What makes fitspiration inspiring? : exploring social influence and persuasion through Instagram

(1)

What makes fitspiration inspiring?

Exploring social influence and persuasion through Instagram

Master’s Thesis

Graduate School of Communication - Persuasive Communication

Name: Sara Jansen

Student Number: 10564195 Supervisor: Stephan Winter

(2)

WHAT MAKES FITSPIRATION INSPIRING

ABSTRACT

The effects of health promotion via Social Networking Sites has yielded contradictory results, particularly for fitspiration, a currently popular online health promotion trend, this is the case. Therefore, the purpose of this research is to explore how the source type, the source perceptions and the fitspiration type influence health encouragement and social influence effects. The corresponding research questions investigated firstly if the effects of nonprofessional versus professional sources on body image and motivation for behavioral change differed, and to what extent this variability could be explained by perceived similarity. Secondly, the impact of source attractiveness and fitspiration type was researched. The sample consisted of female millennials, aged 18-35. The research was conducted via a 2 (nonprofessional versus professional) x 2 (exercise-related fitspiration versus nutrition-related fitspiration) between-subjects design. Results of the study indicated that, overall, source type, source perceptions and fitspiration type had minimal impacts on the fitspiration effects. Additionally, source perceptions did not mediate source type effects. Only perceived similarity was found to positively impact body image, in conflict with our expectations. A multitude of factors was possibly responsible for the meager effects, amongst others the irrelevance of source professionality concerning social influence, the participants self-confidence, and the perceived extremity of the fitspiration message. The main takeaways from the study are firstly that online health promotion may not be as severely damaging as initially thought. Secondly, online social influence regarding health promotion is complex and multifaceted, contingent on a wide variety of internal and external factors that can both limit and strengthen its impact.

(3)

Introduction

The popularity of social networking sites (SNS) has grown immensely over the past decade (Park & Calamaro, 2013). On such platforms, people can build a social identity by communicating their behaviors, likes and dislikes to others (Hillmann & Trier, 2013). With this self-constructed identity, users convey information about what is socially accepted, which impacts the attitudes and behaviors of people in their social network (Bolton et al., 2013; Fishbein, 1979; Zhou, 2011). For this reason, SNS have become a prominent tool for health promotion (Loss, Lindacher & Curbach, 2014). An example of such an online health encouragement strategy, and the focus of the current research, is fitspiration: a trend in which mostly female Instagram users try to

motivate their followers to engage in a healthier lifestyle by posting nutrition or exercise-related images. Extant literature regarding health promotion strategies via SNS, such as fitspiration, has reported conflicting results. For example, such online health promotion was positively linked to self-efficacy, motivation for behavioral change, as well as diet and exercise adherence (Fox & Bailenson, 2009; Latkin & Knowlton, 2015; Resnick, Orwig, Magaziner, & Wynne, 2002). However, these strategies have also been reported to decrease body image, mental and physical well-being, as well as overall quality of life (Carrotte et al., 2017; Ghaznavi, Jannath, Taylor & Laramie, 2015; Holland & Tiggemann, 2017).

Why prior research has found such contradictory results, and what factors may contribute or interfere, is a still remaining question (Brown & Tiggemann, 2016; Wright, 2016). Message effects can, however, vary depending on the type of source, being either professional or

unprofessional (Simoni, Franks, Levahot & Yard, 2011; Wang, Walther, Pingree & Hawkins, 2008). Furthermore, how such a source is perceived can also impact message effects by evoking

(5)

different processes of social influence (Cialdini & Goldstein, 2004). For example, comparisons to a highly attractive source can lead to feelings of powerlessness over one’s own appearance, decreasing motivation for behavioral change (Bandura & Locke, 2003; Mason, 2012). In contrast, comparisons with a similar source could increase motivation, because observing a similar other succeeding in a task enhances the feeling that you can succeed as well (Bandura, 1977). This indicates that both source type and source perceptions could explain the

contradicting results concerning health promotion.

The current research, therefore, investigates whether the (non)professionality of a source matters, as to effects on body image and motivation for behavioral change. Furthermore, it is investigated to what extent the effects can be explained by the source attribute perceived

similarity. The study also touches upon the impact of perceived attractiveness of the source, and the type of fitspiration, being either exercise or nutrition-related. This is because source

attractiveness, as well as the appearance-focus of exercise-related fitspiration, could be

detrimental to body image (Perloff, 2014; Tiggemann & Zaccardo, 2015). Investigating this topic could provide a better understanding regarding social influence via SNS, in the context of health promotion. Additionally, the outcomes of the current study could potentially contribute to a healthier experience of SNS, by eliciting elements of social influence that can be detrimental to one’s self-evaluation and self-efficacy.

(6)

Theoretical framework

Social Media & Health Encouragement

The popularity of SNS has grown tremendously over the past decade (Park & Calamaro, 2013). Illustrating this is that as of 2016, roughly 80% of the U.S. population had at least one active SNS profile ("Statistics and facts about social media usage", n.d.). SNS are web-based services on which users can create a public profile, establish a list of social connections from fellow users, and interact with these connections (Boyd & Ellison, 2010). These profiles mainly

function to communicate, promote, and validate the user’s social identity (Boyd & Ellison, 2010; Tiggemann & Zaccardo, 2015) by visually showcasing behaviors, likes, and dislikes (Hillmann & Trier, 2013). With this self-constructed identity, SNS users communicate information about what is socially accepted to their followers. In turn, this can alter can alter their attitudes and behaviors (Bolton et al., 2013; Zhou, 2011). This influence caused SNS to become prominent tools of health promotion and encouragement. Plenty of health-related games, interventions, communities and profiles have emerged over the past years (Loss et al., 2014). Among those, is a widespread health promotion trend called “fitspiration” (Tiggemann & Zaccardo, 2015).

Fitspiration attempts to inspire and motivate Instagram users to engage in a healthy lifestyle by promoting healthy nutrition and exercise-related habits and preferences via images. To illustrate the magnitude of the trend, a search of the “fitspiration” hashtag on Instagram returned over 3.3 million images. Moreover, fitspiration images mainly comes from, and is directed to women. Approximately 67% of such messages portrays women and 29% portrays men (Carrotte et al., 2017). Therefore, the current study is female-centered.

(7)

Health Encouragement Implications

Both positive (Latkin & Knowlton, 2015) and negative (Huang, Norman, Zabinski, Calfas & Patrick, 2007) effects of online health encouragement on body image and motivation for behavioral change have been found. For example, the visibility of people who display healthy behavior within one’s online social network should reinforce healthy behavior (Wright, 2016). This influence is due to modelling, the process by which people learn behaviors by observing others successfully perform such behaviors (Fox & Bailenson, 2009). Moreover, online health motivation should enhance the feeling of social support (Kravitz, 2011; Syed-Abdul, Lau & Gabarron, 2016), as SNS enable users to connect, communicate and encourage one another regarding health (Veale et al., 2015). Subsequently, perceived social support increases self-efficacy, which is the extent that someone believes in one’s own ability to succeed (Bandura, 1977). In turn, self-efficacy strengthens motivation for behavioral change, as well as exercise and diet adherence (Resnick et al., 2002).

Health encouragement via SNS has also raised some concerns. Fitspiration has been argued to overlap with “thinspiration” because both trends contain thematically equal content (Boepple & Thompson, 2016; Chasler, 2016). Thinspiration encompasses the portrayal of extremely thin women as motivation to lose weight (Chasler, 2016). Like thinspiration, fitspiration employs very thin people as a source of motivation (Jett, LaPorte, & Wanchisn, 2010). Furthermore, both thinspiration and fitspiration frequently contain guilt-inducing

messages about the food or exercise habits of their viewers (Carrotte, Vella & Lim, 2015). There is a multitude of negative implications of structural exposure to thinspiration, like decreased body image, mental and physical health, and overall quality of life (Ghaznavi et al., 2015). It is

(8)

unsurprising, given the similarity between fitspiration and thinspiration, that fitspiration was found to result in comparable negative effects (Boepple & Thompson, 2016). For example, fitspiration can directly decrease mood, body satisfaction and self-esteem (Tiggemann & Zaccardo, 2015) because women compare themselves with, and feel the need to live up to the portrayed beauty standards (Brown & Tiggemann, 2016; Tiggemann & Zaccardo, 2015; Zuo, 2014). These standards, however, are highly unrealistic since fitspiration contains flattering and often edited or photoshopped images, strengthening the aforementioned negative effects (Brown & Tiggemann, 2016). Furthermore, exposure to fitspiration enlarges the likelihood of developing disordered eating, because it promotes an extreme attitude towards body, food and exercise, stimulating an unhealthy relationship with the latter (Adams, 2014; Holland & Tiggemann, 2017). A commonly used motivational quote, illustrating this extremity is“Fit is the new skinny” ("5 Fitness Quotes That Are The Absolute Worst", 2013). Quite recently, online health

promotion was also linked to a relatively new form of disordered eating, named “orthorexia” (Kiberd, 2015; Marsh & Campbell, 2016), which is an extreme fixation on healthy living and eating. Orthorexia leads to the elimination of necessary calories and food groups, and excessive exercising (Koven & Abry, 2015).

The Perceptions of Source Types

There are two types of health encouragement sources, namely professional (e.g. a licenced fitness coach) and nonprofessional (an average SNS user), also called laypeople (Hu & Sundar, 2010; Neubaum & Krämer, 2015; Schiavo, 2013). In health communication, people traditionally prefer the opinions of professionals as they are thought to possess more expert knowledge

(9)

(Cialdini & Goldstein, 2004; Klucharev, Smidts & Fernández, 2008). On social media, however, users commonly prefer nonprofessional, peer-to-peer opinions (Constantinides, 2013; Taquini, 2016). SNS users can, in this digital era, freely and actively seek out and exchange information, increasing the power of user-to-user communication. This makes users not only more critical of message content, but also more trusting of user generated content and distrusting of information from professionals (Hum, 2014; Lenderman, 2006; O’Brien, 2011).

The preference towards laypeople can also be attributed to perceived similarity. Relative to professionals, laypeople evoke more perceived similarity, a concept which explains the degree to which someone feels to share similar opinions, appearance, or personality traits, as well as a similar background or lifestyle with a source (Simoni et al., 2011; Wissler, Cialdini &

Schweitzer, 2002). Laypeople are perceived to share more characteristics with a message

receiver, because people feel they possess more common values, beliefs, and identities with their peers (Hawkins, Kreuter, Resnicow, Fishbein & Dijkstra, 2008; Taquini, 2016). They are also thought to have comparable resources, abilities, and a more similar lifestyle (Fardouly,

Diedrichs, Vartanian & Halliwell, 2015; Jones, 2001). Additionally, they are more relatable; the information they share reflects health-related experiences in the context of everyday life (Korp, 2006). Drawing from the former section, it can be hypothesized that nonprofessional fitspiration sources are perceived to be more similar, relative to professional fitspiration sources. Based on this, the following hypothesis was derived.

H1: Nonprofessional sources are perceived to be more similar in comparison to professional sources

(10)

motivation for behavioral change remains indefinite. This uncertainty can be explained by the social comparison theory, a concept that describes how people construct self-evaluations, based on comparisons to others in their social network (Festinger, 1954). A key element in the social comparison theory is the similarity hypothesis, predicting that people generally tend to compare themselves more to people similar, rather than dissimilar, to themselves(Wood, 1989).

Particularly for adolescents, perceived similarity concerning abilities, resources and lifestyles spark social comparisons regarding attributes such as weight and social status (Jones, 2001). This connects to the idea that people prefer comparing themselves to similar individuals because it provides better diagnostic information (Strahan, Wilson, Cressman & Buote, 2006). Simply put, when a source is perceived to be capable of similar things, their appearance and social status feels like a more relevant comparison source (Fardouly et al., 2015). These comparisons can evoke both assimilation and contrast effects regarding body image. For instance, comparing oneself to an attractive source who shares similar abilities, resources and lifestyle features, can lead to assimilations of self-evaluations, positively impacting one’s body image (Festinger, 1954; Corcoran et al., 2011; Wanic, 2011). In contrast, such comparisons may also negatively impact body image via the contrast effect; how people construct negative self-evaluations, after

comparison to a source who is superior on a certain attribute (Corcoran, Crusius & Mussweiler, 2011; Wheeler & Suls, 2007). This contrast effect results from social upward comparisons; the process of comparing oneself to another who is evaluated as more attractive (Lewallen & Behm-Morawitz, 2016). Considering that fitspiration sources are generally more attractive than most, it is likely for them to be regarded as superior concerning appearance and evoke social upward comparisons and, in turn, the contrast effect (Corcoran et al., 2011; Jett et al., 2010). It has been cited time and time again that social upward comparisons and the contrast effect lead to negative

(11)

ratings of self-attractiveness (Perloff, 2014; Pila, Stamiris, Castonguay & Sabiston, 2014; Tiggemann & Zaccardo, 2015). Perceived similarity, therefore, negatively impacts body image because (i) it facilitates social comparisons (ii) that feel highly relevant to the message receiver (iii) which causes a contrast effect, (iiii) since fitspiration sources are highly attractive.

Perceptions of similarity to a fitspiration source can simultaneously impact motivation for behavioral change positively. People evaluate others more positively when they perceive them to possess traits that are similar to their own (Byrne, Bond, & Diamond, 1969; Byrne, Griffitt, & Stefaniak, 1967). These familiar and positive perceptions of a source decrease scrutiny of messages they produce, increasing the source’s persuasiveness (Garcia-Marques & Mackie, 2000; Packard, Gershoff & Wooten, 2016). Additionally, perceived similarity is an important factor in improving self-efficacy (Schunk,1987; 1995). In turn, self-efficacy plays a major role in motivation for behavioral change. Comparing oneself to a similar, succeeding person enhances the message recipient’s belief that he or she can also succeed (Schunk, 1995). A fitspiration source that is perceived as similar to the message recipient should, thus, be able to positively impact motivation for behavioral change.

Since laypeople should be perceived to possess more similar features than professionals, assumptions about source type effects can be made. To be more precise, the greater similarity of a layperson increases their relevance as a subject for social upward comparisons and,

subsequently, enhances the likelihood of a contrast effect to occur (Corcoran et al., 2011; Fardouly et al., 2015; Hawkins et al., 2008; Jones, 2001). Therefore, the negative effects of similarity on body image should be more pronounced in the nonprofessional condition.

(12)

Therefore, the positive effects of perceived similarity on motivation for behavioral change should be more pronounced in the nonprofessional condition. Based on the former section, the following hypotheses were constructed.

H2: Perceived similarity negatively impacts body image

H3: Perceived similarity positively impacts motivation for behavioral change

H4a: Nonprofessional sources impact body image more negatively in comparison to professional sources

H4b: The negative impact of nonprofessional sources on body image is mediated by perceived similarity

H5a: Nonprofessional sources impact motivation for behavioral change more positively in comparison to professional sources

H5b: The positive impact of nonprofessional sources on motivation for behavioral change is mediated by perceived similarity

The Role of Attractiveness

Another source perception attribute, perceived attractiveness is another element that can have significant effects on body image and motivation for behavioral change. However, as theoretical assumptions cannot be made regarding which source type would be rated as more attractive, the effects of this source attribute will be treated as a separate research question.

Rating of a source’s attractiveness have been found to significantly impact the

persuasiveness of the message (Brinol & Petty, 2009) since people commonly identify with, or aspire to identify with attractive sources, subsequently leading to message conformity (Kelman, 1958) and motivation for behavioral change (Wilson & Sherrell, 1993). More recent literature in the context of health promotion, however, has highlighted several unwanted effects of perceived attractiveness. In fact, identification with an attractive health promotion source was found to be detrimental to one’s body image (Perloff, 2014; Tiggemann & Zaccardo, 2015). This body dissatisfaction results predominantly from social upward comparisons with a source (Lewallen &

(13)

Behm-Morawitz, 2016; Perloff, 2014) and internalization of the thin-ideal, promoted in health encouragement (Van Vonderen & Kinnaly, 2012, Watson, Livingston, Cliette & Eaton, 2015). The more frequent the exposure to these attractive sources, the more likely it is for body

dissatisfaction to occur (Grabe, Ward & Hyde, 2008; Harper, Sperry & Thompson, 2008) since repeated exposure heightens message persuasion and internalization (McHugh, & Chrisler, 2015). Attractive health promotion sources can also negatively impact motivation for behavioral change. High levels of perceived attractiveness induce a sense of inadequacy and, ultimately, powerlessness over a woman’s perceived ability to change her appearance (Mason, 2012;

Peterson, Grippo & Tantleff-Dunn, 2008). Since it is common for people to shun tasks when they are perceived to exceed their coping capabilities, this induced powerlessness negatively affects motivation for behavioral change (Bandura & Locke, 2003). Based on the aforementioned section, the second research question investigates whether source attractiveness negatively impacts body image and motivation for behavioral change.

Fitspiration Type

A number of studies have indicated that viewing appearance-focused media content increases the importance for women to meet the thin-ideal (Simpson & Mazzeo, 2016) and consequently increases body dissatisfaction (Stanley, Barnes & Short, 2015, Tiggemann, 2006). The reason for this is that appearance-focused contents on SNS evoke social upwards comparisons, and the detrimental effects on body image that follow (Perloff, 2014; Tiggemann & Zaccardo, 2015). Particularly Instagram facilitates these effects because of the platform’s visual nature. Images lend a certain “mythical truth” (Marwick, 2015) to the thin-ideal portrayed in fitspiration, which

(14)

or figure (Fardouly et al., 2011). In the case of fitspiration, images evoke the idea that being thin is highly important since they commonly contain and promote this ideal. Both the imagery of such an idea, and the truth that people attribute to it can lead to internalization of such message content. In turn, this internalization of the thin-ideal is connected to an increase in body

dissatisfaction (Watson, Wells & Hudson, 2011).

Fitspiration mainly manifests in two categories, namely nutrition and exercise-related content. For example, a search for the hashtag “cleaneating” (nutrition-related) showed over 28 million results, and the hashtag “exercise” (exercise-related) generated over 25 million results. Exercise-related fitspiration, however, was observed to be more appearance focused than nutrition-related fitspiration, containing more thin and fit bodies, and more nudity. Although there is no specific literature describing the impact of these visual differences in fitspiration posts, it can be hypothesized that the appearance focus in the exercise condition results in more detrimental effects on body image. This assumption lead to the following hypothesis.

H6: The negative fitspiration effects on body image are more pronounced in the exercise condition, in comparison to the nutrition condition

(15)

Figure 1. Conceptual model

Method

Research Design

To answer the research question about whether the effects of professional versus nonprofessional fitspiration sources differ, and whether variability can be explained by perceived similarity and perceived attractiveness, a 2 (professional versus nonprofessional) x 2 (exercise message versus nutrition message) between-subjects research design was employed. The independent variable comprises the type of source with two additional levels, being either exercise-related or nutrition-related fitspiration, whereas there are two dependent variables, being body image and motivation for behavioral change. Furthermore, perceived similarity was treated as a mediating variable. The second research question, instead, investigates the effect of perceived attractiveness on body image and motivation for behavioral change, to uncover whether a similar mediating relationship

(16)

can be observed. Therefore, the current research consists out of four different conditions.

Table 1.

Conditions

Source Type Fitspiration Type Condition Professional Exercise-related 1

Professional Nutrition-related 2 Nonprofessional Exercise-related 3 Nonprofessional Nutrition-related 4

Research units & Selection research units

The target group for the current research is female millennials, more specifically, women aged 18 to 35. Across extant research there is a wide variety of age ranges attributed to the millennial-cohort. The chosen age range, however, was based on current literature about millennials’ social media behavior, as this is most in line with the current research topic (Rodney & Wakeham, 2016). Female millennials are appropriate to target for several reasons. Firstly, millennials are highly active on, and interconnected via SNS (Winograd & Hais, 2011). Additionally,

fitspiration predominantly targets women, who are also more likely to be affected by health and appearance-related social media messages (Lewallen & Behm-Morawitz, 2016; Tiggemann, 2015). The participants were selected through a non-probability convenience sampling procedure

(17)

in which female students were drawn from a population of undergraduate students, enrolled in the Erasmus University Rotterdam. The recruitment took place at the cafeteria of the university, increasing the likelihood of reaching large groups of female millennials without interfering with in-class hours. Furthermore, the recruitment procedure also included a snowball sampling strategy, through which participants were firstly recruited, and secondly asked to share the survey with their peers via SNS, being Twitter and Facebook. The snowball sampling strategy increases the likelihood of obtaining a non-probability sample, as the survey link could be distributed to groups outside of the researcher’s own social network.

Sample

In total there were 194 respondents, ranging from 18 to 77 years old. From these respondents 4.7% were male (N=9) and 79.6% female (N=183) (M=0.95, SD=0.21). Since the current study focuses on female millennials, males (N=9) as well as respondents above the age of 35 (N=3) were excluded from the sample. Furthermore, participants who refused consent for their data to be used for research purposes were also excluded from the survey (N=2), leaving a total of 181 respondents. There were four experimental conditions to which participants were randomly assigned, 1- professional x exercise-related fitspiration, 2- professional x nutrition-related fitspiration, 3- nonprofessional x exercise-related fitspiration, and 4- nonprofessional x nutrition-related fitspiration. From the total sample, 25.4% resided in the professional x exercise-nutrition-related fitspiration condition (N=46), 23.8% in the professional x nutrition-related fitspiration condition (N=43), 22.7% in the nonprofessional x exercise-related fitspiration (N=41), and 28.2% in the nonprofessional x nutrition-related fitspiration condition (N=51). Moreover, there were four

(18)

different categories regarding the frequency of Instagram use. Established through a multiple choice question items, with the predefined answers: 1-Never, 2-Monthly, 3-Weekly, 4-Daily. From the valid sample, 11% never used Instagram (N=20), 7.7% used Instagram on a monthly basis (N=14), 13.3% percent used Instagram on a weekly basis (N=24) and 68% used Instagram on a daily basis (N=123) (M=3.38, SD=1.03).

Materials

For each of the conditions, a different Instagram post was constructed accordingly, with the help of Adparlor, an online Instagram advertisement generator ("AdParlor Mockup Generator", n.d.). There were several indicators to illustrate the (non)professionality of the source. For the

professional source, the profile name “AnytimeFitness_official”, a supposedly professional fitness coach, rather than a public health institution, was chosen to communicate expertise of the source with regard to exercise and nutrition for the following reasons. In the current

environment, fitness coaches are perceived as valuable and credible assets that take up a multitude of roles, under which the role of nutritionist, personal life consultant and weight controller (De Lyon, Neville, & Armour, 2016). Furthermore, an Instagram search exposed that public health organizations rarely engage in Fitspiration. The profile picture of the professional source contains a logo of the Anytime Fitness gym, also indicating the credibility of the source (see Appendix 1A). Additionally, the description of the photo contains a link to the gym’s

website and Hashtags such as “gymcoach”, “health” and “energy”, also signaling professionality. For the nonprofessional source, the chosen profile name was ‘Kaylee_1991_Xx’ to communicate that the participant is dealing with a “normal” layperson. The profile picture of the

(19)

nonprofessional source shows a selfie, seemingly taken at a party. On SNS, it is very common for ordinary users to use such a photo as a profile picture. Therefore, it could be a sufficient indicator of non professionality. Moreover, the photo description and Hashtags use exclamation marks, capitals, slang and abbreviations like ‘fitspo’, ‘fitgirl’ and ‘workit’, also signaling non-professionality (see Appendix 1C). In the professional condition, a link to a website of a gym was posted to communicate expertise and professionality. As for the Fitspiration type, being either nutrition or exercise-related, two different images were chosen, which were used for both the professional and nonprofessional condition to ensure internal validity. The exercise-related Fitspiration image shows a fit, blonde girl in a gym (see Appendix 1B). The nutrition-related Fitspiration image shows, again, a fit, blonde girl who is eating healthy food (see Appendix 1D). Both images were found on Google and contained a person sharing similar appearance features to decrease the chances of visual elements interfering with the social influence these sources may exert over participants.

Procedure

The respondents participated in an online experiment which could be accessed via any mobile device. First, participants were obliged to provide their consent regarding the nature of the study, and that the collected data is used for research purposes. After providing consent, candidates were randomly assigned to one of the four different conditions. Respondents were then exposed to the manipulated fitspiration material, about which they were required to answer a set of questions that exposed their opinions about the fitspiration source’s attractiveness and similarity. Following this, the survey contained five questions, which functioned as a distraction of the

(20)

actual research purpose to prevent biased answers. Subsequently, several questions were posed through which the participants’ motivation for behavioral change and ratings of body image were exposed. Before completion of the survey, a set of demographic questions followed, consisting of gender, age and frequency of Instagram use. Furthermore, the research subjects were given the option to fill out their e-mail address, for the chance to win a Bol.com gift cards. Three gift cards of ten Euro’s were randomly given out in total. The survey was concluded with a short

debriefing, and a thank you from the researcher.

Measures

In total, there were 27 question items, all employing a 7-point likert scale with predefined answers to the questions, ranging from 1- strongly disagree, to 7- strongly agree (for interview guide, see Appendix 2). Question items which were exactly the same across all conditions were added to form the sum of one single item. For the construct attractiveness, for example, all “I think this person is attractive” questions were added, together forming the sum of question item one. Subsequently, these sums of question items were merged to form a scale, of which the unidimensionality and reliability was then tested (for an illustration of this procedure, see table 2 below).

(21)

Table 2.

Procedure Scale Construction; sum of question items (QI) across conditions (C)

C1** C2 C3 C4 Sum QI1* I think this

person is attractive + I think this person is attractive + I think this person is attractive + I think this person is attractive = Sum attractiveness item 1 QI2 I think this

person is classy + I think this person is classy + I think this person is classy + I think this person is classy = Sum attractiveness item 2 QI3 I think this

person is sexy + I think this person is sexy + I think this person is sexy + I think this person is sexy = Sum attractiveness item 3 QI4 I think this

person is elegant + I think this person is elegant + I think this person is elegant + I think this person is elegant = Sum attractiveness item 4 Sum item 1 + 2 + 3 + 4 = Attractiveness Scale Attractiveness

Participants’ perception of the source’s attractiveness was assessed using four items based on an existing scale constructed by Ohanian (1990) (see Appendix 3A) and used a 7-point Likert scale (1- Strongly disagree, 7- Strongly agree). For example, one question item in Ohanian’s (1990) scale asked: “I think this person is..”, with answers ranging from 1-attractive, to 7-unattractive. An exploratory factor analysis indicated that the scale was unidimensional, with a single

component with an Eigenvalue above 1.00, explaining 58.68% of the variance in the four items. The 4-item scale also showed to be reliable, as indicated by a Cronbach’s Alpha of .76 (M=4.58,

(22)

SD=1.03), which exceeds the minimum requirement of .7 (Hair, Anderson, Tatham & Black,

1998).

Perceived Similarity

Participants’ perception of the source’s similarity was assessed using four items based on an existing scale, constructed by Kozar (2012) (see Appendix 3B) and used a 7-point Likert scale (1-Strongly disagree, 7- Strongly agree). Results of an exploratory factor analysis showed that the scale was unidimensional, because there was only one component with an Eigenvalue above 1.00, explaining 55.82% of the variance in the four items. Furthermore, the 4-item scale was found to be reliable, with a Cronbach’s Alpha of .72 (M=2.91, SD=1.05).

Distraction Questions

Furthermore, a couple of self-constructed distraction questions were employed, to help disguise the true purpose of the study (Appendix 3C). Like the other question items, the questions used a predefined set of answers (1- Strongly disagree, 7- Strongly agree).

Motivation for Behavioral Change

Participants’ motivation for behavioral change was assessed using three items, based on Ajzen’s (2006) intention measurement items (see Appendix 3D) and used a 7-point Likert scale (1-

Strongly disagree, 7- Strongly agree). An exploratory factor analysis showed that the scale was

unidimensional one component had an Eigenvalue larger than 1.00, explaining 74.04% of the variance in the three items. The 3-item scale was found reliable, since the Cronbach’s Alpha

(23)

scored .82 (M=5.25, SD=1.14).

Body Image

Participants’ perception of their own body image was assessed using three items based on “The Multidimensional Body-Self Relationship Questionnaire (Cash, Fleming, Alindogan, Steadman & Whitehead, 2002) (see Appendix 3D) and used a 9-point Likert scale (1- Extremely

Dissatisfied to 9- Extremely satisfied). The scale was suggested to be unidimensional with an

exploratory factor analysis, as only one component had an Eigenvalue above 1.00, explaining 86.17% of the variance in the three items. The 3-item scale also showed to be reliable, which was suggested by a Cronbach’s Alpha of .92 (M=5.84, SD=1.92).

Results

Data Analysis

The process of collecting data started on the 19th of April, 2017 and ran until the 15th of May, 2017. The minimum amount of respondents for this research was roughly 150. In total, 230 participants were acquired. Before analyzing the gathered data, unfinished surveys and surveys by respondents that did not comply with the sample requirements were taken out. After this, filtering out 49 surveys, 181 surveys remained to be analyzed via SPSS.

Source Type & Perceived Similarity

(24)

which the participants were exposed, an independent samples t-test was conducted. More

specifically, the means of both source type categories (nonprofessional versus professional) were compared. Results indicated that the mean regarding perceived similarity, displayed in the nonprofessional condition (M=2.77, SD=1.00) did not significantly differ from the professional condition (M=3.06, SD=1.09) , t(179)= 1.87, p= .06. This means that nonprofessional sources are not seen as significantly more similar in comparison to professional sources. In fact, professional sources were even found to be more similar than nonprofessional sources, opposing what was expected. Therefore, H1 is not supported.

Perceived Similarity, Body Image & Motivation for Behavioral Change

To test H2 and H3, and see whether perceived similarity significantly predicts body image and motivation for behavioral change, two separate linear regression analyses were conducted. Firstly, results of the linear regression analysis showed that there was a moderate significant relationship between perceived similarity and body image (F(1,179)55.22, b= .50, p< .001). The

model explains 23.6% of variance in body image (R²=.236, adjusted R²=.231). The participant’s predicted body image is equal to 3.246 + .892 (perceived similarity). For each unit with which perceived similarity increases, body image increases with .892. Contradicting the initial

proposition, this suggests that the more similar a consumer perceives herself to be to the source, the more positive she will evaluate her body image. Therefore, H2 is not supported.

Second, results of the linear regression analysis showed that there was no significant relationship between perceived similarity and motivation for behavioral change (F(1,179)3.519,

(25)

This means that perceived similarity does not impact motivation for behavioral change positively. Therefore, H3 is not supported.

Source Type, Fitspiration Type, Body Image & Motivation for Behavioral Change

To test H4a, H5a, and H6, to see whether scores of body image and motivation for behavioral change differed, depending on the type of source, or the type of fitspiration to which the participants were exposed, two separate two-way ANOVA’s were conducted.

Firstly, results from the two-way ANOVA with source type and fitspiration type as independent variables and body image as dependent variable revealed no significant main effect of source type, F(1, 177)=.43 p=.51, η2= .00, nor a significant main effect of fitspiration type, F(1, 177)= 1.44, p= .88, η2= .00.

Second, results from the two-way ANOVA with source type and fitspiration type as independent variables and motivation for behavioral change as dependent variable revealed no significant main effect of source type, F(1, 177)=1.15 p= .29, η2= .01, nor a significant main effect of fitspiration type, F(1, 177)= .03 p= .87, η2 = .00. Nonprofessional sources, thus, do not impact body image more negatively in comparison to professional sources, nor do they impact motivation for behavioral change more positively. Therefore, H4a and H5a were not supported. Additionally, exercise-related fitspiration did not have a significantly more negative effect on body image than nutrition-related fitspiration.Therefore, H6 was not supported.

Mediation analysis

(26)

mediation effect to occur was not met. This means that perceived similarity could not be held accountable for the effect of source type on body image and motivation for behavioral change. However, the indirect effect was still tested using a bootstrap estimation approach with 10000 samples. Results indicated that there was, indeed, no significant indirect effect of source type on body image through perceived similarity, ab= -.26, BCa CI [-0.55, 0.00]. There was also no significant indirect effect of source type on motivation for behavioral change through perceived similarity, ab= -,04, BCa CI [-0.14, 0.00]. Therefore, H4b and H5b were not supported.

Figure 2. Mediation Body Image

(27)

Perceived Attractiveness, Body Image & Motivation for Behavioral Change

To test the second research question, and see whether perceived attractiveness significantly predicts body image and motivation for behavioral change, two linear regression analyses were conducted. Firstly, results of the linear regression analysis showed that there was no significant relationship between perceived attractiveness and body image (F(1,179)0.082, b= -.02, p=.78). The model explains 0.00% of variance in body image (R²=.000, adjusted R²=-.005). This means perceived attractiveness of a source does not negatively impact body image.

Second, results of the linear regression analysis showed that there was no significant relationship between perceived similarity and motivation for behavioral change (F(1,179)0.009,

b= .01, p=.92.The model explains .00% of variance in motivation (R²=.000, adjusted R²=-.006).

This means that perceived attractiveness of a source has no significant negative effect on motivation for behavioral change.

Discussion

Discussion & Conclusion

The goal of this research was to uncover whether the impacts of health encouragement on body image and motivation for behavioral change differed, depending on source type, source

perceptions and fitspiration type. Therefore, a number of statistical analyses was conducted to answer the hypotheses and, subsequently, the research questions. The first hypothesis

investigated whether the nonprofessional source was regarded to be more similar, in comparison to the professional source. In contrast to the literature, describing how laypeople are thought to share more relatable and comparable values, beliefs, identities, abilities and lifestyles (Hawkins

(28)

et al., 2008; Jones, 2001; Korp, 2005; Simoni et al., 2011; Wissler et al., 2002), results indicated that nonprofessional sources were not perceived as significantly more similar. Based on this, the literature is not per se proven to be inaccurate. The current results do, however, suggest that mere cues of non-professionality concerning health may not be enough to evoke feelings of similarity, or that they are outweighed by other factors that decrease the nonprofessional source’s similarity.

The second hypothesis tested whether perceived similarity negatively impacts body image. Results showed that there was, in fact, a significant relationship. However, this relationship was positive instead of negative, suggesting that perceived similarity positively impacts body image. This result conflicts with the literature, describing that perceptions of similarity with an attractive source spark social upwards comparisons that, in turn, negatively impact body image via the contrast effect (Corcoran et al., 2011; Strahan et al., 2006; Tiggemann & Zaccardo, 2015; Wood, 1989). The results of the current research are more in line with the idea that perceived similarity leads to positive self-evaluations, through the process of

assimilation (Corcoran et al., 2011; Festinger, 1954; Wanic, 2011). Perhaps this can be explained by a number of moderators, which were regarded to go beyond the scope of the current research, which influence whether assimilation or contrast effects occur. The participant’s already

existing self-evaluation, as well as the extremity of the promoted standard, are examples of such moderators (Corcoran et al., 2011). For example, it could be the case that participants in this study were relatively self-confident, or did not perceive the promoted beauty ideal as very extreme, weakening the negative effects on body image. This indicates that similarity effects do not apply unconditionally, but are dependent on personal characteristics of the message receiver, as well as contextual factors. Furthermore, the unexpected results could imply that the employed question items were not sufficient to measure perceived similarity. The items lack a focus on

(29)

resources and abilities and are, therefore, not corresponding with the literature, possibly causing participants to respond differently than expected.

In order to investigate H3, it was examined whether perceived similarity positively impacts motivation for behavioral change. Results showed that perceived similarity did not significantly predict changes in motivation for behavioral change and, therefore, results conflict with the literature. Perceived similarity should decrease message scrutiny and subsequently reinforce the persuasiveness of the source (Garcia-Marques & Mackie, 2000; Packard et al., 2016). Furthermore, perceived similarity of a succeeding source should enhance the message recipient’s self-efficacy, and, subsequently, motivation for behavioral change (Schunk, 1995). The insignificant results do not necessarily invalidate the literature, but imply that on SNS, perceptions of similarity are not sufficient to decrease message scrutiny and increase self-efficacy. Perhaps this is related to how, via SNS, people have become increasingly empowered, and therefore critical of message content (Lenderman, 2006; O’Brien, 2011). As people become more critical, it might be more difficult to persuade them to engage in new ways.

The fourth and fifth hypotheses examined if nonprofessional sources impact body image more negatively, and motivation for behavioral change more positively in comparison to

professional sources. Furthermore, it was researched whether perceived similarity was accountable for this effect. No significant relationship between source type, body image and motivation for behavioral change, nor an interaction effect with perceived similarity was found. Thus, current results do not correspond with the literature, which argues that nonprofessional sources should impact body image more negatively, and motivation for behavioral change more positively in comparison to professional sources, as they are considered to be more similar to the

(30)

Since nonprofessional sources were not perceived as more similar than professional sources, it is no surprise that no direct, nor interaction effect was found, considering the pivotal role of

similarity. Current results may also indicate that, in the context of SNS, (non)professionality is an irrelevant factor in whether someone is likely to engage in social upward comparisons, or whether someone’s self-efficacy is likely to be increased.

The second research question investigated whether perceived attractiveness negatively impacts body image and motivation for behavioral change. The results revealed nonsignificant relationships, meaning that perceived attractiveness does not predict changes in body image or motivation. These results conflict with the literature, describing that source attractiveness negatively impacts body image via identification and social upward comparisons (Lewallen & Behm-Morawitz, 2016; Perloff, 2014; Tiggemann & Zaccardo, 2015; Van Vonderen & Kinnaly, 2012), and motivation for behavioral change through enhanced powerlessness and decreased self-efficacy (Bandura & Locke, 2003; Mason, 2012; Peterson et al., 2008). The insignificant results can be explained by how message effects are typically larger when the message is

repeated, as this reinforces message persuasion and internalization (Grabe et al., 2008; Harper et al., 2008; McHugh, & Chrisler, 2015). A single exposure to the thin-ideal in the current research, according to the aforementioned theory, could thus not have been enough to cause a significant negative effect on body image and motivation for behavioral change. The literature is thus not per se incorrect, however, message repetition may be necessary for source attractiveness to have a meaningful impact.

The sixth hypothesis tested whether negative fitspiration effect on body image was more pronounced in the exercise, versus the nutrition condition. Ultimately, no significant difference between the two conditions was found. This means that exercise-related fitspiration did not

(31)

impact body image more negatively than nutrition-related fitspiration. Quite possibly this is, again, a repetition-related issue. The thin-ideal, prominent in the exercise condition, should increase body dissatisfaction via the imagery of this ideal, and its internalization that follows (Watson et al., 2011). Message internalization can, subsequently, be strengthened by repetition (McHugh, & Chrisler, 2015). Therefore, the aforementioned literature is not invalid, but possibly dependent on repetition. Simply put, significant results of appearance-focused content may become prominent if internalization of the message is enhanced by message repetition.

The findings shed light on the research questions, regarding the responses to health promotion source types, the mediating role of perceived similarity, and the effects of source attractiveness as well as fitspiration types. Results indicated that there was no significant difference in how either a nonprofessional or a professional source affected body image and motivation. Furthermore, although perceived similarity did positively impact body image, no interaction effect of source type and perceived similarity was found. This means that differences in responses to the types of sources could not be explained by perceptions of similarity.

Moreover, both fitspiration type and source attractiveness did not affect body image and motivation.

Practical contributions

The results of the current study bring about some scientific, societal and practical implications. First and foremost, it can be concluded that the negative effects of health encouragement via SNS may not be as threatening as extant literature poses, considering the lack of significant effects. This lack of significant results, however, points out that repetition might be an interesting

(32)

moderator to include in future research, as it could be an important factor in whether health encouragement is detrimental to body image and motivation for behavioral change. This, however, is speculative, as we do not know for certain whether implications become more pronounced with repetition. Nonetheless, it is scientifically interesting as it indicates that,

concerning SNS, social influence is a process that could progress with exposure. From a practical point of view, it highlights that regulations which limit the amount of health promotion content could potentially be effective in preventing its negative effects if repetition were to be proven to be a significant moderator. Additionally, the results also contribute to the importance of

understanding which factors regulate whether social comparisons lead to either assimilations, or contrast effects. This is important, as the contrast effect leads to body dissatisfaction, which should be prevented, contrary to assimilation. Such a better understanding facilitates SNS users to generate psychologically healthy content, and a general comprehension regarding which type of content could be damaging to whom.

Limitations & Suggestions for Further Research

Limitations

Several limitations concerning the design of the current study should be mentioned. The most major one concerns the fact that participants were exposed to the fitspiration message one single time, before filling out the survey. In the real world, however, exposure to health promotion and the thin-ideal is constant and repeated because the content is widely spread and accessible. Therefore, results of a single exposure can not be generalized, as they do not represent real-life effects. Another limitation was that there was no pretest, documenting the participant’s

(33)

already-WHAT MAKES FITSPIRATION INSPIRING

established self-evaluations. In hindsight, this knowledge could have provided clarity on how social influence works for whom, making the results easier to interpret, leaving less of a theoretical gap. Furthermore, there was a disconnect between the survey and the literature. The literature emphasized how perceived similarity in terms of lifestyle, abilities and resources could influence effects on body image in particular. However, the survey employed an already-existing scale, highlighting similarity mainly in terms of appearance. It would, thus, be fair to say that the scale did not measure what needed to be measured. Subsequently, this could have caused for participants to focus on appearance aspects regarding similarity, possibly preventing processes that were described in the literature to be activated (e.g. increasing the likelihood of relevant, social upward comparisons to happen), steering the results away from what was expected.

Suggestions for Further Research

From the discussion, conclusion and limitations section, a number of suggestions for future research were derived. Social influence via SNS is a process that is likely to manifest with repetition and, thus, time. Therefore, it would be advised for future research to undertake a longitudinal study, as results would then be a more accurate representation of social influence mechanisms in real life. Rather than a single exposure, participants should thus be exposed to online health promotion several times. A further suggestion for future research would be to investigate what factors contribute for an individual to either engage in assimilation or contrast effects after a social comparison with a SNS source. This entails that participants should take a pretest that measures suggested moderators, such as already-established self-evaluations, and perceived extremity of the promoted standard in the message. A final suggestion would be to

(34)

scale should focus on perceived similarity in terms of lifestyle, abilities and resources, rather than appearance features. As of now, no definite statements regarding similarity effects can be made. However, a more relevant measurement of the variable, suitable to the current research topic, could provide more accurate results.

Overall, the effects of health encouragement, depending source type, source perceptions and fitspiration type that were found were meager. With exception to perceived similarity, none of the aforementioned variables had a compelling main, nor mediating effect on body image and motivation for behavioral change. As mentioned, perceived similarity had a significant effect on body image. This was, however, in the opposite direction of what was expected, being that the relationship was positive rather than negative. Why the results were found to be so minimal, could be dependent on a variety of factors. For example, (non)professionality could be irrelevant for social influence. Additionally, moderators like already established self-evaluation and the perceived extremity of the fitspiration post could have limited the strength of social influence. From this, it can be concluded that social influence via online health promotion is a complex process, subjected to a multitude of internal and external factors that interfere with or contribute to its possible implications on body image and motivation for behavioral change.

(35)

Bibliography

5 Fitness Quotes That Are The Absolute Worst. (July 19th, 2013). The Huffington Post. [Blog

post]. Retrieved from http://www.huffingtonpost.com/2013/07/19/worst-fitness-quotes_n_3618176.html

Adams, R. (July 7th, 2014). Why ‘Fitspo’ Should Come With A Warning Label. Huffington

Post. [Blog post]. Retrieved from

http://www.huffingtonpost.com/2014/07/17/fitspo-fitspiration_n_5574150.html

AdParlor Mockup Generator. Admocks.adparlor.com. Retrieved from http://admocks.adparlor.com/#instagram/app_installs/carousel Ajzen, I. (2006). Constructing a theory of planned behavior questionnaire.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.

Psychological Review, 84(2), 191-215. http://dx.doi.org/10.1037//0033-295x.84.2.191

Bandura, A., & Locke, E. (2003). Negative self-efficacy and goal effects revisited. Journal Of

Applied Psychology, 88(1), 87-99. http://dx.doi.org/10.1037/0021-9010.88.1.87

Boepple, L., & Thompson, J. (2015). A content analytic comparison of fitspiration and thinspiration websites. International Journal Of Eating Disorders, 49(1), 98-101. http://dx.doi.org/10.1002/eat.22403

Bolton, R., Parasuraman, A., Hoefnagels, A., Migchels, N., Kabadayi, S., & Gruber, T. et al. (2013). Understanding Generation Y and their use of social media: a review and research agenda. Journal Of Service Management, 24(3), 245-267.

http://dx.doi.org/10.1108/09564231311326987

Boyd, D., & Ellison, N. (2010). Social network sites: definition, history, and scholarship. IEEE

Engineering Management Review, 38(3), 16-31.

http://dx.doi.org/10.1109/emr.2010.5559139

Brinol, P., & Petty, R. E. (2009). Source factors in persuasion: A self-validation approach. European Review of Social Psychology, 20(1), 49-96.

Brown, Z., & Tiggemann, M. (2016). Attractive celebrity and peer images on Instagram: Effect on women's mood and body image. Body Image, 19, 37-43.

Byrne, D., Bond, M. H., & Diamond, M. J. (1969). Response to political candidates as a function of attitude similarity-dissimilarity. Human Relations, 22(3), 251-262.

Byrne, D., Griffitt, W., & Stefaniak, D. (1967). Attraction and similarity of personality characteristics. Journal of Personality and social Psychology, 5(1), 82.

Carrotte, E., Prichard, I., & Lim, M. (2017). “Fitspiration” on Social Media: A Content Analysis of Gendered Images. Journal Of Medical Internet Research, 19(3), e95.

http://dx.doi.org/10.2196/jmir.6368

Carrotte, E., Vella, A., & Lim, M. (2015). Predictors of “Liking” Three Types of Health and Fitness-Related Content on Social Media: A Cross-Sectional Study. Journal Of Medical

(36)

Internet Research, 17(8), e205. http://dx.doi.org/10.2196/jmir.4803

Cash, T., Fleming, E., Alindogan, J., Steadman, L., & Whitehead, A. (2002). Beyond Body Image as a Trait: The Development and Validation of the Body Image States Scale.

Eating Disorders, 10(2), 103-113. http://dx.doi.org/10.1080/10640260290081678

Chasler, J. K. (2016). Fitspiration: Empowering or Objectifying? The Effects of Fitspiration and Self-Objectification on Exercise Behavior.

Cialdini, R. B., & Goldstein, N. J. (2004). Social influence: Compliance and conformity. Annu.

Rev. Psychol., 55, 591-621.

Constantinides, E. (2013). Social Media Marketing: Challenges and Opportunities. Consumer

Information Systems and Relationship Management: Design, Implementation, and Use: Design, Implementation, and Use, 51.

Corcoran, K., Crusius, J., & Mussweiler, T. (2011). Social comparison: Motives, standards, and mechanisms. Theories in social psychology, 119-139.

De Lyon, A. T., Neville, R. D., & Armour, K. M. (2016). The Role of Fitness Professionals in Public Health: A Review of the Literature. Quest, 1-18.

Fardouly, J., Diedrichs, P., Vartanian, L., & Halliwell, E. (2015). Social comparisons on social media: The impact of Facebook on young women's body image concerns and mood.

Body Image, 13, 38-45. http://dx.doi.org/10.1016/j.bodyim.2014.12.002

Festinger, L. (1954). A Theory of Social Comparison Processes. Human Relations, 7(2), 117-140. http://dx.doi.org/10.1177/001872675400700202

Fishbein, M. (1979). A theory of reasoned action: some applications and implications.

Fox, J., & Bailenson, J. (2009). Virtual Self-Modeling: The Effects of Vicarious Reinforcement and Identification on Exercise Behaviors. Media Psychology, 12(1), 1-25.

http://dx.doi.org/10.1080/15213260802669474

Gabriel Taquini, R. (2016). The consumer-to-consumer persuasion: application of rhetorical

appeals in consumer’s reviews (Master's thesis)

Garcia-Marques, T., & Mackie, D. M. (2000). The positive feeling of familiarity: Mood as an information processing regulation mechanism. The message within: The role of subjective

experience in social cognition and behavior, 240-261. Philadelphia: Psychology Press

Ghaznavi, J., & Taylor, L. (2015). Bones, body parts, and sex appeal: An analysis of #thinspiration images on popular social media. Body Image, 14, 54-61. http://dx.doi.org/10.1016/j.bodyim.2015.03.006

Grabe, S., Ward, L., & Hyde, J. (2008). The role of the media in body image concerns among women: A meta-analysis of experimental and correlational studies. Psychological

Bulletin, 134(3), 460-476. http://dx.doi.org/10.1037/0033-2909.134.3.460

Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate Data Analysis. Upper Saddle River, NJ: Prentice Hall

Harper, K., Sperry, S., & Thompson, J. (2008). Viewership of pro-eating disorder websites: Association with body image and eating disturbances. International Journal Of Eating

(37)

Disorders, 41(1), 92-95. http://dx.doi.org/10.1002/eat.20408

Hawkins, R., Kreuter, M., Resnicow, K., Fishbein, M., & Dijkstra, A. (2008). Understanding tailoring in communicating about health. Health Education Research, 23(3), 454-466. http://dx.doi.org/10.1093/her/cyn004

Hillmann, R. & Trier, M. (2013), Influence and dissemination of sentiments in social network communication patterns. In

Proceedings of the 21st European Conference on Information Systems

Holland, G., & Tiggemann, M. (2016). “Strong beats skinny every time”: Disordered eating and compulsive exercise in women who post fitspiration on Instagram. International Journal

Of Eating Disorders, 50(1), 76-79. http://dx.doi.org/10.1002/eat.22559

Huang, J., Norman, G., Zabinski, M., Calfas, K., & Patrick, K. (2007). Body Image and Self-Esteem among Adolescents Undergoing an Intervention Targeting Dietary and Physical Activity Behaviors. Journal Of Adolescent Health, 40(3), 245-251.

http://dx.doi.org/10.1016/j.jadohealth.2006.09.026

Hu, Y., & Shyam Sundar, S. (2009). Effects of Online Health Sources on Credibility and Behavioral Intentions. Communication Research, 37(1), 105-132.

http://dx.doi.org/10.1177/0093650209351512

Hum, S. (May 15th, 2014). Marketing to Millennials: Why User-Generated Content Works Best. [Blog post]. Referral Candy. Retrieved from

https://www.referralcandy.com/blog/marketing-millennials-user-generated-content-works-best/

Jett, S., LaPorte, D., & Wanchisn, J. (2010). Impact of exposure to pro-eating disorder websites on eating behaviour in college women. European Eating Disorders Review, 18(5), 410-416. http://dx.doi.org/10.1002/erv.1009

Jones, D. C. (2001). Social Comparison and Body Image: Attractiveness Comparisons to Models and Peers Among Adolescent Girls and Boys. Sex Roles, 45(9-10), 645–664. http://dx.doi.org/10.1023/A:1014815725852

Kaplan, A., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59-68.

http://dx.doi.org/10.1016/j.bushor.2009.09.003

Kiberd, R. (Aril 20th, 2015). Fitspo: how strong became the new skinny. [Blog post] The

Guardian. Available at:

https://www.theguardian.com/lifeandstyle/2015/apr/20/fitspo-strong-skinny-social-media-food-abs-better-living-body-fascism

Korp, P. (2005). Health on the Internet: implications for health promotion. Health Education

Research, 21(1), 78-86. http://dx.doi.org/10.1093/her/cyh043

Kozar, J. (2012). Effects of Model Age on Adult Female Consumers’ Purchase Intentions and Attitudes for an Age-Specific Product, Clothing. International Journal Of Marketing

(38)

New Mexico.

Kelman, H. (1958). Compliance, identification, and internalization three processes of attitude change. Journal Of Conflict Resolution, 2(1), 51-60.

http://dx.doi.org/10.1177/002200275800200106

Klucharev, V., Smidts, A., & Fernández, G. (2008). Brain mechanisms of persuasion: how ‘expert power’ modulates memory and attitudes. Social Cognitive And Affective

Neuroscience, 3(4), 353-366. http://dx.doi.org/10.1093/scan/nsn022

Koven, N., & Abry, A. (2015). The clinical basis of orthorexia nervosa: emerging perspectives.

Neuropsychiatric Disease And Treatment, 385. http://dx.doi.org/10.2147/ndt.s61665

Latkin, C., & Knowlton, A. (2015). Social Network Assessments and Interventions for Health Behavior Change: A Critical Review. Behavioral Medicine, 41(3), 90-97.

http://dx.doi.org/10.1080/08964289.2015.1034645

Lenderman, M. (2006). Experience the message: How experiential marketing is

changing the brand world. (pp. 139-144) New York: Carrol & Graf Publishers

Lewallen, J., & Behm-Morawitz, E. (2016). Pinterest or Thinterest?: Social Comparison and Body Image on Social Media. Social Media + Society, 2(1), 205630511664055. http://dx.doi.org/10.1177/2056305116640559

Loss, J., Lindacher, V., & Curbach, J. (2014). Online social networking sites—a novel setting for health promotion?. Health & Place, 26, 161-170.

http://dx.doi.org/10.1016/j.healthplace.2013.12.012

Marsh, S. & Campbell, D. (October 1st, 2016). Clean eating trend can be dangerous for young people, experts warn. [Blog post] The Guardian. Retrieved from

https://www.theguardian.com/society/2016/oct/01/clean-eating-trend-dangerous-young-people-food-obsession-mental-health-experts

Mason, S. E. (2012). The Effect of Thin-Ideal Media On Body Image: An Experiment Using the

Solomon Four-Group Design (Doctoral dissertation)

Marwick, A. (2015). Instafame: Luxury Selfies in the Attention Economy. Public Culture, 27(1 75), 137-160. http://dx.doi.org/10.1215/08992363-2798379

McHugh, M. C., & Chrisler, J. C. (Eds.). (2015). The Wrong Prescription for Women: How

Medicine and Media Create a" Need" for Treatments, Drugs, and Surgery: How Medicine and Media Create a" Need" for Treatments, Drugs, and Surgery. (p. 166).

ABC-CLIO.

Neubaum, G., & Krämer, N. C. (2015). Let’s blog about health! Exploring the persuasiveness of a personal HIV blog compared to an institutional HIV website. Health communication,

30(9), 872-883. http://dx.doi.org/10.1080/10410236.2013.856742

O'Brien, C. (2011). The emergence of the social media empowered consumer. Irish marketing

review, 21(1/2), 32.

Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers' perceived expertise, trustworthiness, and attractiveness. Journal of advertising, 19(3),

What makes fitspiration inspiring? : exploring social influence and persuasion through Instagram