heide, B. (2012). A social network as information: The effect of system generated reports of connectedness on credibility on Twitter.
30
Appendix
31
Photo 1 Photo 2
Photo 3 Photo 4
32
Photo 1 Photo 2
