Content uploaded by Yilang Peng
Author content
All content in this area was uploaded by Yilang Peng on Jul 05, 2020
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=hhth20
Health Communication
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/hhth20
Fitspiration on Instagram: Identifying Topic
Clusters in User Comments to Posts with
Objectification Features
Volha Murashka , Jiaying Liu & Yilang Peng
To cite this article: Volha Murashka , Jiaying Liu & Yilang Peng (2020): Fitspiration on Instagram:
Identifying Topic Clusters in User Comments to Posts with Objectification Features, Health
Communication, DOI: 10.1080/10410236.2020.1773702
To link to this article: https://doi.org/10.1080/10410236.2020.1773702
Published online: 16 Jun 2020.
Submit your article to this journal
View related articles
View Crossmark data
Fitspiration on Instagram: Identifying Topic Clusters in User Comments to Posts with
Objectication Features
Volha Murashka
a
, Jiaying Liu
a
, and Yilang Peng
b
a
Department of Communication Studies, University of Georgia, Athens;
b
Department of Financial Planning, Housing and Consumer Economics,
University of Georgia, Athens
ABSTRACT
This study aimed to examine features of objectied images in popular tspiration accounts on social
media, identify the most prevalent user discussion topics about these images, and investigate the linkages
between specic objectication cues and discussion topics. We employed content analysis to identify
gender-specic objectication elements in tspiration images (N = 2,000) on Instagram and unsupervised
topic modeling to uncover topics in user comments (N = 35,263). Multilevel regressions were conducted
to examine how gender and objectication features associate with the topics. Results revealed that one-
third of the images were objectied. Objectication features such as sexual posing and the presence of
objectied chest or abdomen were associated with the occurrence of two out of the three most
prominent topic clusters (inspiration, body, and image-related discussions). We concluded that objectied
tspiration may distract users from health goals.
Introduction
Fitspiration is an increasingly popular social media trend that
promotes a healthy lifestyle and fitness by showing inspiring
examples (Easton et al., 2018). Despite good intentions, fitspira-
tion images may produce negative consequences on viewers’ body
image as such images often contain elements of objectification
(Tiggemann & Zaccardo, 2018), i.e., degrading someone to the
status of merely a body that is mainly valued for its use to and
pleasure of others (Fredrickson & Roberts, 1997). Objectified
fitspiration content may promote the perceptions that maintain-
ing an attractive body, instead of health well-being, is the main
goal of exercising (Deighton-Smith & Bell, 2018). Given that
exposure to exercise messages that emphasize the importance of
physical appearance may increase body shame (Aubrey, 2010), it
is important to critically analyze objectification cues presented in
fitspiration posts to minimize potential risks to viewers’ body
image.
Social media users leave an abundance of digital traces com-
menting on the objectified fitspiration content, providing
a valuable source for us to understand how body image percep-
tions are influenced by such content. Prior research has sepa-
rately examined objectification features in fitspiration images
(e.g., Deighton-Smith & Bell, 2018) and user discussions of
these images on social media (Santarossa et al., 2019). It remains
less clear how different objectification features, e.g., image cues
that focus on specific body parts, include sexual posing, or depict
an individual’s head or face in a less visible way (Tiggemann &
Zaccardo, 2018), may have promoted different discussion topics
among users. Moreover, although women are often subject to
objectification in media, research shows that objectified male
images have also become prevalent (e.g., Carrotte et al., 2017).
To develop more targeted interventions against negative body
perceptions, it is also important to understand whether these
objectification features would influence user discussions simi-
larly or differently for male and female fitspiration images.
This study aims to enrich gender-specific knowledge about
fitspiration on several fronts. First, this study describes objec-
tification cues in the images of the most popular female and
male fitspiration accounts on Instagram. Previous fitspiration
research has examined predominantly female images from the
most recent Instagram posts (e.g., Deighton-Smith & Bell,
2018). However, given the potentially different gender-
specific impacts of fitspiration content, and the widespread
influence of the popular fitspiration accounts, it is critical to
examine both female and male images from the most popular
accounts. In addition, building upon prior research (Santarossa
et al., 2019), this study attempts to examine the associations
between objectification features in the images and discussion
themes that emerged from the corresponding user comments.
A close scrutiny of the naturally occurring online user com-
ments to popular fitspiration posts enables us to better under-
stand whether objectified fitspiration content may lead its
followers to focus more on objectification cues and appearance
concerns, rather than health and fitness goals.
Literature review
Objectication in media and gender roles
The objectification theory posits that our body perceptions are
constructed not only from biology but also through discourses
and practices that prevail in our society. When a body or a body
part is valued mainly for its use to and pleasures of others
CONTACT Jiaying Liu jiaying.liu@uga.edu Department of Communication Studies, University of Georgia, Athens, GA, USA
HEALTH COMMUNICATION
https://doi.org/10.1080/10410236.2020.1773702
© 2020 Taylor & Francis Group, LLC
(Fredrickson & Roberts, 1997), which reduces an individual from
a human being to the status of an object, the body or the body part
is considered objectified (Bartky, 1990). Objectification can occur
in interpersonal and social encounters as well as in media pre-
sentations (Calogero et al., 2011). Objectification experiences may
lead to self-objectification, i.e., perceiving oneself as an object to be
viewed and evaluated by others (Fredrickson & Roberts, 1997).
Objectification cues in media can be understood as instances
of thematic descriptions or visual presentations depicting and
overemphasizing bodies and appearances (Aubrey et al., 2020).
Thematic descriptions of objectification highlight the values of
competence in externally perceivable traits (e.g., appearance)
over internal traits (e.g., accomplishments). For instance, pre-
vious content analysis found that headlines of women’s health
magazines included objectifying phrases focusing on achieving
a desired body or avoiding an unattractive body (Aubrey, 2010).
Objectification cues can also be manifested in visual depictions
through image cropping, highlighting close-ups of specific body
parts with no focus on a person, posing/posturing that guides
visual attention to a body part, and excessive skin exposure or
strategic undress (Aubrey & Hahn, 2016; Carrotte et al., 2017;
Tiggemann & Zaccardo, 2018).
It is important to distinguish between two forms of objecti-
fication: literal objectification and sexual objectification. Literal
objectification is conceptualized as perceiving an individual as an
object without considering a person’s sexuality whereas sexual
objectification refers to the explicit sexualization of an individual
(Heflick & Goldenberg, 2014). Sexual objectification is the most
prevalent form of objectification; however, objectification can
take various forms in treating a body as an object without
necessarily relating it to sexual pleasure (e.g., viewing a body as
a vessel for reproduction) (Smolak & Murnen, 2011). Thus,
objectified content in media emphasizes body and appearance
without a specific focus on the body as being sexual whereas
sexualized content primarily emphasizes the sexual appeal of
body parts (Calogero et al., 2011; Heflick & Goldenberg, 2014).
Previous research operationalized sexualized cues in media as
instances of sexually suggestive facial expressions, postures/pos-
ing in a sexy manner, provocative clothing, activities, make-up,
the amount of skin exposure, connotation of sexual activity, and
camera angles emphasizing specific body parts (e.g., crotch,
buttocks, cleavage) (Carrotte et al., 2017; Speno & Aubrey,
2018; Stankiewicz & Rosselli, 2008; Tiggemann & Zaccardo,
2018). Given that the intention of fitspiration content is not
focused on sexualization but can nevertheless contain sexualized
cues, a broader definition of objectification that takes various
forms, including sexual objectification, will be applied in the
current study.
Gender socialization is an important aspect of objectification
experiences (Fredrickson & Roberts, 1997). From childhood girls
learn that their bodies’ main function is to attract others (Stephens
et al., 1994). They perceive themselves as objects to be evaluated by
others and internalize the cultural standards of an idealized female
body (Thompson et al., 1999), which can lead to negative percep-
tions of their own bodies (e.g., Morry & Staska, 2001). In line with
these standards, women are likely to focus on losing fat with
a particular concentration on areas from the “waist down” (e.g.,
buttocks; Andersen et al., 2000).
In addition to its main focus on women’s experiences, the
objectification theory can also be applied to men’s experiences
with the consideration of societal and cultural standards of male
attractiveness (Moradi, 2010). The rapidly developing muscular-
ity culture socializes men into idealizing muscular bodies (Leit
et al., 2001) with a particular focus on gaining muscles in body
parts from the “waist up” (e.g., arms; Andersen et al., 2000). As
a result, men may feel pressured to comply with the ideal
muscular body depicted in the mass media (Barlett et al.,
2008). Thus, considering the relevance of objectification experi-
ences for both men and women, this study investigates objecti-
fication cues in the images of both genders through the lens of
the objectification theory.
Fitspiration on social media
While previous research on body perceptions has primarily
focused on the influence of media use in general (e.g.,
Tiggemann, 2003), recent research has indicated the crucial
role of social media platforms in shaping body perceptions. An
important area that warrants further exploration is fitspiration
or user-generated content on social media that encourages
users to work out and stay healthy (i.e., fit plus inspiration)
(Easton et al., 2018). By promoting strength and empower-
ment, fitspiration is considered a positive alternative to thin-
spiration (Tiggemann & Zaccardo, 2015), a trend that inspires
individuals to lose weight and promotes eating disorders
(Lewis & Arbuthnott, 2012). Fitspiration content is extremely
popular on Instagram (Carrotte et al., 2017). Our own search of
the hashtag ‘fitspiration’ and ‘fitspo’ on Instagram returned
18.74 million and 69.3 million images in March 2020.
Despite its good intentions, fitspiration has also led to some
concerns. Cumulative evidence has revealed negative unin-
tended consequences associated with viewing fitspiration
images including greater negative mood and higher body dis-
satisfaction (e.g., Prichard et al., 2018). To understand why
fitspiration content can be detrimental, it is important to
examine how fitspiration influencers present themselves to
their viewers. Self-presentation refers to the process by which
individuals package and edit images of themselves to produce
a desired impression to their audience (Goffman, 1959; Toma
& Hancock, 2010). Many online users aim to present them-
selves in an attractive way and minimize their imperfections
(e.g., Toma & Hancock, 2010). By selectively posting content,
fitspiration influencers promote their positive self-image
online by showcasing their strength, healthy eating, daily train-
ing, and their “perfect” body image as an outcome. Such self-
presentation strategies may seem benign on the surface.
However, in the long run they can cumulatively perpetuate
pervading body image ideals and elevate the importance of
physical appearance above that of health among viewers
(Pilgrim & Bohnet-Joschko, 2019).
Previous research shows gender differences in social media
users’ self-presentation. Compared to men who tend to high-
light their power and strength, women are more likely to
emphasize their physical attractiveness (Manago et al., 2008).
Emphasis on physical and sexual attractiveness has been
observed among female fitspiration influencers who have
2V. MURASHKA ET AL.
been found to use more words related to sexuality than men
(e.g., “butt,” “sexy”) (Auxier et al., 2019). Given the gender-
specific self-presentation goals, researchers have started to
examine appearance-related characteristics specific to women’s
and men’s visual self-presentation in fitspiration content.
Deighton-Smith and Bell (2018) found that women were fea-
tured more frequently than men in fitspiration content. Men
were more likely than women to demonstrate a muscular or
toned body, show half of their body, and display arms, face, and
chest; women, on the other hand, were more likely to show full
body, to emphasize buttocks and legs, and to wear sexualized
clothing. Tiggemann and Zaccardo (2018) also identified
a greater prevalence of muscular body shape among male
images and greater thinness among female images. Sexualized
fitspiration content was found more common for female than
male images (Carrotte et al., 2017).
Audience reactions to objectied tspiration images
While most existing studies have focused on quantifying objec-
tification features of fitspiration images, it is less known how
audience reacts to these images. Santarossa et al. (2019) exam-
ined user comments to recent fitspiration images (N = 10,000)
on Instagram and identified “good feelings” as one of the most
popular topics, based on word frequencies (54% of all words).
It was concluded that this finding supports the assumption that
fitspiration inspires people to follow a healthy lifestyle.
Similarly, Raggatt et al. (2018) found that viewers of fitspiration
content felt inspired to exercise and eat healthily. Based on the
empirical evidence and considering the inherently inspira-
tional nature of fitspiration (Easton et al., 2018), it is likely
that online users will express how such posts have inspired
them to work out and stay healthy.
Furthermore, comments on fitspiration images can be
understood from the perspective of a heterosexual script,
which defines a set of complementary but unequal roles for
women and men to follow in their romantic and sexual inter-
actions. It also describes how different genders are expected to
think, feel, and behave in these encounters to “appropriately”
reflect gendered power dynamics (Kim et al., 2007). For exam-
ple, it is expected that men actively pursue, sexually objectify,
and value women for their appearance, and attract women
using power and resources. At the same time, women are
expected to show interest in men in a passive way (e.g., using
their appearance to attract men) (Kim et al., 2007; Ward et al.,
2016). Considering the sexualized nature of many fitspiration
visuals, such content can trigger thoughts and communicative
behaviors that align with the heterosexual script – e.g., user
comments may focus on men’s physical prowess and resources
while on women’s sexuality and appearance.
Appearance conversations refer to discussions that reinforce
idealized appearance (e.g., thinness for women and muscularity
for men) and they are common for both women and men
(Becker et al., 2013; Jankowski et al., 2014). Appearance discus-
sions in offline settings have been found to negatively influence
both women’s and men’s body concerns (e.g., Jankowski et al.,
2014; Tiggemann & Boundy, 2008); however, there is a lack of
research on appearance comments in online settings such as on
social media (Feltman & Szymanski, 2018). Considering that
fitspiration content on social media often presents idealized
and physically attractive appearances (Tiggemann & Zaccardo,
2018), online users are likely to engage in appearance-related
discussions when commenting on the fitspiration images
(Santarossa et al., 2019). Given that individuals pay more atten-
tion to the physical appearance of women than men (Seidman &
Miller, 2013), such comments are expected to be more prevalent
for female fitspiration images.
Given that female and male bodies are often depicted differ-
ently in fitspiration content (e.g., Carrotte et al., 2017), user
discussions about female and male fitspiration images are likely
to differ in their focus on specific features or body parts. In
addition, stereotypical gender perceptions that dictate how
females and males are expected to dress, behave, and present
themselves may lead viewers to produce drastically different
interpretations even when the same features are presented in
the images (e.g., females with a masculine body may be consid-
ered aggressive, inappropriate, or less graceful thus inhibiting
viewers’ feelings of inspiration). Furthermore, to gain more
relevant information about exercising, dieting, and to improve
their appearance (Easton et al., 2018; Raggatt et al., 2018), female
users are more likely to follow female accounts and male users
tend to follow male fitspiration accounts. Considering that sam-
ples consisting of predominantly females perceive fitspiration as
an important platform to share health and fitness-related inter-
ests (e.g., Raggatt et al., 2018) whereas men are often less inclined
to discuss their imperfections on social media (Palmer, 2015),
differences are expected in prevalent discussion topics respond-
ing to female and male fitspiration posts. Understanding online
users’ reactions to female and male fitspiration images will
enrich our knowledge of problematic and health-promoting
aspects of gender-specific fitspiration content in general and
will be beneficial for future interventions focusing on individuals
of a specific gender.
Finally, despite the general agreement on the powerful
influence of fitspiration content, prior studies were either
focused exclusively on content analysis of image features
(Tiggemann & Zaccardo, 2018), comments (Santarossa et al.,
2019), or on self-reported perceptions (Prichard et al., 2018).
No study has examined actual associations between exogen-
ously content-coded image features and the corresponding
online users’ reactions. To fill this gap, this study investigates
in a real-world fitspiration context how different content-
analyzed objectification features are associated with viewers’
perceptions as reflected in their comments.
The present study
In light of the previous research evidence on fitspiration
reviewed above, the present study seeks to take a comparative
perspective of gender-specific fitspiration and further the fol-
lowing three lines of inquiry: 1) identify and compare the most
prevalent objectification cues in content posted by popular
male and female fitspiration Instagram accounts; 2) uncover
gender-specific comment topics emerged from Instagram user
discussions responding to fitspiration posts; 3) explore associa-
tions between gender-specific objectification features and like-
lihood of comment topic occurrence. First of all, based on prior
literature that identified and compared gender-specific
HEALTH COMMUNICATION 3
fitspiration image objectification features (e.g., Carrotte et al.,
2017; Deighton-Smith & Bell, 2018), we propose the following
hypothesis:
H1: Female fitspiration images are objectified more frequently
than male fitspiration images.
Given that recent content analytical research on fitspiration
images observed a higher prevalence of less clearly visible face/
head and sexual posing among female fitspiration posts (e.g.,
Tiggemann & Zaccardo, 2018), we put forth the following
hypothesis:
H2: (a) Sexual posing, (b) not clearly visible face, and (c)
absence of head will be more characteristic of female fitspira-
tion images than male fitspiration images.
Informed by previous research suggesting that female fit-
spiration images had a higher degree of objectification than
male images (Tiggemann & Zaccardo, 2018), we further aimed
to examine gender-specific differences in objectified body parts
which were considered the focus of the images. Considering
that previous gender-specific content analysis of fitspiration
either only explored very few body parts (e.g., stomach and
buttocks; Carrotte et al., 2017) or only examined body parts
that were visible, i.e., not necessarily the main focus of the
image (Deighton-Smith & Bell, 2018), this study aims to
further investigate gender-specific differences in objectified
body parts (including arms, legs, chest, abdomen, back, and
buttocks) that were the focus of the images in the most popular
fitspiration accounts. The following research question is
proposed:
RQ1: Which objectified body parts are more characteristic
of female and male fitspiration images respectively?
Second, based on the inspirational nature of fitspiration con-
tent (Tiggemann & Zaccardo, 2015), we speculate that online
users will show their fitness inspiration provoked by images in
fitspiration accounts through verbal expressions of admiration
and motivation to workout. Furthermore, considering prior
findings regarding body-related comments (Santarossa et al.,
2019), we expect that body-related discussions may appear in
user comments to images in the popular fitspiration accounts.
Thus, the following hypothesis is advanced:
H3: Comments to the objectified fitspiration images will be
predominantly focused on (a) fitness inspiration and (b) body
objectification related discussions.
Considering that very limited research has examined this
question and we may not be exhaustive in hypothesizing all the
possible topic cluster patterns, we also ask:
RQ2: What are the other prevalent topics that may emerge
from the comments?
Finally, to identify how gender-specific objectification elements
may elicit different discussion topics in user comments, we
propose to examine the following research questions:
RQ3: What are the differences in the most prevalent discus-
sion topics that emerge from the comments to the images with
objectification features depending on the gender of the fitspira-
tion account owners?
RQ4: How are the following objectification features, (a) sex-
ual posing, (b) objectified body parts, (c) not clearly visible face,
and (d) absence of head, associated with the likelihood of occur-
rence for each prevalent topic emerged from the comments?
Method
Data collection
This study employed a gender-comparison approach to explore
popular female and male fitspiration accounts. Considering that
a systematic sampling frame of Instagram profiles is not avail-
able, previous studies either obtained samples of fitspiration
images by searching for the most recent posts containing the
“fitspiration” or “fitspo” hashtags on Instagram (e.g., Deighton-
Smith & Bell, 2018), or used third-party platforms (e.g., tag-
board.com) that allowed extraction of posts based on time
stamps and overall popularity and then selected the first several
hundreds of images returned for coding (Carrotte et al., 2017).
These extraction methods have resulted in samples with predo-
minantly female images given the unbalanced distribution of
female and male fitspiration images on Instagram. Due to the
importance of gender comparison in this study, to obtain sam-
pling frames for both popular female and male images, we
adopted a different approach starting with identifying popular
Instagram fitspiration accounts for each gender.
Specifically, we resorted to Google News to identify popular
accounts. Using the Google advanced search operators to con-
nect four groups of search terms (intitle: fitspiration OR intitle:
fitspo) AND (account OR profile) AND (best OR popular)
AND (Instagram), we were able to locate 80 articles that con-
tained either “fitspiration” or “fitspo” in the title, and the other
three groups of keywords either in title or text published within
five years of October 2017 (i.e., the time our search was con-
ducted) from a variety of online news media outlets. We then
manually examined each article and compiled a list of fitspira-
tion accounts mentioned in these articles. We checked each
account’s profile name, description, profile picture, number of
followers, and recent posts. Eligible (non-commercial/institu-
tional, non-celebrity, identifiable, and non-mixed gender) and
relevant (those that contained more than 60% of posts relevant
to the fitness or fitspiration theme) publicly available accounts
that were mentioned more than three times across media
recommendations were selected (eligible N = 208, out of
a total of 413 accounts retrieved). The focus on identifying
the popular accounts was also based on our intention to obtain
the highest possible number of comments for the subsequent
topic modeling analysis.
Next, we coded fitspiration accounts by gender based on
profile information (85% were female accounts) and ranked the
accounts in each gender pool by the number of followers. To
facilitate gender-based comparison, we chose 20 top-ranked
accounts for each gender. For each of the 40 accounts chosen,
50 images posted in 2017 were randomly selected, resulting in
4V. MURASHKA ET AL.
a total sample of 2000 images (N = 1,000 images for each
gender). Selenium WebDriver (Muthukadan, 2018) was used
to systematically retrieve all the comments (N = 35,263) asso-
ciated with the selected images. Selenium WebDriver is an
automated browsing software that can be used to extract data
from online social networks (Diaz-Granados et al., 2015).
Coding procedures
The images were first coded by an overall content category
“people.” Images containing people were further coded for gen-
der (male, female, or both) and objectification (objectified, non-
objectified). Following prior practice (Carrotte et al., 2017;
Tiggemann & Zaccardo, 2018), we coded an image as objectified
when there was a focus on a body part, or body parts, a specific
body turn or pose, cropping of the image, or setting a camera
from a particular angle. Expanding Tiggemann and Zaccardo
(2018) codebook, we coded the objectified images with the
following features: focus on a specific body part, posing in
a sexy manner, clearly visible face, and depicted head. If an
image was focused on body parts, we coded specific information
of the body part(s) including arms, legs, chest, abdomen, back,
and buttocks. For images that had multiple people, our objecti-
fication coding applied to the focal person, or the first person
from the left side if it was unclear who was the main focus of the
post as it would be the first person seen if an image was viewed
from left to right (consistent with Carrotte et al., 2017).
Before coding the overall sample of 2000 images, we ran-
domly chose 200 images from the accounts of interest that were
not included in our final sample. Two coders, who were graduate
students of female and male genders, worked collaboratively on
coding and discussing 50 images randomly chosen from the trial
sample of 200 images. After resolving all disagreements and
refining coding strategies through discussion, they coded
another set of 150 images independently and high intercoder
reliability was achieved (average Krippendorff’s α = 0.95). After
that, 2000 images were split by the two coders, i.e., each coder
coded 500 female and 500 male images. See Table 1 for coding
criteria, inter-coder reliabilities, and example figures.
Data analysis
To identify the prevalence of objectification across gender-
specific fitspiration images (H1) and to explore differences in
gendered objectified images (H2 and RQ1), we performed
a content analysis on the fitspiration images and conducted
a one-way multivariate analysis of variance (MANOVA). To
uncover the most prevalent discussion topics about the objec-
tified fitspiration images (H3 and RQ2) and explore gender
differences in the most prevalent discussion topics (RQ3), we
conducted unsupervised topic modeling analyses applying
Latent Dirichlet Allocation (LDA) on the user-generated com-
ments associated with the objectified images in our sample (for
details on LDA, see Blei, 2012; Blei et al., 2003).
LDA is a statistical model of language that assumes that the
textual dataset consists of a number of topic clusters, and each
topic cluster is manifested by a collection of words. Rather than
imposing categories of interest beforehand, it utilizes modeling
assumptions and patterns of word co-occurrences to group texts
into topic clusters, each of which is interpretable in the form of
a probability distribution over words (Grimmer & Stewart,
2013). In each topic, terms are considered prominent when
they occur significantly more frequently in this topic than in
any other topic. The topics are estimated simultaneously with
the analysis of frequently occurring words and phrases that
belong to a particular topic. Because topic modeling understands
various cases of term usage depending on the context, the same
word may appear in different topics (DiMaggio et al., 2013). The
Python packages NLTK (Bird et al., 2009), lda (Riddell et al.,
2016), and Gensim (Řehůřek & Sojka, 2010) were used to con-
duct the topic modeling analyses.
Finally, to understand whether objectification features
(RQ4) were associated with any prevalent discussion topics
while accounting for the nested structure of our data, multi-
level regression analyses were conducted in R with the percen-
tage of each topic (as identified by topic modeling) appearing
in the comments as the dependent variable. Objectification
features were specified as fixed effects and fitspiration account
was treated as a random effect. The number of comments and
the number of likes associated with each image were controlled
in the analyses.
Results
Content analysis identified 377 female and 301 male images
with objectification features resulting in 678 objectified images
(33.9%) out of 2000 images. For female images, 16 out of 377
cases depicted more than one person; among these, no clear
focus on any specific individual was identified in 7 images (less
than 2% of the female sample). For male images, 30 out of 301
images depicted more than one person; among these, no clear
focus on any specific individual was identified in 14 images
(less than 5% of the male sample). We obtained 26,487 com-
ments for the female objectified images and 8,776 for the male
objectified images. The average number of comments and likes
was 141 (SD ¼366.65) and 8,543 (SD ¼15,758.07) respec-
tively, for female images and 54 (SD ¼93.38) and 4,950 (SD
¼7,098.16) respectively, for male images. Significantly more
female images (37.7%) than male images (30.1%) were identi-
fied with objectification features (χ
2
(1) ¼12.88, p <.001;
H1). Next, we compared the proportions of each feature occur-
ring in the objectified images, as identified by the content
analysis, for female and male fitspiration images, respectively
(see Table 2). Across female images, the most featured body
parts were abdomen (48.5%), arms (39.8%), and legs (39%). In
regard to male images, the most featured body parts were arms
(80.7%), abdomen (39.5%), and chest (37.5%).
To further examine gender difference in objectification fea-
tures (H2 and RQ1), we conducted a one-way MANOVA with
sexual posing, objectified body parts (arms, legs, chest, abdo-
men, back, and buttocks), not clearly visible face, and absence of
head as dependent variables. The results suggested a significant
overall gender difference (Wilks Λ ¼.730; F(9, 665) ¼27.264,
p ¼.001, partial η
2
¼.270).
Given an overall significant effect of gender, we looked at
the results of each dependent variable separately. We applied
the Bonferroni adjustment to control for Type 1 error, with p
¼.0056 as the cutoff value. We observed significant gender
HEALTH COMMUNICATION 5
Table 1. Coding scheme for the fitspiration images on instagram.
Variable Description Coding Details Reliability
a
Example Images
b
People People or other Images that include at least one person = 1; other = 0 1
Gender Genders of individuals in
the image
Male =1, female = 2, or both = 3 (if individuals of both genders are considered the focus of the image) .99
Objectification
c
Presence of objectification
features
1. Images with a focus on a specific body part =1; images without a focus on a specific body part = 0 .92
2. Posing in a sexy manner = 1 (e.g., lifting one’s arm overhead while leaning or sitting; alluring or sultry gaze; winking; lying down; revealing
clothing); no posing in a sexy manner = 0
1
3. Clearly visible face (> 50% of the face) = 1; not clearly visible face (< 50% of the face) = 0 .94
4. Presence of head = 1; absence of head = 0 .96
Body Parts Focus on specific body
parts
The body part is a focus of the image = 1; the body part is not a focus of the image = 0 (coding applies to all specific body parts as listed below):
1. Arms (bare bicep and shoulder on show)
2. Legs (thigh or more than 50% of bare lower leg)
3. Chest (cleavage for females and pectoral muscles for males)
4. Abdomen (bare)
5. Back (bare)
6. Buttocks (revealing or emphasis through specific turn/posing)
.96
.96
.91
.96
.89
.89
Buttocks
Legs
Chest
Abdomen
Arms
a
.Inter-coder reliability was calculated with Krippendorff’s alpha.
b
.Faces in the images were intentionally blurred.
c
.Considering that one image may contain multiple objectification features, the four objectification features were coded as four dummy variables for each image.
6V. MURASHKA ET AL.
differences on sexual posing (F(1, 673) ¼23.69, p <.001),
absence of head (F(1, 673) ¼10.34, p ¼.001), arms (F(1,
673) ¼118.05, p <.001), legs (F(1, 673) ¼34.49, p
<.001), abdomen (F(1, 673) ¼8.78, p ¼.003), and but-
tocks (F(1, 673) ¼63.93, p <.001). Comparison of the
frequencies of objectification features across female and male
images suggested that female images were significantly more
likely to contain sexual posing, less likely to show an indivi-
dual’s head (H2a and H2c), and more likely to include objecti-
fied legs, abdomen, buttocks. Male images were significantly
more likely to show objectified arms (RQ1). There was no
significant difference for the visible face (H2b) across female
and male images (see Table 2).
Our next hypothesis (H3) examined the discussion
topics that emerged from the comments responding to the
objectified images identified by content analysis. Given the
nature of our text corpus (i.e., Instagram comments which
are all short texts), it is unlikely for the LDA to generate
a large number of topics that are often observed in larger
text corpora that contain much longer texts such as news
articles or broadcast transcripts. We thus ran LDA models
with the number of topics ranging from 2 to 10. To deter-
mine the optimal topic solution, we used the topic coher-
ence measure C
V
, which is an established measure to assess
topic quality, as it provides the closest scores to human
evaluation in comparison to other commonly used mea-
sures of topic coherence (Röder et al., 2015). Based on the
results, the topic number of three produced the highest C
V
of 0.63, and it was chosen as the optimal topic solution for
discussions associated with female images. For discussions
associated with male images, the highest scores were iden-
tified for the three-topic (C
V
= 0.71) and the five-topic (C
V
= 0.72) solutions. Manual examination of the top-ranked
keywords for the suggested solutions revealed an overlap of
topics in the five-topic solution. Thus, the three-topic solu-
tion was chosen for the discussions associated with male
images. Although slightly different in their specific focus
and prevalence distributions, as we will elaborate below, the
three topics for both comments’ corpora could be summar-
ized as (1) fitness inspiration, (2) body-related discussions,
and (3) image-related discussions. Thus, our hypothesis
(H3) was supported and a new topic (image-related discus-
sions) was identified (R2). The keywords for each of the
topics with relevant examples and prevalence information
are presented in Table 3.
As can be seen from Table 3, the fitness inspiration topic
accounted for 74.6% and 25.4% of comments to female and
male objectified images. This group of comments included
positive feedback about the inspirational nature of the posts,
encouraging fitspiration accounts to continue sharing motiva-
tional content, expressing appreciation for the shared content,
and gratitude for motivation to pursue fitness goals. Such
examples were found in comments associated with images of
both genders and our close reading of the comments highly
associated with this topic (based on topic keyword occur-
rences) did not show differences in the way online users
expressed inspiration to female and male fitspiration
images (RQ3).
The topic of body-related discussions was identified in both
comments’ corpora (11.7% for female images and 27.6% for
male images). Close scrutiny of the keywords and texts most
associated with this topic revealed that body-related discus-
sions for female images were almost exclusively focused on the
sexually objectified body, as reflected in the most frequently
appearing keywords such as “perfect,” “sexiest,” and “hot”
(Table 3). For male images, comments associated with this
topic were also focused on the sexually objectified body as
can be seen in the keywords such as “hot,” “abs,” and “super”
(e.g., “supermodel”). In addition, the keywords for the same
topic included masculine words (e.g., “man,” “brother”). This
pattern was observed to a lesser extent in body-related discus-
sions to female images; specifically, only one female-related
word (“woman”) was identified in body-related discussions.
In other words, from the perspective of gender differences in
this topic (RQ3), body-related discussions heavily centered on
body objectification for both female and male images; however,
gender-specific language was observed more in comments
responding to male images.
In addition to the two hypothesized topics, we also identi-
fied a third topic, image-related discussions, which accounted
for 13.6% and 46.9% for the comments in female and male
images, respectively. This topic generally focused on the ele-
ments of posts (e.g., location, image caption, or image
esthetics). We also observed some gender-specific characteris-
tics associated with this topic (RQ3). Specifically, male image
comment corpus was focused more on the image, video, cap-
tion, or message presented in the image while female image
comment corpus emphasized various items of fitness clothing
or accessories worn by an individual in the image. Some
examples of these conversations included compliments on
Table 2. Objectification features across female and male fitspiration images.
Objectification features
Women
(N = 377)
Men
(N = 301) F(1, 673) statistic p Value partial η
2
Body parts Arms 150 (39.8%) 243 (80.7%) 118.05 p <.001 .149
Legs 147 (39.0%) 62 (20.6%) 34.49 p <.001 .049
Chest 121 (32.1%) 113 (37.5%) 0.64 p =.423 .001
Abdomen 183 (48.5%) 119 (39.5%) 8.78 p =.003 .013
Back 45 (11.9%) 23 (7.6%) 2.33 p =.127 .003
Buttocks 106 (28.1%) 19 (6.3%) 63.93 p <.001 .087
Sexual posing 70 (18.6%) 14 (4.7%) 23.69 p <.001 .034
Face not clearly visible 88 (23.3%) 55 (18.3%) 5.58 p = .180 .008
Head is not depicted 20 (5.3%) 14 (4.7%) 10.34 p =.001 .015
Notes. Numbers in cells represent the count and percentage of occurrences of the corresponding objectification feature in the female and male objectified fitspiration
images, respectively.
HEALTH COMMUNICATION 7
specific items of clothing, questions about brands of fitness
clothes, and expressing the need to have a similar item of
fitness clothing.
To answer RQ4, we conducted multilevel regression analyses,
to explore whether objectification cues identified by our content
analysis were associated with the likelihood of occurrence of each
topic identified by the topic modeling analysis.
1
Multilevel analyses
did not identify any significant effect of objectification features on
the likelihood of topic one (fitness inspiration) occurrence for
female and male images. For topic two (body-related discussions),
significant positive effects of sexual posing (γ ¼.16, p ¼.003)
and chest (γ ¼.14, p ¼.005) on the occurrence of this topic were
observed for female images. For male images, objectification of
chest (γ ¼.16, p ¼.003) and abdomen (γ ¼.13, p ¼.025) was
positively associated with body-related discussions. For topic three
(image-related discussions), none of the objectification features
was significantly associated with the occurrence of this topic for
female images. For male images, the analyses identified
a significant negative effect of chest objectification on the occur-
rence of this topic (γ ¼− .13, p ¼.002) (see Table 4). In
summary, for female images, sexual posing and objectified chest
were positively associated with body-related discussions. For male
Table 3. Topic clusters, interpretations, and highest-ranking terms generated from the 3-topic solution using LDA topic modeling of comments associated with
fitspiration images.
Image corpora (by gender)
Topic, interpretation,
and prevalence (%) Highest-ranking terms in the topic
a
Examples
Female Topic 1:
Inspiration
(74.6%)
Inspir*, fit*, workout*, follow*, want*,
help, share*, start, hard, goal*
-You’ve been my fitness inspiration since I was 16
- I want to thank you for helping encouraging and inspiring
women to workout
- So inspiring the way you help ppl
- You are so inspiring and I have been following you and
have loved your workouts over the last couple years
Topic 2:
Body-related
(11.7%)
Babe, goal*, ab*, perfect, sexi*, leg*, hot*,
fit, stun*, woman
- What a Nice abs babe
- Butt goals
- The perfect looking woman
- You have the sexiest legs ever
Topic 3:
Image-related
(13.6%)
Leg*, cool, top, wear*, tight*, want, style,
activ*, pant*, photo
- I want those leggings
- Need these tights
- I want these shoes
- Great photo
Male Topic 1:
Inspiration
(25.4%)
Inspir*, motiv*, fit*, follow*, believ*,
help*, start*, share, appreci*, video*
- You are so motivating and inspiring
- You are a fitness inspiration
- It helps on those days when you feel down and no
motivation
- May you continue to share all your knowledge and
inspiration with us all
Topic 2:
Body-related
(27.6%)
Man, bro*, hot, beast, ab*, strong, goal*,
super, fit*, face
- Fittest man alive
- Amazing Abs My Brother
- Your face looks hot here
- Hot and Strong
Topic 3:
Image-related
(46.9%)
Word*, incred*, said, man, gorgeous,
stun*, omg* caption*, dude, thought
- Wise words man
- So incredibly true and well put
- Simple and beautifully said
- That is truly a stunning picture
Notes. a. Words with an asterisk (*) were truncated with Snowball stemmer.
Table 4. Multilevel regression analyses of gender-specific objectification features predicting topic occurrences.
Objectification features
Topic 1
(Inspiration)
Topic 2
(Body-related)
Topic 3
(Image-related)
Female Male Female Male Female Male
Sexual posing −.08
†
−.04 .16** −.08 −.02 .10
†
Visible face −.09
†
.07 .07 −.01 .05 −.05
Head cropped .03 .11 .02 −.04 −.05 −.04
Objectified body parts Arms .05 −.01 −.05 −.03 −.02 .04
Legs −.03 .01 .01 −.09 .03 .08
Chest −.08
†
−.04 .14** .16** −.01 −.13**
Abdomen −.04 −.08 .04 .13* .03 −.07
Back .09
†
−.06 −.09
†
−.03 −.05 .07
Buttocks −.01 −.08 −.02 −.03 .03 .08
Notes. ** p <.01; * p <.05;
†
p <.1. Standardized multilevel regression coefficients γ are reported. We examined how objectified image features and body parts were
associated with the occurrence of each topic in the two comments corpora (i.e., responding to female and male images) separately. “Female” and “male” in the table
present each of the comments corpus, respectively. All objectification features were coded as dummy variables as multiple features could be present in the same
image, e.g., having various objectification elements in one image. The multilevel analyses specified message features associated with each image as fixed effects, and
the fitspiration accounts as random effects. N = 20 for account-level analysis in both gender pools; N = 377 and N = 301 for image-level analysis in female and male
pools, respectively. We controlled for the number of comments and likes associated with each image in the analyses. For female and male images, the number of likes
was significantly associated only with Topic 1. No significant effect was observed for the number of comments on the occurrence of any of the topics. Detailed results
associated with the two control variables are available upon request.
8V. MURASHKA ET AL.
images, objectified chest and abdomen were positively associated
with body-related discussions; objectified chest was negatively
associated with image-related discussions.
Discussion
This study aimed to explore the prevalence and characteristics
of objectification across female and male fitspiration images on
Instagram, and to identify the most prevalent topics that
emerge from comments to fitspiration posts. In particular, we
were interested in understanding how gender and objectifica-
tion features were linked to the likelihood of comment topic
occurrence within this context. Our study is among the first to
analyze a large corpus of Instagram discussions in order to
provide empirical evidence for the types of discussions that
objectified fitspiration images can elicit. Our results revealed
that close to one-third of all fitspiration images were objecti-
fied, corroborating the proposition that fitspiration content
represents an objectified perception of fitness (Deighton-
Smith & Bell, 2018). Although female images were objectified
more frequently than male images, almost one-third of male
images contained objectification cues. Because objectified con-
tent can foster self-objectification among male and female
individuals (e.g., Aubrey, 2006), inquiries about objectification
for both genders may inform interventions targeted at women
and men to reduce negative body perceptions.
For gender-specific objectification features, we observed
that female images included more depictions of the body in
a sexual pose, which supported previous findings on fitspira-
tion (Tiggemann & Zaccardo, 2018) and corroborated predic-
tions based on the objectification theory (Fredrickson &
Roberts, 1997). In line with the research on “face-ism” suggest-
ing that male images are more focused on face or head com-
pared to female images (Archer et al., 1983), absence of head
and not clearly visible head were more common among female
images; however, the difference reached significance only for
the images with the absence of the head. Our results aligned
with previous findings emphasizing the prevalence of objecti-
fied buttocks, legs, and abdomen in female images and arms in
male images (Carrotte et al., 2017; Deighton-Smith & Bell,
2018). Because women desire to modify areas from the “waist
down” (e.g., buttocks) and men desire to gain muscles in the
areas from the “waist up” (e.g., arms) (Andersen et al., 2000),
popular female and male fitspiration accounts may have dif-
ferent foci when posting images of the idealized body to
respectively attract viewers of each gender.
Echoing the most prevalent #fitspo themes (i.e., feeling good
and appearance-related discussions) identified by Santarossa
et al. (2019), our study also observed that feeling inspired and
body-related discussions were among the most common topics
in user comments. Furthermore, consistent with findings from
Raggatt et al. (2018), which concluded that the majority of
individuals (90.3%) who accessed fitspiration content reported
feelings of inspiration to follow a healthy lifestyle, our study
further confirmed, through analysis of user comments, that
individuals not only felt inspired but also shared their feelings
of inspiration in their comments.
Findings from the previous literature were mixed regarding the
potential of fitspiration to inspire users to follow a healthy lifestyle.
Tiggemann and Zaccardo (2015) found that viewing fitspiration
(versus traveling) images was associated with inspiration to eat
healthy food and improve fitness. However, after entering pre-
exposure inspiration scores, the effect of image type on inspiration
was not significant. Robinson et al. (2017) found that viewing
athletic (versus thin) image ideals was associated with greater
inspiration; however, there was no influence on actual exercising.
Considering that a substantial part of the comments to fitspiration
posts is related to inspiration, it is important to further explore the
factors that facilitate actual healthy behaviors after feeling inspired
by fitspiration images. Interestingly, inspirational discussions in
our sample were less prevalent among the comments to male
images. One possible explanation for this finding may be related
to the dominance of masculine behavior in the online muscle-
building community (Smith & Stewart, 2012); consequently, men
may be less comfortable showing their vulnerabilities by expres-
sing inspiration of others’ physique and exercise routines. More
research is warranted to explore gender differences in inspira-
tional discussions.
Furthermore, considering that in real life viewers of fitspira-
tion posts are likely to be exposed to multiple comments, it is
critical for future research to examine the effects of viewing
comments with different topics to the same fitspiration post
(e.g., inspirational followed by appearance-related discus-
sions). For example, although observing how other online
users express their inspiration in the comments may trigger
positive emotions, subsequent exposure to phrases emphasiz-
ing appearance can contribute to focusing on appearance-
related reasons for exercising, which can be related to negative
health outcomes (Aubrey, 2010; DiBartolo et al., 2007). The
ultimate effect may depend on a combination of factors such as
the distribution of frequencies associated with comments of
different topics, the order of the topics appearing in the com-
ment board during exposure, argument strength associated
with each comment topic, the viewers’ individual characteris-
tics, etc., and warrants future investigations.
Our analysis also revealed some negative effects of objecti-
fied fitspiration images. While previous studies inferred poten-
tially negative impacts of objectification through content
coding of images, our study directly examined viewers’ reac-
tions and confirmed that audiences indeed pick up the objecti-
fication cues in images and focus their discussion on the body
instead of health or fitness-related goals. Research showed that
viewing objectification words can increase body shame and
appearance anxiety (Roberts & Gettman, 2004), and exposure
to appearance-related comments on social media triggers body
dissatisfaction (Tiggemann & Barbato, 2018). In line with these
prior findings, this study shows that user comments to objecti-
fied fitspiration posts include themes that can further promote
objectifying perceptions of fitness.
We also observed that discussions following female images
focused heavily on appearance particularly with reference to
clothing. Considering the positive association between the
perceived importance of wearing fashionable clothes at fitness
centers with self-objectification (Prichard & Tiggemann, 2005),
discussions of clothing to female fitspiration posts may be
HEALTH COMMUNICATION 9
concerning as they may trigger an objectification focus by
emphasizing wearing attractive fitness clothes. In addition,
discussion of clothing in fitspiration accounts may be interest-
ing from the perspective of ‘influencer marketing’ that refers to
the collaboration between different brands and social media
influencers (De Veirman et al., 2017). Because popular fitspira-
tion accounts have a substantial number of followers, fitness
influencers may collaborate with commercial brands to pro-
mote their products. By advertising fitness clothing, fitness
influencers may further emphasize appearance focus among
their followers. Future research is invited to explore the asso-
ciation between appearance-related advertisement in fitspira-
tion accounts and viewers’ focus on appearance and self-
objectification.
An important finding of this study is related to the features
that generate objectification-related discussions. Our results sug-
gested that body-related discussions were more likely to be
generated when chest and abdomen were featured in male
images and sexual posing and chest in female images. For male
images, the observed greater occurrences of body-related discus-
sions in response to objectified upper body parts aligned well
with the pervading cultural standards of the idealized muscular
male body, specifically upper body parts (Andersen et al., 2000).
For female images, it is possible that objectified chest (i.e.,
a female body part that is often strongly associated with sexu-
ality) and sexual posing highlighted sexual attractiveness of the
depicted female body, resulting in more appearance-related and
sexualized discussions.
This study is not without limitations. We analyzed objectifi-
cation elements across 2,000 fitspiration posts from the popular
fitspiration accounts, which may not be representative of all
fitspiration content on Instagram and other social media. In
addition, following prior practice (Carrotte et al., 2017), for
images that depicted multiple individuals without an explicit
focus, we coded the first person from the left. This coding
decision may contain some degree of subjectivity. Future
research is recommended to examine how different coding deci-
sions for images that contain multiple people may affect conclu-
sions differently. Additionally, our sample was collected in 2017
and it is possible that fitspiration content and users’ reactions
may be affected by recent movements such as #MeToo (O’Neil
et al., 2018) and body positivity (Cohen et al., 2019) on social
media since then. Finally, this study did not include systematic
human coding in the comment analyses to further corroborate
the topic clusters identified by the automated topic modeling
process. Future research may benefit from including both
human coding and automated topic modeling in their analyses.
Conclusion
This study employed a data-driven approach to identify general
patterns in a large-scale social media data corpus on fitspiration.
To our knowledge, this is the first study that identified the main
topics of fitspiration discussions using the unsupervised topic
modeling. The findings of this study shed light on understanding
why fitspiration content may be detrimental to viewers’ body
perceptions. At the same time, considering the potential positive
impact of fitspiration content that may inspire users to follow
a healthy lifestyle (Raggatt et al., 2018), we encourage scholars to
investigate how to effectively leverage user-generated fitspiration
content to promote positive health outcomes while minimizing
the undesirable side-effects of objectification.
Note
1. Regression analyses using the overall dataset (i.e., combining male
and female datasets) revealed significant moderation effects of
gender on objectified back in predicting fitness inspiration discus-
sions (γ =.21, p =.04) as well as on objectified chest (γ =.34, p <.001
and γ = −.35, p <.001) in predicting body-related and image-related
discussions respectively.
ORCID
Volha Murashka http://orcid.org/0000-0002-7212-9255
Jiaying Liu http://orcid.org/0000-0001-8461-0954
Yilang Peng http://orcid.org/0000-0001-7711-9518
Disclosure of potential conict of interest
No potential conflict of interest was reported by the author(s).
References
Andersen, A., Cohn, L., & Holbrook, T. (2000). Making weight: Men’s
conicts with food, weight, shape, and appearance. Gurze Books.
Archer, D., Iritani, B., Kimes, D. D., & Barrios, M. (1983). Face-ism: Five
studies of sex differences in facial prominence. Journal of Personality
and Social Psychology, 45(4), 725–735. https://doi.org/10.1037/0022-
3514.45.4.725
Aubrey, J. S. (2006). Effects of sexually objectifying media on
self-objectification and body surveillance in undergraduates: Results
of a 2-year panel study. Journal of Communication, 56(2), 366–386.
https://doi.org/10.1111/j.1460-2466.2006.00024.x
Aubrey, J. S. (2010). Looking good versus feeling good: An investigation of
media frames of health advice and their effects on women’s body-
related self-perceptions. Sex Roles, 63(1–2), 50–63. https://doi.org/10.
1007/s11199-010-9768-4
Aubrey, J. S., & Hahn, R. (2016). Health versus appearance versus body
competence: A content analysis investigating frames of health advice in
women’s health magazines. Journal of Health Communication, 21(5),
496–503. https://doi.org/10.1080/10810730.2015.1103328
Aubrey, J. S., Speno, A. G., & Gamble, H. (2020). Appearance framing
versus health framing of health advice: Assessing the effects of
a YouTube channel for adolescent girls. Health Communication, 35
(3), 384–394. https://doi.org/10.1080/10410236.2018.1564955
Auxier, B., Buntain, C., & Golbeck, J. (2019). Analyzing sentiment and
themes in fitness influencers’ twitter dialogue. In N. G. Taylor,
C. Christian-Lamb, M. H. Martin, & B. Nardi (Eds.), Information in
contemporary society (pp. 429–435). Springer.
Barlett, C. P., Vowels, C. L., & Saucier, D. A. (2008). Meta-analyses of the
effects of media images on men’s body-image concerns. Journal of
Social and Clinical Psychology, 27(3), 279–310. https://doi.org/10.
1521/jscp.2008.27.3.279
Bartky, S. L. (1990). Femininity and domination: Studies in the phenom-
enology of oppression. New York, NY: Routledge.
Becker, C. B., Diedrichs, P. C., Jankowski, G., & Werchan, C. (2013). I’m not
just fat, I’m old: Has the study of body image overlooked “old talk”?
Journal of Eating Disorders, 1(1), 1–12. https://doi.org/10.1186/2050-
2974-1-6
10 V. MURASHKA ET AL.
Bird, S., Klein, E., & Loper, E. (2009). Natural language processing with
Python: Analyzing text with the natural language toolkit. O’Reilly
Media, Inc.
Blei, D. M. (2012). Probabilistic topic models. Communications of the
ACM, 55(4), 77–84. https://doi.org/10.1109/msp.2010.938079
Blei, D. M., Ng, A. Y., & Jordan, M. I.Latent dirichlet allocation (2003).
Journal of Machine Learning Research, 3, 993–1022. https://doi.org/10.
5555/944919.944937
Calogero, R. M., Tantleff-Dunn, S., & Thompson, J. K. (2011).
Objectification theory: An introduction. In R. M. Calogero, S. Tantleff-
Dunn, & J. K. Thompson (Eds.), Self- objectification in women: Causes,
consequences, and counteractions (pp. 3–21). American Psychological
Association. https://doi.org/10.1037/12304-001
Carrotte, E. R., Prichard, I., & Lim, M. S. C. (2017). “Fitspiration” on social
media: A content analysis of gendered images. Journal of Medical
Internet Research, 19(3), e95. https://doi.org/10.2196/jmir.6368
Cohen, R., Irwin, L., Newton-John, T., & Slater, A. (2019). #bodypositivity:
A content analysis of body positive accounts on Instagram. Body Image,
29(2019), 47–57. https://doi.org/10.1016/j.bodyim.2019.02.007
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. https://doi.org/10.1080/02650487.2017.1348035
Deighton-Smith, N., & Bell, B. T. (2018). Objectifying fitness: A content
and thematic analysis of #fitspiration images on social media.
Psychology of Popular Media Culture, 6(1), 1–18. https://doi.org/10.
1037/ppm0000143
Diaz-Granados, M., Diaz-Montes, J., & Parashar, M. (2015). Investigating
insurance fraud using social media. In Proceedings of 2015 IEEE
International Conference on Big Data (pp. 1344–1349). Santa Clara, CA:
IEEE Computer Society. https://doi.org/10.1109/BigData.2015.7363893
DiBartolo, P. M., Lin, L., Montoya, S., Neal, H., & Shaffer, C. (2007). Are
there “healthy” and “unhealthy” reasons for exercise? Examining indi-
vidual differences in exercise motivations using the function of exercise
scale. Journal of Clinical Sport Psychology, 1(2), 93–120. https://doi.org/
10.1123/jcsp.1.2.93
DiMaggio, P., Nag, M., & Blei, D. (2013). Exploiting affinities between
topic modeling and the sociological perspective on culture: Application
to newspaper coverage of US government arts funding. Poetics, 41(6),
570–606. https://doi.org/10.1016/j.poetic.2013.08.004
Easton, S., Morton, K., Tappy, Z., Francis, D., & Dennison, L. (2018).
Young people’s experiences of viewing the fitspiration social media
trend: Qualitative study. Journal of Medical Internet Research, 20(6),
e219. https://doi.org/10.2196/jmir.9156
Feltman, C. E., & Szymanski, D. M. (2018). Instagram use and
self-objectification: The roles of internalization, comparison, appear-
ance commentary, and feminism. Sex Roles, 78(5–6), 311–324. https://
doi.org/10.1007/s11199-017-0796-1
Fredrickson, B. L., & Roberts, T. A. (1997). Objectification theory: Toward
understanding women’s lived experiences and mental health risks.
Psychology of Women Quarterly, 21(2), 173206. https://doi.org/10.
1111/j.1471-6402.1997.tb00108.x
Goffman, E. (1959). The presentation of self in everyday life. Anchor Books.
Grimmer, J., & Stewart, B. M. (2013). Text as data: The promise and
pitfalls of automatic content analysis methods for political texts.
Political Analysis, 21(3), 267–297. https://doi.org/10.1093/pan/mps028
Heflick, N. A., & Goldenberg, J. L. (2014). Seeing eye to body: The literal
objectification of women. Current Directions in Psychological Science,
23(3), 225–229. https://doi.org/10.1177/0963721414531599
Jankowski, G. S., Diedrichs, P. C., & Halliwell, E. (2014). Can appearance
conversations explain differences between gay and heterosexual men’s
body dissatisfaction? Psychology of Men & Masculinity, 15(1), 68–77.
https://doi.org/10.1037/a0031796
Kim, J. L., Lynn Sorsoli, C., Collins, K., Zylbergold, B. A., Schooler, D., &
Tolman, D. L. (2007). From sex to sexuality: Exposing the heterosexual
script on primetime network television. Journal of Sex Research, 44(2),
145–157. https://doi.org/10.1080/00224490701263660
Leit, R. A., Pope, H. G., & Gray, J. J. (2001). Cultural expectations of
muscularity in men: The evolution of Playgirl centerfolds. International
Journal of Eating Disorders, 29(1), 90–93. https://doi.org/10.1002/1098-
108x(200101)29:1<90::aid-eat15>3.0.co;2-f
Lewis, S. P., & Arbuthnott, A. E. (2012). Searching for thinspiration: The
nature of internet searches for pro-eating disorder websites.
Cyberpsychology, Behavior and Social Networking, 15(4), 200–204.
https://doi.org/10.1089/cyber.2011.0453
Manago, A. M., Graham, M. B., Greenfield, P. M., & Salimkhan, G. (2008).
Self-presentation and gender on MySpace. Journal of Applied
Developmental Psychology, 29(6), 446–458. https://doi.org/10.1016/j.
appdev.2008.07.001
Moradi, B. (2010). Addressing gender and cultural diversity in body
image: Objectification theory as a framework for integrating theories
and grounding research. Sex Roles, 63(1–2), 138–148. https://doi.org/
10.1007/s11199-010-9824-0
Morry, M. M., & Staska, S. L. (2001). Magazine exposure: Internalization,
self-objectification, eating attitudes, and body satisfaction in male and
female university students. Canadian Journal of Behavioural Science, 33
(4), 269–279. https://doi.org/10.1037/h0087148
Muthukadan, B. (2018). Selenium with Python. https://selenium-python.
readthedocs.io
O’Neil, A., Sojo, V., Fileborn, B., Scovelle, A. J., & Milner, A. (2018). The#
MeToo movement: An opportunity in public health? The Lancet, 391
(10140), 2587–2589. https://doi.org/10.1016/s0140-6736(18)30991-7
Palmer, L. (2015). “Poppin’ bottles, getting wheysted.” Exploring young
men’s engagement with fitspiration content and its consequential influ-
ences on attitudes and behaviour. Journal of Promotional
Communications, 3(3), 425–445. https://promotionalcommunications.
org/index.php/pc/article/viewFile/66/79
Pilgrim, K., & Bohnet-Joschko, S. (2019). Selling health and happiness
how influencers communicate on Instagram about dieting and exercise:
Mixed methods research. BMC Public Health, 19(1), 1–9. https://doi.
org/10.1186/s12889-019-7387-8
Prichard, I., McLachlan, A., Lavis, T., & Tiggemann, M. (2018). The impact
of different forms of# fitspiration imagery on body image, mood, and
self-objectification among young women. Sex Roles, 78(11–12), 789–798.
https://doi.org/10.1007/s11199-017-0830-3
Prichard, I., & Tiggemann, M. (2005). Objectification in fitness centers:
Self-objectification, body dissatisfaction, and disordered eating in aero-
bic instructors and aerobic participants. Sex Roles, 53(1–2), 19–28.
https://doi.org/10.1007/s11199-005-4270-0
Raggatt, M., Wright, C. J., Carrotte, E., Jenkinson, R., Mulgrew, K.,
Prichard, I., & Lim, M. S. (2018). “I aspire to look and feel healthy like
the posts convey”: Engagement with fitness inspiration on social media
and perceptions of its influence on health and wellbeing. BMC Public
Health, 18(1), 1002–1013. https://doi.org/10.1186/s12889-018-5930-7
Řehůřek, R., & Sojka, P. (2010). Software framework for topic modeling
with large corpora. In R. Witte, H. Cunningham, J. Patrick,
E. Beisswanger, E. Buyko, U. Hahn, K. Verspoor, & A. R. Coden
(Eds.), Proceedings of the LREC 2010 workshop on new challenges for
NLP frameworks (pp. 45–50). Valletta, Malta: University of Malta.
Riddell, A., Hopper, T., & Grivas, A. (2016). Lda: 1.0.4. Zenodo. https://
doi.org/10.5281/zenodo.57927
Roberts, T. A., & Gettman, J. Y. (2004). Mere exposure: Gender differences
in the negative effects of priming a state of self-objectification. Sex Roles,
51(1–2), 17–27. https://doi.org/10.1023/b:sers.0000032306.20462.22
Robinson, L., Prichard, I., Nikolaidis, A., Drummond, C., Drummond, M.,
& Tiggemann, M. (2017). Idealized media images: The effect of fitspira-
tion imagery on body satisfaction and exercise behavior. Body Image,
22(2017), 65–71. https://doi.org/10.1016/j.bodyim.2017.06.001
Röder, M., Both, A., & Hinneburg, A. (2015). Exploring the space of topic
coherence measures. In Proceedings of the eighth ACM international
conference on web search and data mining (pp. 399–408). Shanghai,
China: Association for Computing Machinery. https://doi.org/10.1145/
2684822.2685324
Santarossa, S., Coyne, P., Lisinski, C., & Woodruff, S. J. (2019). # fitspo on
Instagram: A mixed-methods approach using Netlytic and photo ana-
lysis, uncovering the online discussion and author/image characteris-
tics. Journal of Health Psychology, 24(3), 376–385. https://doi.org/10.
1177/1359105316676334
HEALTH COMMUNICATION 11
Seidman, G., & Miller, O. S. (2013). Effects of gender and physical attrac-
tiveness on visual attention to Facebook profiles. Cyberpsychology,
Behavior and Social Networking, 16(1), 20–24. https://doi.org/10.
1089/cyber.2012.0305
Smith, A. C., & Stewart, B. (2012). Body perceptions and health behaviors
in an online bodybuilding community. Qualitative Health Research, 22
(7), 971–985. https://doi.org/10.1177/1049732312443425
Smolak, L., & Murnen, S. K. (2011). The sexualization of girls and women as
a primary antecedent of self-objectification. In R. M. Calogero,
S. Tantleff-Dunn, & J. K. Thompson (Eds.), Self-objectification in
women: Causes, consequences, and counteractions (pp. 3–21). American
Psychological Association. https://doi.org/10.1037/12304-003
Speno, G. A., & Aubrey, J. S. (2018). Sexualization, youthification, and
adultification: A content analysis of images of girls and women in
popular magazines. Journalism & Mass Communication Quarterly, 95
(3), 625–646. https://doi.org/10.1177/1077699017728918
Stankiewicz, J. M., & Rosselli, F. (2008). Women as sex objects and victims
in print advertisements. Sex Roles, 58(7–8), 579–589. https://doi.org/10.
1007/s11199-007-9359-1
Stephens, D. L., Hill, R. P., & Hanson, C. (1994). The beauty myth and
female consumers: The controversial role of advertising. Journal of
Consumer Aairs, 28(1), 137–153. https://doi.org/10.1111/
j.1745–6606.1994.tb00819.x
Thompson, J. K., Heinberg, L. J., Altabe, M., & Tantleff-Dunn, S. (1999).
Exacting beauty: Theory, assessment, and treatment of body image
disturbance. American Psychological Association.
Tiggemann, M. (2003). Media exposure, body dissatisfaction and disordered
eating: Television and magazines are not the same! European Eating
Disorders Review, 11(5), 418–430. https://doi.org/10.1002/erv.502
Tiggemann, M., & Barbato, I. (2018). “You look great!”: The effect of
viewing appearance-related Instagram comments on women’s body
image. Body Image, 27(2018), 61–66. https://doi.org/10.1016/j.
bodyim.2018.08.009
Tiggemann, M., & Boundy, M. (2008). Effect of environment and
appearance compliment on college women’s self-objectification,
mood, body shame, and cognitive performance. Psychology of
Women Quarterly, 32(4), 399–405. https://doi.org/10.1111/j.1471-
6402.2008.00453.x
Tiggemann, M., & Zaccardo, M. (2015). “Exercise to be fit, not skinny”:
The effect of fitspiration imagery on women’s body image. Body Image,
15(2015), 61–67. https://doi.org/10.1016/j.bodyim.2015.06.003
Tiggemann, M., & Zaccardo, M. (2018). ‘Strong is the new skinny’:
A content analysis of# fitspiration images on Instagram. Journal of
Health Psychology, 23(8), 1003–1011. https://doi.org/10.1177/
1359105316639436
Toma, C. L., & Hancock, J. T. (2010). Looks and lies: The role of physical
attractiveness in online dating self-presentation and deception.
Communication Research, 37(3), 335–351. https://doi.org/10.1177/
0093650209356437
Ward, L. M., Lippman, J., Erickson, S., & Giaccardi, S. (2016). Sexual
media content and effects. In J. F. Nussbaum (Ed.), Oxford research
encyclopedia of communication (pp. 1–39). Oxford University Press.
12 V. MURASHKA ET AL.
