Psychology of Popular Media

© 2022 American Psychological Association 2024, Vol. 13, No. 2, 249–255

ISSN: 2689-6567 https://doi.org/10.1037/ppm0000443

BRIEF REPORT

I’ll See Your Beautified Photo and Raise You One: An Experimental
Investigation of the Effect of Edited Social Media Photo Exposure
Wendy L. Wolfe and Leah Yakabovits
Department of Psychology, Georgia Southern University
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Social networking sites (SNSs) provide users with the opportunity to view photos and posts shared by celeb-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

rities and friends as well as share their own images. These unique aspects of SNSs may increase the risk of
negative psychological effects of social comparisons. Upward social comparisons on SNSs are particularly
likely, given the use of photo editing to create “beautified” images before posting. Although relatively new as
an area of experimental investigation, there is indication that exposure to edited SNS photos harms women’s
body image. The findings from experimental studies on the effect of editing photos of one’s self have been
more inconclusive. This experimental study of 95 undergraduate women examined both the effect of expo-
sure to edited photos of women embedded within an SNS mock-up and the subsequent effect on their photo
taking and editing. The effect of photo editing on perceived physical attractiveness and mood was also
assessed. Although type of SNS exposure did not affect perceived physical attractiveness, mood, or number
of photos taken, participants who viewed an SNS page with edited photos were significantly more likely to
edit their photos after being asked to take a selfie than participants who viewed an SNS page with unedited
photos of the same women and a control group. Across conditions, photo editing was associated with
adverse changes in perceived attractiveness and mood. It is recommended that future research continues
to examine experimental manipulations of SNS exposure and SNS activity that more closely reflect the inter-
active ways that people engage with SNSs.

Public Policy Relevance Statement

As photo editing applications proliferate and become integrated as a basic function within phone cam-
eras, we are likely to increasingly encounter edited images on social media platforms. This study of col-
lege women finds that exposure to edited images increases the likelihood of editing one’s own photos,
and also that photo editing has a negative effect on perceived physical appearance and mood. It is impor-
tant to better understand the complex interactive nature with which people engage with social media, and
to develop and study interventions to mitigate the adverse effect of exposure.

Keywords: social media, social networking, photo editing, photo manipulation, physical appearance

Multiple studies indicate that media exposure can negatively However, these unique aspects of SNSs also may increase the risk of
affect women’s body image (e.g., Grabe et al., 2008; Groesz et al., the negative psychological effects of social comparisons because the
2002). Increasingly, social media platforms have usurped traditional comparisons being made are within one’s peer group. Consistent
media outlets, particularly in use by young adults. Social networking with social comparison theory, comparisons to others perceived as
sites (SNSs) provide users with the opportunity to view photos and similar are more likely to affect self-evaluation (Tiggemann et al.,
posts shared by both celebrities and friends, share their own images 2020). A neighbor or former classmate who posts a beautiful selfie
and updates, and interact with others through “likes” and comments. or who regularly receives many “likes” and positive comments

This article was published Online First November 28, 2022. upon request.
Wendy L. Wolfe https://orcid.org/0000-0002-4785-5458 Wendy Wolfe was the lead author of the article. Leah Yakabovits assisted
The authors would like to thank Sarah Deslauriers, Alice Ionescu, Takera with writing the method section and with proofreading and providing sug-
Molton, Amber Reagan, and Ariyana Smith for their assistance with the study. gested changes to drafts of the article.
The authors have no known conflicts of interest to disclose, nor are there Correspondence concerning this article should be addressed to Wendy
funding sources to report. L. Wolfe, Department of Psychology, Georgia Southern University, 11935
Data, methods, and materials used to conduct this research are available Abercorn Street, Savannah, GA 31419. Email: wlwolfe@georgiasouthern.edu

249
 250 WOLFE AND YAKABOVITS

may, therefore, be more threatening to the maintenance of a positive (regardless of editing). However, participants who edited their pho-
self-evaluation than a celebrity who seems to have a very different tos without posting them experienced a decrease in body/weight pre-
life and resources. occupation and dissatisfaction, whereas participants who edited and
Holland and Tiggemann (2016) conducted a systematic review of posted their photos experienced an increase. Finally, Tiggemann
research on SNS use and body image and disordered eating out- et al. (2020) exposed female undergraduates to thin or average-sized
comes. Across correlational, experimental, and longitudinal studies, images of women on Instagram, followed by selfie-taking and
results consistently illustrated an association between SNS use and optional editing. Although photo exposure was not associated with
increased body image concerns and disordered eating. Moreover, a change in appearance dissatisfaction, mood, or photo taking or
among SNS activities, photo-related activities have been found to editing behavior, selfie editing was associated with an increase
be most strongly related to body image concerns. In their cross- in negative mood and appearance dissatisfaction, regardless of
sectional survey of adolescent females, Meier and Gray (2014) condition.
found that of different types of activity on Facebook, only photo- Thus far, the relationship between photo-related activity on SNSs
related activity was significantly correlated with weight dissatisfac- and adverse psychological impact has primarily been examined
tion, drive for thinness, thin-ideal internalization, and self- cross-sectionally. Among the experimental studies that have been
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

objectification. Fox and Vendemia (2016) examined different conducted, exposure to edited photos (as opposed to thin vs.
types of SNS activity in adult women and men and found that average-sized images of different women, as in the Tiggemann
women were significantly more likely than men to report that view- et al., 2020 study) and editing of photos of one’s self have not
ing attractive photos of the same sex (upward social comparisons) been examined in combination. There is some experimental support
left them feeling worse about their own bodies. Tiggemann et al. for the adverse effect of edited photo exposure on body image
(2018) assessed the effect of thin-ideal versus average Instagram (Kleemans et al., 2018). One unexamined effect of edited photo
images of women in an experimental design and found that females exposure is its effect on the recipient’s photo behaviors. Exposure
exposed to the thin-ideal images were more likely to experience an to beautified photos may prompt upward social comparisons, poten-
increase in body dissatisfaction after viewing the photos. tially increasing the likelihood of negative self-evaluation and a
Upward social comparisons with peers on SNSs are particularly more critical view towards one’s own appearance in photos. This
likely, given the use of photo editing to create beautified images may result in an increased tendency to take multiple selfies in
before posting (sometimes referred to in the literature as “photo order to select one deemed worthy of sharing (a behavioral manifes-
manipulation”). Kleemans et al. (2018) randomly assigned adoles- tation of photo investment, see McLean et al., 2015) and/or an
cent girls to either view 10 original photos (face and full body) or increased likelihood of editing (and time spent editing) one’s own
10 edited versions of the originals, presented to participants in the photos. Whether photo selection and editing then serve to further
form of Instagram selfies. Participants who viewed the edited photos increase negative feelings of one’s own attractiveness or serves
were more likely to experience a negative state body image, and this a reparative function is unclear, given the disparate findings in
was particularly true for participants with higher social comparison prior research. However, this question too is worth exploring,
tendencies. The effect of editing one’s own photos has also been given that it more closely reflects the interactive way that individuals
investigated, although the findings have been less consistent. use social media. We hypothesized that female participants exposed
McLean et al. (2015) surveyed adolescent girls about their SNS to control (non-human) images and to unedited female images on
activity, body image, and dietary restraint. They found that photo a mock-up of an SNS would experience a less adverse change
editing and photo sharing were positively correlated with body dis- in physical appearance and mood ratings (Hypothesis 1) and
satisfaction and dietary restraint. Similarly, Chang et al. (2019) would be less likely to engage in photo selection behavior (taking
found that photo editing was negatively associated with body esteem multiple face and full-body selfies) and photo editing (Hypothesis
in a survey of adolescent females in Singapore and that the relation- 2), compared with participants exposed to edited images of female
ship was fully mediated by peer appearance comparisons. Three faces and bodies. We also explored the effect of photo editing on
studies have examined the effect of photo editing experimentally. changes in mood and perceived physical attractiveness (Research
Mills et al. (2018) randomly assigned female undergraduate SNS Question).
users to either take and upload an unedited selfie to their own SNS
profile; take, edit, and upload a selfie; or read an SNS article about
a non-appearance-related topic. Comparison of pre- and post- Method
manipulation Visual Analog Scale (VAS) ratings illustrated that tak- Participants
ing and posting a selfie, regardless of editing, was associated with
significantly greater decreases in feelings of attractiveness, than for Participants were 123 female undergraduates at a public university
control participants. However, participants who edited their photos in the southeastern United States who reported current or previous
prior to posting them were not significantly different from control social media use. Twenty-eight were excluded due to either quick
group participants on change in VAS ratings of anxiety and confi- (n = 3) or random responses (n = 1), accurate identification of the
dence. Wick & Keel (2020) randomly assigned male and female study’s hypothesis (n = 7), or because of outlier responses poten-
undergraduates who reported editing photos on Instagram to tially due to error on measures of the dependent variables (n = 10)
either post an unedited full body photo, edit and post a full body or technical/other issues that prevented study completion (n = 7).
photo, edit a full body photo (no posting), or complete a control The reasonable duration for completing the survey while paying suf-
task after the photo was taken. Comparison of pre- and post- ficient attention to the questions was determined by pilot research
manipulation VAS ratings highlighted increases in body/weight pre- instructing survey respondents to complete the survey and research
occupation and dissatisfaction for those who posted their photos tasks “as quickly as possible.” On this basis, a completion time of
 EFFECT OF EDITED PHOTO EXPOSURE 251

20 min was determined to be the minimum time needed to complete and duration of use, and type of usage to provide descriptive infor-
the survey and experimental tasks. Due to concern that participants mation about participants’ SNS use. This English language ques-
with faster completion times likely did so by not reading instructions tionnaire has been published and validated by using a Chinese
or survey items, those completing the study in less than 20 min were college student sample (Shi et al., 2014). Questions were modified
excluded from data analysis. Random responses were determined in this study to assess SNSs that are particularly relevant to an
from an obvious lack of attention to the anchors on the VAS, such American college sample (e.g., Facebook, Instagram, Twitter,
that all items in a scale were rated the same way (e.g., indicating Snapchat). Participants also completed a demographic questionnaire
both a high level of happiness and a high level of boredom by plac- that included questions about age and race.
ing the VAS slider to the far left of the scale). Outlier responses were
determined based on inter-quartile range × 1.5 for the number of Materials
photos taken and total time spent editing photos, because extreme
high responses may have indicated user error with the equipment The survey was administered on a desktop computer. Photo taking
and software. After exclusions, 95 participants remained (n = 23 was completed on an iPad by using the YouCam Perfect™ app
control, 39 edited, 33 unedited), with no significant differences in (2019). Optional editing features were available to all participants
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

proportion eliminated or retained across condition (χ2 (2) = 5.97,

This document is copyrighted by the American Psychological Association or one of its allied publishers.

within YouCam Perfect™ (2019), which offers dozens of ways to

p . .05). Of these, 51% self-identified as Caucasian/White, 29.5% enhance images, including slimming and other modifications of
as African American/Black, 9.5% as Hispanic/Latina, 7.4% as body shape, skin-tone changes, complexion clearing, and others.
Asian/Pacific Islander, and 2.2% as other. Participants ranged
from 18 to 49 years of age (M = 21.51, SD = 4.93). All participants Procedure
received class credit for participation.
The Institutional Review Board-approved investigation was con-
ducted from March 2019 to March 2020. Before the participant
Measures and Materials entered the lab, a researcher turned on the screen-recording feature
Measures of the iPad. A discreet piece of tape with the lab room number cov-
ered the red light indicating that the recording was in progress.
VASs were used to measure participants’ perception of their Female researchers wore lab coats to minimize the impact of the
physical appearance (Physically Unattractive–Physically Attractive), researcher’s gender and appearance on participants. A researcher
mood (Happy–Sad), and distractor variables such as hunger verbally reviewed the informed consent document with each partic-
(Hungry–Full) and boredom (Bored–Interested). Participants rated ipant, which presented the study as examining social media prefer-
how they were feeling at the moment by sliding an indicator across ences and interactions with different types of stimuli (photo and
an unmarked linear scale (corresponding to 0–100) with opposing text) on social media-type platforms and informed participants of
adjectives as anchors at either end. Because these measures were the possibility that photo and video recording might be used.
repeated several times in the survey, the VAS format for item presen- Participants completed a survey on a desktop computer in a private
tation (as opposed to a Likert-type scale with numbers or descriptors, research lab with the researcher available outside of the room to
such as 4 = somewhat true) decreased the likelihood that participants assist, if necessary. Each survey was administered in Qualtrics and
would recall their prior responses. Scores from the VAS were used to began with the first series of VASs. Next, participants advanced to
measure the change in perceived physical attractiveness and mood a screen containing a condition-specific mock-up of a generic social
from baseline, after exposure to the SNS mock-up, and after taking media newsfeed (not branded with Facebook, Instagram, etc.) based
a selfie. on their random assignment of either “Edited Photo,” “Unedited
The 11-item, Physical Appearance Comparison Scale-Revised Photo,” or “Control” condition (a portion of the Edited Photo social
(PACS-R; Schaefer & Thompson, 2014) was used to measure how media newsfeed is provided in Figure 1). All topographical features
often participants “generally compare themselves to others” in var- and text content of the pages were identical between the conditions
ious circumstances (e.g., at work, in public, at a social gathering) other than the photos on the page. Participants in the edited photo
using a 5-point Likert-type scale ranging 0 (never) to 4 (always). condition viewed a page containing edited facial (n = 6) and full-
We anticipated that participants’ general tendency to compare body (n = 4) photos of young adult females, posted as either status
their appearance to others would influence their reactions to viewing updates with text (e.g., “I’m ready for the weekend!”) or as profile
social media images in this study, as has been found in other inves- photo updates, interspersed with non-photo memes (n = 2) to
tigations (Chang et al., 2019; Fox & Vendemia, 2016). Because enhance ecological validity. Edited photos were “beautified” using
physical appearance comparison tendency was not possible to assess several photo editing applications to enhance complexion and facial
and match across conditions prior to random assignment, it was features and adjust body features in alignment with typical, online
anticipated that PACS-R scores would be used as a covariate if beauty “norms” (lengthen legs, narrow waist, enhance bust).
experimental conditions significantly differed on this variable. The Participants in the unedited photo condition viewed a page contain-
PACS-R has been validated in samples of female college students ing original, “unedited” versions of the same photos of young adult
and has shown excellent internal consistency, and convergent valid- females as in the edited photo condition, along with the same non-
ity with measures of body satisfaction and self-esteem (Schaefer & photo memes. Images of ethnically diverse females and females
Thompson, 2014). In the present study, Cronbach’s alpha for the within the typical student age range were selected from websites fea-
measure was .95. turing non-copyright images in the public domain to maximize the
The Social Networking Sites Usage Questionnaire (Shi et al., likelihood of personal comparisons by our student sample.
2014) was used to assess the SNSs participants’ use, frequency Although the edited versus unedited comparison photos were not
 252 WOLFE AND YAKABOVITS

Figure 1 body and other stereotypical Western standards of female beauty.

Portion of the Social Media Newsfeed in the Edited Photo Participants in the control condition viewed a page with photos
Condition. of animals, objects, scenery, and the two memes (n = 12).
Participants were instructed to take as much time as desired while
scrolling from the top of the page to the bottom of the page and to
respond to each prompt for rating. To enhance the credibility of
the study as investigating, “Social Media Preferences and Inter-
actions,” participants were prompted to rate the likelihood of giving
the post a “like” or “thumbs up” after each of the “posts” using a
Likert-type rating scale from 0 to 100 (very unlikely–very likely).
The number of “likes” from “others” was not provided, nor were
comments. An additional task intended to increase the credibility
of the declared investigation topic was presented, which asked the
participants to record their impressions of the person to which the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

social media newsfeed belonged, with another slider-style scale

from 0 to 100. Sample impression items include popularity (unpop-
ular–popular), education (educated–uneducated), and personality
(introverted–extroverted).
After interacting with the social media newsfeed, participants
completed similar VAS items as at the beginning of the survey. To
measure the effect of the photo exposure on participants’ photo
selection and editing, all participants were provided with a prompt
explaining that they had been selected to take and share two ‘selfie’
photos. Participants were directed to open the YouCam Perfect™
(2019) photo app on an iPad provided next to the computer, take a
headshot, then stand at a marked distance across from the iPad and
use the timer feature for a full body shot. Participants were informed
that they could take as many photos as desired and that they could
use the YouCam Perfect™ (2019) features to edit their photos if
desired. They were instructed to select one headshot and one body
shot to “share.” The directive to share the photo was intended to sim-
ulate posting a “selfie” in order to encourage editing behaviors that
may occur under the presumption that the shared photos may
become public. Photos were shared by selecting one headshot and
one body shot in the “photos” folder on the iPad, which the partic-
ipant was instructed to upload to the Google drive associated with
the research team. Upon saving their photo in the YouCam
Perfect™ (2019) app with the date and time of participation, partic-
ipants proceeded with the rest of the survey.
Following the photo taking and “sharing” process, participants
completed the final set of VAS items, the PACS-R (positioned
after measures of all dependent variables to mitigate demand charac-
teristics), the SNSs Usage Questionnaire, a demographic question-
naire, and two final manipulation check questions inquiring about
participants’ beliefs about the purpose and hypotheses of the
study. When finished, participants were prompted to alert the
researcher to the completion of their survey. During debriefing, par-
ticipants were made aware that screen recording of the iPad was uti-
lized in order to measure the number of photos taken and time spent
editing photos before sharing, and that recording would be deleted
after coding.
The photo-taking, editing, and selection process were recorded by
Note. Second image has been altered for publication, but original study using the screen recording feature on the iPad. After the participant
materials are available upon request. See the online article for the color was debriefed and excused, the researcher reviewed the recording
version of this figure. and documented the number of photos taken and the time spent edit-
ing photos. The participants’ selected photos and (if edited) their
piloted, four female undergraduate research assistants were in agree- original photos were saved to a password-protected file folder. All
ment with the authors that the photo pairs were visibly different screen recordings from the experimental session were deleted imme-
and that the edited photos were more consistent with the thin-ideal diately after review.
 EFFECT OF EDITED PHOTO EXPOSURE 253

Results two exceptions. When a covariate was not used, the MANCOVA
for time spent editing face photos indicated a significant difference
Preliminary Analyses across conditions, F(2, 91) = 3.53, p = .03, ηp 2 = .07, and for body
As anticipated, participants from this college student sample are photos a marginally significant difference, F(2, 91) = 3.11, p = .05,
regular users of social media. Ninety percent reported SNSs use ηp 2 = .06, both with regard to greater time editing by participants
daily or multiple times daily, with the most frequently used “favor- in the edited condition.) A comparison of participants randomly
ite” platforms being Instagram (33%), Snapchat (31%), and assigned to the three conditions illustrated that the conditions did
Facebook (24%). Of the more photo-based SNS platforms, 96% of not significantly differ in age, F(2, 92) = 0.50, p = .61, race,
participants reported being current or past users of Instagram (92% χ2(10) = 8.31, p = .60, SNS frequency, F(2, 92) = 2.06, p = .13, typ-
for Facebook), and 70% reported sharing photos on SNS at least ical daily duration of use, F(2, 92) = 1.71, p = .19, baseline VAS
monthly. Also as expected, PACS-R scores were significantly nega- physical attractiveness, F(2, 92) = 0.68, p = .51, or mood ratings,
tively correlated with VAS physical attractiveness ratings at baseline F(2, 92) = 0.003, p = .99.
(r = –.27, p = .004), were significantly positively correlated with
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

total time participants spent editing their photos (r = .27, p = .004), Primary Analyses
This document is copyrighted by the American Psychological Association or one of its allied publishers.

and differed significantly across conditions, F(2, 92) = 3.72, p = .03,

such that participants in the edited condition (M = 2.34, SD = 0.95) To test Hypothesis 1, a mixed model analysis of covariance
were significantly higher in appearance comparison tendency than (ANCOVA) was conducted, with VAS attractiveness ratings at
participants in the unedited condition (M = 1.73, SD = 0.91), and Time 1 (baseline) and Time 2 (post-social media exposure) as a
with control participants at an intermediary level (M = 2.08, SD = repeated measure, condition as a between groups variable, and
0.99). Therefore, PACS-R scores were used as a covariate in the PACS-R as a covariate. Results did not support a significant time
primary analyses. (Data analyses conducted without inclusion by condition interaction, F(2, 91) = 2.48, p = .09, ηp 2 = .05. A simi-
of PACS-R as a covariate yielded similar findings for the primary lar mixed model ANCOVA was used to test the effect of the different
analyses, in terms of significance versus non-significance, with forms of social media exposure on participants’ VAS mood ratings.

Table 1
Means and Estimated Marginal Means (EM) Using PACS-R as Covariate
Condition
Control Edited Unedited
Measurement time
and editing status M (SD) EM (SE) M (SD) EM (SE) M (SD) EM (SE)
VAS appearance
Time 1 61.61 (19.69) 61.72 (5.25)a 54.28 (28.61) 56.23 (4.11)a 59.52 (26.37) 57.13 (4.48)a
56.98 (5.09)a; 57.21 (3.98)a; 60.86 (4.35)a;
Time 2 56.83 (22.29) 56.73 (5.18)b 54.54 (27.06) 58.51 (5.48)b 64.12 (26.99) 63.19 (4.95)b
Edit Yes 57.54 (18.33) 58.33 (6.84)b 51.59 (27.03) 55.33 (4.51)b 60.88 (22.90) 58.49 (5.08)b
Edit No 55.90 (27.66) 55.13 (7.80)b 65.17 (25.67) 61.69 (10.12)b 72.78 (35.93) 67.88 (8.35)b
Time 3 54.91 (19.75) 55.79 (5.51)b 46.61 (28.37) 50.81 (5.83)b 63.27 (31.07) 64.06 (5.26)b
Edit Yes 49.15 (19.17) 49.89 (7.27)b 44.72 (29.19) 48.21 (4.79)b 58.25 (26.67) 56.02 (5.41)b
Edit No 62.40 (18.80) 61.68 (8.29)b 56.67 (23.06) 53.42 (10.76)b 76.67 (39.23) 72.10 (8.88)b
VAS mood
Time 1 72.96 (17.62) 73.04 (3.96)a 72.95 (22.61) 74.37 (3.09)a 73.27 (16.48) 71.54 (3.38)a
67.87 (4.11)a; 73.94 (3.21)a; 73.53 (3.51)a;
Time 2 67.78 (22.90) 67.13 (4.17)b 72.41 (20.86) 73.02 (4.41)b 75.39 (17.51) 73.84 (3.98)b
Edit Yes 72.15 (18.37) 72.64 (5.50)b 71.22 (21.78) 73.51 (3.62)b 74.58 (15.35) 73.12 (4.09)b
Edit No 62.10 (27.73) 61.63 (6.27)b 74.67 (15.11) 72.53 (8.14)b 77.56 (23.26) 74.56 (6.72)b
Time 3 72.35 (16.39) 72.35 (3.96)b 66.24 (23.13) 72.66 (4.18)b 76.55 (18.37) 76.34 (3.77)b
Edit Yes 72.38 (16.50) 72.93 (5.22)b 63.31 (23.71) 65.90 (3.44)b 73.29 (16.95) 71.64 (3.88)b
Edit No 72.30 (17.13) 71.77 (5.95)b 81.83 (11.29) 79.42 (7.72)b 85.22 (20.19) 81.83 (6.37)b
# Photos
Face 2.39 (1.50) 2.39 (0.33) 2.24 (1.34) 2.21 (0.26) 2.58 (1.82) 2.60 (0.28)
Body 2.22 (1.73) 2.22 (0.32) 2.34 (1.24) 2.29 (0.25) 2.39 (1.64) 2.46 (0.27)
Total 4.61 (2.84) 4.61 (0.53) 4.58 (2.08) 4.50 (0.42) 4.97 (2.81) 5.06 (0.45)
Edit time
Face 43.87 (69.54) 43.64 (16.69) 97.84 (92.96) 92.37 (13.23) 61.70 (75.10) 68.16 (14.25)
Body 29.00 (48.79) 28.86 (13.03) 66.39 (77.34) 63.07 (10.33) 37.45 (52.69) 41.38 (11.13)
Total 72.09 (104.5) 71.71 (25.66) 165.26 (149.02) 156.39 (20.34) 98.03 (108.95) 108.51 (21.91)
Note. EM = estimated marginal means; PACS-R = Physical Appearance Comparison Scale-Revised; VAS = Visual Analog Scale. VAS appearance and VAS
mood were scaled such that higher scores indicate a more positive view of one’s physical appearance and a more positive (“happy”) mood. Edit time reflects time
editing in seconds.
a
Estimated marginal mean and associated standard error scores are from the ANCOVA comparing Time 1 (baseline) with Time 2 (post-social media exposure)
scores.
b
Estimated marginal mean and associated standard error scores are from the ANCOVA comparing Time 2 (post-social media exposure) with Time 3 (post-photo
taking and editing) scores.
 254 WOLFE AND YAKABOVITS

Participants in the three conditions did not significantly differ in their their photos experienced an increase in happiness (Time 2 [M =
change in VAS ratings on the mood item from baseline to post-social 70.68, SD = 23.84], Time 3 (M = 79.24, SD = 17.54)], resulting in
media exposure, F(2, 91) = 1.90, p = .16, ηp 2 = .04. Descriptive sta- a significant difference between those who edited and those who
tistics are presented in Table 1. did not F(1, 91) = 8.40, p = .005, ηp 2 = .08.
To test Hypothesis 2, a multivariate analysis of covariance
(MANCOVA) was used to determine the effect of condition on Exploratory Analyses
photo selection and photo editing, with PACS-R scores again used
as a covariate. There were no significant differences in overall Given the age diversity of our sample, and the greater likelihood
photo selection, F(2, 90) = 0.43, p = .65, ηp 2 = .01, as measured by of traditionally aged college students making appearance compari-
the total number of photos taken, or by the number of face selfies, sons with the similarly aged images on the SNS newsfeed, a series
F(2, 90) = 0.50, p = .61, ηp 2 = .01, or body selfies, F(2, 90) = 0.19, of post hoc analyses were conducted with age as an additional covar-
p = .83, ηp 2 = .004. Participants in the edited condition took as iate (age was not significantly correlated with PACS-R scores: r(93)
many selfies after social media exposure as participants in the un- = –.11, p = .28). Results of these analyses paralleled the above find-
edited and control conditions. There was a significant effect of con- ings. When using age as the only covariate, results also paralleled the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

dition on overall photo editing, F(2, 90) = 3.48, p = .03, ηp 2 = .07, above findings, with one exception. Taking into account age as a
This document is copyrighted by the American Psychological Association or one of its allied publishers.

such that participants exposed to the edited photos on the social covariate in testing Hypothesis 2 resulted in a significant difference
media page spent significantly more time editing their photos than across the conditions in time spent editing face, F(2, 90) = 3.59, p
participants exposed to unedited photos or participants in the control = .03, ηp 2 = .07, and body, F(2, 90) = 3.12, p = .049, ηp 2 = .07, pho-
condition. As can be seen in Table 1, although this same pattern tos, when considered separately, as well as time spent editing all
emerged when face, F(2, 90) = 2.65, p = .08, ηp 2 = .06, and body, photos, F(2, 90) = 4.68, p = .01, ηp 2 = .09, with participants exposed
F(2, 90) = 2.28, p = .11, ηp 2 = .05, photo editing was examined sep- to the edited photos subsequently spending more time editing both
arately, the differences were not statistically significant. their face selfies and their full-body photos.
Sixty-nine participants edited their selfies and 25 participants did
not. We explored whether photo editing served a detrimental or Discussion
reparative function for our participants (research question) by con-
ducting a mixed model ANCOVA using VAS attractiveness ratings This is the first known study to investigate both the effects of
at Time 2 (post-social media exposure) and Time 3 (post-photo tak- exposure to edited photos and spontaneous editing behavior of par-
ing/editing) as a repeated measure, condition and use of editing soft- ticipants using an experimental design. Results suggest a potential
ware as between groups variables, and PACS-R score as a covariate. impact of exposure to edited SNS photos and a detrimental impact
There was a marginally significant main effect of photo editing on of photo editing. Although viewing edited (beautified) photos
change in participant self-ratings of attractiveness, F(1, 87) = 3.52, embedded within an SNS newsfeed had no impact on participant
p = .06, ηp 2 = .04, but not a significant main effect of condition, F mood or appearance ratings, this exposure was subsequently associ-
(2, 87) = 2.01, p = .14, ηp 2 = .04, or interaction between condition ated with greater editing of selfies taken by participants. Viewing
and editing on change in VAS attractiveness ratings, F(2, 87) = edited photos may normalize this behavior. More problematically,
1.57, p = .21, ηp 2 = .04. Collapsing across conditions and comparing exposure to such images may prompt upward social comparisons
the change in VAS attractiveness ratings between participants who that females counter through creating an altered image of themselves
edited their photos with those who did not highlighted that regard- to “share” with others on SNS platforms. Instead of serving a repar-
less of the type of social media images participants were exposed ative function, photo editing was associated with decreased ratings
to before taking facial and full-body selfies, those who edited their of personal attractiveness and a worsening of mood state for partic-
selfies were more likely to experience a decrease in self-reported ipants, consistent with findings reported by Tiggemann et al. (2020).
physical attractiveness after taking and editing their photos (Time This was true regardless of the type of images participants viewed in
2 [M = 55.94, SD = 24.24], Time 3 [M = 50.26, SD = 27.04]). the SNS mock-up. For this sample of frequent SNS users, engrained
Participants who did not edit their photos reported a slight increase tendencies to manipulate photos and greater familiarity with doing
in perceived attractiveness [Time 2 (M = 64.20, SD = 30.19), Time 3 so may have prompted some of our participants to edit their photos
(M = 66.16, SD = 28.74)], resulting in a significant difference after taking them, even after viewing unedited photos or neutral
between those who edited and those who did not, F(1, 91) = 4.73, images. However, regardless of condition, after taking selfies and
p = .03, ηp 2 = .05. A similar mixed model ANCOVA was used to then editing them, participants experienced a detrimental effect on
examine the effect of editing on change in VAS mood scores from their perception of their physical attractiveness and on their mood.
Time 2 to Time 3. There was a significant main effect of photo edit- Photo editing may increase the salience of perceived imperfections
ing on change in participant self-ratings of mood, F(1, 87) = 6.47, such that even as participants beautify their image, they reconfirm
p = .01, ηp 2 = .07, but not a significant main effect of condition, that there is a need to make such adjustments.
F(2, 87) = 0.54, p = .59, ηp 2 = .01, or interaction between condition The minimal impact of SNS exposure to beautified photos on
and editing on change in VAS mood ratings, F(2, 87) = 0.16, p = .85, physical attractiveness and mood ratings was surprising and runs
ηp 2 = .004. Collapsing across conditions and comparing the change counter to Kleemans et al.’s (2018) findings, as well as experiments
in VAS mood ratings between participants who edited their photos involving exposure to thin-ideal images versus comparison images,
with those who did not highlighted that those who edited their selfies such as Tiggemann et al.’s (2018). VASs were used to measure
were more likely to experience a decrease in happiness after taking attractiveness and mood in this investigation in order to minimize
and then editing their photos (Time 2 [M = 72.57, SD = 18.92], Time the potential for the influence of memory effect bias for variables
3 [M = 68.49, SD = 20.62]), whereas participants who did not edit measured repeatedly, as well as demand characteristics alerting
 EFFECT OF EDITED PHOTO EXPOSURE 255

participants to the hypotheses of the study. However, this choice Grabe, S., Ward, L. M., & Hyde, J. S. (2008). The role of the media in body
may have prevented a robust assessment of these variables of inter- image concerns among women: A meta-analysis of experimental and cor-
est. In contrast, Kleemans et al. (2018) used the 11-item Body Image relational studies. Psychological Bulletin, 134(3), 460–476. https://
State Scale in their between groups design investigation. Noting doi.org/10.1037/0033-2909.134.3.460
Groesz, L. M., Levine, M. P., & Murnen, S. K. (2002). The effect of exper-
the advantages of VAS items when conducting repeated measure-
imental presentation of thin media images on body satisfaction: A meta-
ments, Tiggemann et al. (2018) also measured dependent variables
analytic review. International Journal of Eating Disorders, 31(1), 1–16.
by using VASs but included separate VAS items to assess weight
https://doi.org/10.1002/eat.10005
dissatisfaction, appearance dissatisfaction, and facial appearance Holland, G., & Tiggemann, M. (2016). A systematic review of the impact
dissatisfaction. of the use of social networking sites on body image and disordered eating
It should be noted that there was a trend for condition to impact outcomes. Body Image, 17, 100–110. https://doi.org/10.1016/j.bodyim
change in physical attractiveness ratings in this study, as well as .2016.02.008
an almost statistically significant effect of condition on time editing Kleemans, M., Daalmans, S., Carbaat, I., & Anschütz, D. (2018). Picture per-
both face and body photos (considered separately, rather than com- fect: The direct effect of manipulated Instagram photos on body image in
bined). It is possible that these trends would have reached statistical adolescent girls. Media Psychology, 21(1), 93–110. https://doi.org/10
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

significance with a larger sample size. However, data collection was .1080/15213269.2016.1257392
interrupted by the COVID-19 pandemic, introducing a lapse of McLean, S. A., Paxton, S. J., Wertheim, E. H., & Masters, J. (2015).
in-person research for a year, new confounds, and new required Photoshopping the selfie: Self photo editing and photo investment are
associated with body dissatisfaction in adolescent girls. International
IRB protocols that prevented reasonable resumption of the investiga-
Journal of Eating Disorders, 48(8), 1132–1140. https://doi.org/10.1002/
tion. (An a priori G-power analysis by using a moderate effect size
eat.22449
predicted a sample size of N = 251 to provide sufficient power Meier, E. P., & Gray, J. (2014). Facebook photo activity associated with body
for an ANCOVA with three conditions and one covariate.) image disturbance in adolescent girls. Cyberpsychology, Behavior, and
Nevertheless, this study adds to the literature by testing both the Social Networking, 17(4), 199–206. https://doi.org/10.1089/cyber.2013
effect of exposure to edited photos and photo selection and manip- .0305
ulation behavior by using an ethnically diverse sample of young- to Mills, J. S., Musto, S., Williams, L., & Tiggemann, M. (2018). “Selfie” harm:
middle-aged adult females. It is recommended that future research Effects on mood and body image in young women. Body Image, 27,
investigations continue to examine experimental manipulations of 86–92. https://doi.org/10.1016/j.bodyim.2018.08.007
both SNS exposure and SNS activity that more closely reflect the Schaefer, L. M., & Thompson, J. K. (2014). The development and validation
interactive ways that people engage with social media. Moreover, of the Physical Appearance Comparison Scale-Revised (PACS-R). Eating
as evidence mounts regarding the potential adverse impact of Behaviors, 15(2), 209–217. https://doi.org/10.1016/j.eatbeh.2014.01.001
Shi, Y., Luo, Y. L. L., Yang, Z., Liu, Y., & Cai, H. (2014, June 22–27). The
photo editing on mood and perceptions of personal appearance,
development and validation of the social networking sites (SNSs) usage
research should extend into examining the specific aspects of
questionnaire [Paper presentation]. International Conference on Social
photo editing that are most problematic. For example, Wick and Computing and Social Media, Heraklion, Crete, Greece. https://doi.org/
Keel (2020) found a differential effect of photo editing on body dis- 10.1007/978-3-319-07632-4_11
satisfaction, depending on whether the photo was to be shared. Tiggemann, M., Anderberg, I., & Brown, Z. (2020). Uploading your best
Additionally, as past research highlighting the adverse effect of self: Selfie editing and body dissatisfaction. Body Image, 33, 175–182.
thin-ideal media exposure on women’s body image led to targeted https://doi.org/10.1016/j.bodyim.2020.03.002
interventions, such as media literacy education, future research Tiggemann, M., Hayden, S., Brown, Z., & Veldhuis, J. (2018). The effect of
might explore potential interventions for decreasing photo editing Instagram “likes” on women’s social comparison and body dissatisfaction.
behavior in vulnerable populations. Body Image, 26, 90–97. https://doi.org/10.1016/j.bodyim.2018.07.002
Wick, M. R., & Keel, P. K. (2020). Posting edited photos of the self:
References Increasing eating disorder risk or harmless behavior? International
Journal of Eating Disorders, 53(6), 864–872. https://doi.org/10.1002/eat
Chang, L., Li, P., Loh, R. S. M., & Chua, T. H. H. (2019). A study of .23263
Singapore adolescent girls’ selfie practices, peer appearance comparisons, YouCam Perfect Photo Editor (Version 5.46.2) [Mobile App]. (2019).
and body esteem on Instagram. Body Image, 29, 90–99. https://doi.org/10 Perfect Mobile Corporation. Apple App Store. https://apps.apple.com/us/
.1016/j.bodyim.2019.03.005 app/youcam-perfect-photo-editor-/id768469908
Fox, J., & Vendemia, M. A. (2016). Selective self-presentation and social
comparison through photographs on social networking sites. Received December 15, 2021
Cyberpsychology, Behavior, and Social Networking, 19(10), 593–600. Revision received May 25, 2022
https://doi.org/10.1089/cyber.2016.0248 Accepted September 14, 2022 ▪