Body Image 45 (2023) 401–413

Contents lists available at ScienceDirect

Body Image
journal homepage: www.journals.elsevier.com/body-image

Spanish validation of two social media appearance-related constructs

associated with disordered eating in adolescents: The Appearance- ]]
]]]]]]
]]

related Social Media Consciousness scale (ASMC) and the Critical

Thinking about Media Messages scale (CTMM)
⁎
Marta Rojo ,1, Lucía Beltrán-Garrayo 1, María del Camino del Blanco-Barredo,
Ana Rosa Sepúlveda
Faculty of Psychology, Autonomous University of Madrid, Spain

a r t i cl e i nfo a bstr ac t

Article history: Appearance-related constructs underlying social media are negatively associated with mental health.
Received 17 October 2022 However, their impact on the Spanish population is still unexplored. The present study aimed to validate the
Received in revised form 14 April 2023 Spanish versions of two appearance-related scales: (1) the appearance-related social media consciousness
Accepted 17 April 2023
(ASMC) scale; and (2) the critical thinking about media messages (CTMM) scale. Translation and cultural
Available online 1 May 2023
adaptation of the scales were carried out. The scales’ psychometric properties were assessed using ex­
ploratory and confirmatory factor analyses, measurement invariance across gender (boys vs. girls) and age
Keywords:
Spanish scale validation groups (early adolescents vs. middle adolescents), internal consistency, and convergent validity. The sample
Social media included 803 Spanish secondary school adolescents aged between 12 and 18 (Mage = 15.1, 47.9% girls, 47.2%
Appearance consciousness boys, 4.9% non-binary gender/others). The exploratory factor analyses replicated original one-factor
Critical thinking structures of both scales, which was verified using confirmatory factor analysis. Regarding the ASMC Scale, a
Adolescence re-specified model (allowing for error correlations between Items 1–2) presented an adequate fit. Both
models were invariant across gender and age groups. Excellent internal consistency was found. Bivariate
correlations between the ASMC and eating disorders related variables (body esteem, disordered eating, self-
esteem, sociocultural attitudes towards appearance, and general mental health) supported its convergent
validity and proved ASMC to be a potential target for future preventive eating disorder interventions.
However, the CTMM scale correlated only with sociocultural pressures, thus, further research is needed to
assess the validity of the CTMM in Spanish samples.
© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction The overexposure to physical appearance stimuli on social media

facilitates the development of appearance-focused behaviors, such
As in other Western countries, daily social media use among as constant monitoring of one’s own body (Fardouly et al., 2015;
Spanish adolescents has been dramatically increasing in recent years Karsay et al., 2021) (i.e., self-objectification), which had already been
(Qustodio, 2020; Rideout & Robb, 2018). Specifically, 66.2% of ado­ identified as being typical among the adolescent population by
lescents aged 12–17 have a profile on at least one social media site, Fredrickson and Roberts (1997). Moreover, the positive feedback
rising to 90% over the age of 15 (Ministerio de Interior, 2014). It is expected from peers regarding one’s physical appearance increases
noteworthy that photo-based social media sites (i.e., Instagram, motivation to invest time and effort in caring and worrying about
Tiktok, Snapchat, Facebook, and Twitter) which set the focus on the image transmitted to the social media audience (Ranzini & Hoek,
physical appearance (Deighton-Smith & Bell, 2018; Fox & Vendemia, 2017; Zheng et al., 2019). These appearance-focused experiences
2016), are the most used (Qustodio, 2020; Rideout & Robb, 2018). have been related to several mental health issues (Lin et al., 2016;
Jeri-Yabar et al., 2019; Primack et al., 2017), especially concerning
eating disorder related symptomatology (i.e., body surveillance,
⁎
Correspondence to: Department of Biological and Health Psychology, Faculty of
dietary restriction, appearance comparison and internalization of
Psychology, Autonomous University of Madrid, Ivan Paulov, 6, 28049, Spain. the thin ideal) (Choukas-Bradley et al., 2019; Holland & Tiggemann,
E-mail address: [email protected] (M. Rojo). 2016; Maheux et al., 2022b; McLean et al., 2015; Meier & Gray, 2014;
1
Co-first author.

https://doi.org/10.1016/j.bodyim.2023.04.004
1740-1445/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

Mingoia et al., 2017). These observations are distinctly important messages and photos to reduce its credibility and persuasive influ­
during adolescence, as the intersection between developmental and ence on body dissatisfaction and eating disorders (Halliwell et al.,
socialization processes, and the influence of social media may in­ 2011; McLean et al., 2013, 2016a, 2017). The “Critical Thinking about
crease body image concerns and mental health problems, particu­ Media Messages” scale (McLean et al., 2016a; Scull et al., 2010) ap­
larly among girls (Choukas-Bradley et al., 2022). pears to be the most suitable to assess the frequency of CTMM.
The CTMM scale was first developed as a one dimension-scale by
1.1. Appearance-related Social Media Consciousness (ASMC) Scull et al. (2010) to evaluate to what extent adolescent women from
the Southeastern U.S. claim to think critically about the media
Recent studies have pointed to appearance-related social media messages. The measure presented high levels of internal consistency
consciousness (ASMC) as a risk factor that may underlie the re­ (α = .91), but other psychometric properties were not assessed in the
lationship between social media and mental health issues (Choukas- original study (Scull et al., 2010). The CTMM psychometric properties
Bradley et al., 2020). ASMC refers to the extent to which individuals were later assessed by McLean et al. (2016a) on a sample of ado­
are aware of how attractive they might appear to an online social lescent women from Australia. The confirmatory factor analysis
media audience. Based on this perception, young people might be showed a one-factor structure and the measure obtained good in­
influenced to perform specific appearance-focused social media ternal consistency (α = .90), test-retest reliability (r = .80), and
behaviors such as posing, selecting, and editing photos before construct validity. Moreover, CTMM appeared to be negatively as­
posting them, or checking their body parts and appearance even sociated with social media exposure and eating disorder risk factors,
offline (Chae, 2017; Choukas-Bradley et al., 2020; Fox & Vendemia, such as internalization of the thin ideal and body dissatisfaction
2016; Zheng et al., 2019). A pioneering study conducted by Choukas- (McLean et al., 2016a, 2016b). Particularly, in McLean et al.’s study
Bradley et al. (2019) examined, by means of a four-item survey, (2016b), high levels of critical thinking mitigated the negative im­
ASMC among undergraduate women in the United States (U.S.) pact of high thin internalization and appearance comparisons traits
(Mage = 18.35). The results indicated a high ASMC frequency among on body dissatisfaction. However, the literature presents conflicting
approximately three-quarters of the participants. Moreover, ASMC results regarding its role (McLean et al., 2016c; 2016d). Therefore,
was found to be related to body image concerns, lower self-esteem, while previous studies support the promotion of CTMM as a pro­
and depressive symptoms (Choukas-Bradley et al., 2019). Following, tective factor in the prevention of eating disorders, further research
Choukas-Bradley et al. (2020) developed and validated “The Ap­ is needed. Some of the gaps that need to be addressed include the
pearance-Related Social Media Consciousness Scale” in two high assessment of the psychometric properties of the CTMM scale across
school samples of girls and boys from the Southeastern U.S (sample cultures, gaining insight into its psychometric properties within
1: Mage = 15.72; sample 2: Mage = 16.25). A 13-item, one-factor samples of boys and men, as well as increasing knowledge about the
ASMC scale for assessing this construct emerged. The scale showed association of CTMM with adolescent mental health.
configural, metric and partial scalar gender invariance, strong in­
ternal consistency (sample 1: full sample α = .95, girls α = .94, boys
α = .96; sample 2: full sample α = .92, girls α = .90; boys α = .91), and 1.3. The current study
adequate test-retest reliability after one week (r = .83). Convergent
validity was supported by bivariate correlations between the ASMC Overall, there is growing interest in ASMC and CTMM as two
scale and measures of body vigilance, body shame, self-objectifica­ appearance-related constructs that may underlie the impact of social
tion, and body comparison. Moreover, ASMC was cross-sectionally media on adolescents’ mental health. However, cross-culturally,
associated with adolescents’ disordered eating and depressive scientific evidence on these two constructs is still limited. To our
symptoms, with a significant interaction predicting an effect of knowledge, Spanish-validated versions of the ASMC and CTMM
ASMC on disordered eating only in girls (Choukas-Bradley et al., scales are still not available. Hence, the present study aimed to va­
2020). Subsequently, Maheux et al. (2022b) studied a sample of lidate both scales in a Spanish adolescent population. First, we aimed
high-school adolescents in the Southeastern US over one year and to translate into Spanish and culturally adapt both scales. Second, we
found that the high scores in ASMC was prospectively related to sought to assess their psychometric properties. Specifically, we as­
depressive symptoms after controlling for social media use. It seems sessed the factorial structure of the scales (ASMC and CTMM) via
that this experience is particularly prevalent among adolescents. exploratory and confirmatory factor analyses. Measurement in­
More than 90% of youth presented some degree of ASMC, with variance across gender (boys vs. girls) and age groups (early ado­
higher values among girls (Choukas-Bradley et al., 2020), where it lescence vs. middle adolescence) was also tested to compare mean
might be underlying the impact of social media use in adolescent scores between groups. We also examined internal consistency and
girls’ body image cognitions and concerns (i.e. self-objectification, tested convergent validity by studying the correlations with other
thin-ideal internalization, body shame, body dissatisfaction) eating disorders, body image, and well-being related variables se­
(Choukas-Bradley et al., 2022). Overall, although the evidence is parately between genders (boys vs. girls).
limited as this is a recent construct, there is growing interest in the Based on the previous literature, we expected (1) to replicate the
use of the ASMC scale (Burnell et al., 2021; Maheux et al., 2022a; original one-factor structure in both scales, (2) to support gender
2022b), which has recently been validated among young adults from and age groups invariance of ASMC Scale, while no directional hy­
five English-speaking countries (U.S., U.K., Canada, Australia, New potheses were made regarding CTMM Scale given the lack of pre­
Zealand) (Maheux et al., 2022a), and cross-culturally adapted and vious evidence (3) to find high internal consistency with values
validated among Turkish adolescents, with good psychometric similar to the original validation studies, and (4) to find adequate
properties (Çetinkaya et al., 2022). convergent validity of the measures by gender. Specifically, we ex­
pected higher levels of ASMC to be associated with higher social
1.2. Critical Thinking about Media Messages (CTMM) media exposure, sociocultural pressure towards appearance, and
eating disorder symptomatology, as well as lower body esteem,
Critical thinking about media messages (CTMM), a key element of mental health and general self-esteem. Moreover, we hypothesized
media literacy, can be defined as the ability to critically judge the higher levels of CTMM to be associated with lower scores in social
purpose and influence of media messages (Hobbs & Frost, 2003; media exposure, internalization of the thin-ideal and sociocultural
Silverblatt, 2001). CTMM is thought to promote discussion on the pressure, and, on the other hand, to be positively associated with
impact of unrealistic aesthetic ideals delivered by social media body-esteem. As a final, exploratory hypothesis, higher correlations

402
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

between ASMC and conceptually-related variables were expected for regular school hours in the presence of one or two research team
girls compared to boys. members and a teacher. A week after the survey completion, parti­
cipants attended a workshop on media literacy, as compensation for
2. Method participating in the study. This research was conducted following the
ethical guidelines of the Declaration of Helsinki. Ethics approval was
2.1. Participants given by the Niño Jesús University Children’s Hospital (Ref. 076/18),
and by the Autonomous University of Madrid Ethics Committee (CEI-
The sample comprised 803 adolescents aged between 12 and 18 98–1803).
(Mage = 15.1; SD = 1.40, 47.9% “girls”, 47.2% “boys”, 4.9% non-binary
gender/others) from four public and private urban secondary schools 2.3. Measures
located in the Community of Madrid (Spain). Approximately, 7.4%
(n = 59) of the students were aged 12, 16.9% (n = 135) were 13, 24.8% 2.3.1. Sociodemographic data
(n = 198) were 14, 20.5% (n = 164) were 15 years, 20.4% (n = 163) Participants self-reported information about their date of birth
were 16, 9.1% (n = 73) 17% and 1% (n = 8) were 18. Exclusion criteria (for age calculation), gender identity, school grade and father and
were insufficient knowledge of Spanish, necessary for completing mother education.
the study protocol, and non-availability of parents’ informed consent
or adolescents’ assent. Participants were randomly divided in a 1:1 2.3.2. Social media exposure
ratio into two split-half subsamples. Data was randomly divided into Participants reported their daily time on different social media
a first split-half (n = 401) (Mage =15.01, SD = 1.36, 45.6% girls, 50.4% (Instagram, WhatsApp, TikTok, Facebook, Snapchat, and other social
boys and 4% non-binary gender/others) to perform the exploratory media) using a 5-point Likert scale (1 = Never, 2 = Less than 1 h, 3 =
factor analysis (EFA) and a second split-half (n = 402, Mage = 15.19, SD Between 1 and 2 h, 4 = Between 2 and 5 h, 5 = 5 or more hours). The
=1.42, 50.2% girls, 44% boys, 5.7% non-binary gender/others =0%) maximum value over the daily time spent (average across the days)
subsample for conducting the confirmatory factor analysis (CFA). on social media was calculated. It reflects the largest daily amount of
time spent on social media expressed in hours.
2.2. Procedure
2.3.3. Appearance-related Social Media Consciousness scale
2.2.1. Translation (ASMC, original version by Choukas-Bradley et al., 2020). The
A double forward and backward method was conducted fol­ original ASMC scale consists of 13 items with a single factor struc­
lowing the International Test Commission, 2017. Three independent ture aimed at measuring the frequency of an individual’s thoughts
bilingual translators participated in the process. First, two inter­ and behaviors related to ongoing awareness about their physical
preters whose native language was the target language (Spanish) appearance on social media. Items are measured on a 7-point Likert
translated the original English versions into Spanish. Second, a bi­ scale (1 = Never, 7 = Always). Higher mean scores indicate higher
lingual expert panel specialized in the eating disorder/obesity field, levels of appearance-related social media consciousness. Past ana­
whose native language was Spanish, unified a preliminary Spanish lyzed psychometric properties of this scale are summarized above. In
version. Third, the Spanish versions were back-translated into Eng­ the current study, internal consistency was high (ω = .94).
lish by another independent translator. A bilingual expert-re­
conciliation panel comprising the translators and members of the 2.3.4. Critical Thinking about Media Messages scale
research team, then resolved any discrepancies between the two (CTMM, original version by Scull et al., 2010; McLean et al.,
versions of the instruments and concluded a final Spanish version of 2016a). The CTMM is a 6-item measure with a one-factor structure
the two scales that guaranteed conceptual and semantic equiva­ that examines the degree to which individuals think critically about
lence. Lastly, a pilot study with the translated versions was carried media messages using a 6-point Likert Scale ranging from 1 (Never)
out among ten adolescents within the age and group of interest. to 6 (Always). In our study, participants completed the questionnaire
Specifically, a 5-point Likert scale was used to assess the degree of based on specific instructions “Please answer the following ques­
clarity (1 = not at all clear, 5 = very clear), appropriateness (1 = not at tionnaire about the analyses you made through media messages, noting
all appropriate, 5 = very appropriate), and emotional discomfort (1 = that "media message" refers to those messages through social media
no discomfort, 5 = a lot of discomfort) generated by each item. In (such as Instagram, Tik Tok, among others), TV or other media, which
addition, an open-ended question was used to obtain unstructured provide information about appearance, body image, or specific aesthetic
information. Due to the high percentage of clarity, appropriateness ideals.” that explicitly listed the contexts in which the items should
and low discomfort generated by the items, the final Spanish ver­ have been applied (e.g. the influence of media messages spread from
sions did not need to be modified (see Tables 3 and 4). such as Instagram, TikTok, TV etc. on appearance/body image). The
mean of the scores was calculated, with higher values indicating
2.2.2. Sample recruitment and data collection higher critical thinking about media messages. Past psychometric
The recruitment of participants was conducted through five properties of this scale are summarized above. In the current study,
public and private secondary schools in the urban South-East area of internal consistency was also high (ω = .91).
Madrid, selected using convenience sampling, between April 19th to
June 10th 2021. Prior contact with the headteacher allowed us to 2.3.5. Eating Disorder Examination Questionnaire for Adolescents
establish suitable spots in the students’ schedule for data collection. (EDE-Q-A, original version by Carter et al., 2001, Spanish version
Meetings were held at the centers with the school management by Sepúlveda et al., 2019). This version of the instrument is meant
teams to provide them with all the information about the study. for the adolescent population and includes 36 items aimed at
Parents were informed by the headmasters about the study and measuring the frequency or severity of eating disordered-related
signed the informed consent to participate. At this point, students behaviors and cognitions over the past 14 days. In the present study,
received the link to the online survey (via Qualtrics Software) which only 22 items were used, that evaluate specific eating disorder at­
firstly presents an information page and an assent form. They titudes. Items are rated on a 7-point Likert scale (0 = No days, 6 =
completed the anonymous online survey which lasted around Everyday). While the original instrument counts four subscales
35–40 min using either school computers, or their own electronic (Carter et al., 2001), the Spanish validation confirmed a 2-factor
devices. As previously mentioned, data collection occurred during model: “Restraint” (α = .89), and “Eating, Shape and Weight

403
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

Concerns” (α = .98), unifying the 3-factor original structure. The model of the original version) showed adequate Cronbach’s alpha
average of both subscales constitutes the Global Score of the EDE-Q- values ranging between .77 and .93 and omega coefficient from .79
A (α = .97) (Sepúlveda et al., 2019). In the current study, internal to .93 in a sample of adolescents. In the current sample, the internal
consistencies were excellent (ω = .88 – .96). consistencies of the scale ranged from ω = .88 − .95.

2.3.6. Sociocultural Attitudes Towards Appearance Questionnaire‐4 2.4. Statistical analyses

(SATAQ-4, original version by Schaefer et al., 2015; Spanish ver­
sion by Llorente et al., 2015). This questionnaire includes 22 items 2.4.1. Data cleaning
which assess the perceived sociocultural pressure towards beauty From the 839 initial collected surveys, data from 36 participants
ideals (from family, peers, and media) and the internalization of were excluded in the final analyses. More specifically, 13 participants
beauty standards (thin and muscular). Items are measured on a 7- only completed the demographic information part, and data from 16
point Likert scale (1 = Completely Disagree, 7 = Completely Agree). A participants was removed due to constant response patterns, which
replication of the original five-factor structure was used in the va­ were clearly noticeable despite the inclusion of 12 reverse-scored
lidity analyses of the Spanish-language version of this measure. items in the survey. Moreover, seven cases were excluded due to a
Mean scores of each subscale were computed, with higher scores lack of response to the ASMC or the CTMM scales. Outliers were
indicating a higher degree of endorsement of Western cultural detected using boxplot. No extreme outliers emerged, nor were
standards. The Spanish version showed high reliability (with a these considerably deviated. Outliers detection was also com­
Cronbach’s alpha between .88 and .97). In the current study, the plemented using Z-scores (absolute values above 3.29), although
internal consistencies were similar (ω = .82 − .96). normal distribution was not supported in some variables. The ASMC
and CTMM scores did not reveal any outliers, and those outliers
2.3.7. Body–Esteem Scale for Adolescents and Adults detected in other variables were kept and considered valid as real
(BESAA-S, original version by Mendelson et al., 2001, Spanish cases of population. Multivariate outliers were also assessed using
version by Beltrán-Garrayo et al., 2022). This questionnaire evaluates Mahalanobis Distance Test (Tabachnick and Fidell, 2013). Only 12
an individual’s attitudes and feelings about their own bodies and multivariate outliers were identified (< 1.5% from the total sample).
appearance. The original three-factor structure (BE-Weight, BE-Ap­ Similarly, these were considered real cases from the data with
pearance, and BE-Attribution subscales) was replicated in the normal deviations, as there were not sufficient reasons to consider
Spanish-language measure. The Spanish version reduced the original them invalid (Orr et al., 1991).
scale into 14 items scored on a 5-point Likert scale (0 = Never;
4 = Always). It includes three subscales with adequate internal 2.4.2. Descriptive analyses
consistency in an adolescent population: BE-Appearance (general Means, standard deviations and scales ranges for each measure
feelings about appearance) (α = .88), BE-Weight (weight satisfaction) were calculated for the full sample. Item-level means and standard
(α = .88), and BE-Attribution (evaluations attributed to others toward deviations were also calculated for ASMC and CTMM Scales. Student
one’s body and appearance) (α = .75). The mean of the item scores t-test for independent samples was used to compare the two split-
was calculated for each subscale and the global score was reported halves on quantitative variables and chi-squared tests for categorical
as the mean of the subscale, with higher values indicating greater variables.
body-esteem levels. The internal consistencies in the current study
ranged from ω = .74 − .89. 2.4.3. Factorial structure analyses
An exploratory factor analysis (EFA) (n = 401) and a confirmatory
2.3.8. The Rosenberg Self‐Esteem Scale factor analysis (CFA) (n = 402) were conducted on two randomly split
(RSES, original version by Rosenberg, 1965; Spanish validation by halves of the sample for examining the original factor structure of
Martín-Albo et al., 2007). This scale measures general self-esteem each scale (subsamples were randomly selected using Sample Function
and general perception of self-worth. It is composed of ten items in Rstudio). Swami & Barron’s (2018) guidelines were followed. The
scored on a 4-point Likert scale (1 = Strongly disagree, 4 = Strongly sample size considered for the EFA exceeded the recommended
agree). A total score is calculated summing the item responses participant-to-item ratio of 20:1 (Hogarty et al., 2005), and sample
(ranging from 10 to 40); higher scores indicate higher self-esteem. A adequacy was proved after data analyses with commonalities ≥ .50.
replication of the original one-factor structure was used in the va­ Minimum sample size for CFA was estimated by power analyses
lidity analyses of the Spanish-language version of this measure. The according to the Root Mean Square Error (RMSEA) (power = .95,
original scale has good psychometric properties with a Cronbach’s RMSEA = .05, alpha = .05) (MacCallum et al., 1996). Multivariate
alpha between .77 and .87. The Spanish validation showed adequate normality was rejected after conducting Mardia’s test in each
internal consistencies ranging from .85 to .88 and with a test-retest sample. Univariate normal distribution of each item was considered
coefficient of r = .84. Internal consistency for the current study was with skewness values between [− 3, 3] and kurtosis between [− 10,
similar (ω = .88). 10] (Weston & Gore, 2006). However, the more restrictive criteria by
Curran et al. (1996) was considered for the selection of the estima­
2.3.9. Mental Health Continuum‐Short Form tion method for the EFA and CFA, with skewness values < |2| and
(MHC-ST, original version by Keyes et al., 2008; Spanish valida­ kurtosis < |7|.
tion by Piqueras et al., 2022). This instrument consists of 14 items Barlett’s test of sphericity (Barlett, 1950) and Kaiser-Meyer-
and assesses positive general mental health during the last month. It Olkin’s (KMO) measure (Kaiser, 1974) of sampling adequacy (MSA)
is formed by three different factors: emotional, psychological, and were examined to determine if the data set was suitable for factor
social well-being. Items are measured on a 6-point Likert scale (1 = analysis. A significant test of sphericity indicated appropriateness,
Never, 6 = Everyday). A total score (ranged 14–84) was calculated and (KMO) values are interpreted as above .70 as adequate, be­
from the sum of item responses with higher scores reflecting greater tween .80 – .90 are good and above .90 are excellent. Parallel ana­
levels of mental health (well-being). It has shown good internal lyses (Horn, 1965) were performed to determine the number of
consistency (α > .80), and discriminant reliability across adolescents retain factors. Factors were retained if their eigenvalues of compo­
from different cultures (Keyes, 2009). The Spanish validation (which nents are greater than the eigenvalues of simulated components
presents a bifactor model structure, instead of the three-factor from the random data.

404
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

Table 1
Descriptive statistics of the demographic data for the whole sample and each split-half subsamples.

Total Sample EFA: first split-half (n = 401) CFA: second split-half (n = 402)
(N = 803)

Age M (SD) 15.1 (1.39) 15.01 (1.36) 15.19 (1.42) t (798) = -1.79, p = .073

Gender N(%) 2 (2)= 3.84, p = .146

Boys 379 (47.2) 202 (50.4) 177 (44)
Girls 385 (47.9) 183 (45.6) 202 (50.2)
Non-binary/Others 39 (4.9) 16 (4) 23 (5.7)

School Ggrade N(%) 2 (5)= .75, p = .189

1º ESO (7th Grade) 95 (11.9) 53 (13.3) 42 (10.5)
2º ESO (8th Grade) 203 (25.4) 101 (25.3) 102 (25.5)
3º ESO (9th Grade) 172 (21.5) 95 (23.8) 77 (19.3)
4º ESO (10th Grade) 172 (21.5) 80 (20.1) 92 (23)
1º BACH (11th Grade) 155 (19.4) 70 (17.5) 85 (21.3)
2º BACH (12th Grade) 2 (0.3) 2 (0.5)

Father Education N(%) 2 (4) = 7.17; p = .127

No education/ Unfinished primary education 10 (1.2) 4 (1) 6 (1.5)
Completed primary education (10th Grade) 56 (7) 25 (6.2) 31 (7.7)
Secondary School (12th Grade)/ Vocational training 134 (16.7) 58 (14.5) 76 (18.9)
University Studies 488 (60.8) 262 (65.3) 226 (56.2)
No answer 115 (14.3) 52 (13) 63 (15.7)

Mother Education N(%) 2 (4) = 3.8; p = .434

No education/ Unfinished primary education 9 (1.1) 5 (1.2) 4 (1)
Completed primary education (10th Grade) 45 (5.6) 27 (6.7) 18 (4.5)
Secondary School (12th Grade)/ Vocational training 114 (14.2) 50 (12.5) 64 (15.9)
University Studies 546 (68) 276 (68.8) 270 (67.2)
No answer 89 (11.1) 43 (10.7) 46 (11.4)

Social Media exposure M(SD) 3.55 (.79) 3.58 (.78) 3.53 (.79) t (801) = .83, p = .405
Note. 2: chi-square; t: t-Student test.

The EFA model for the ASMC Scale was computed using gen­ 2.4.4. Measurement invariance
eralized least squares (GLS) estimation (for being item 7 skewness Multi-group CFA was performed to assess measurement in­
and kurtosis slightly deviated), and maximum likelihood (ML) esti­ variance at the configural, metric, and scalar levels across gender,
mation was used in the case of the CTMM Scale. Factor loadings were and across age groups for both scales. Gender invariance was only
considered adequate with a cut-off value of .40 (Stevens, 1992). To assessed between boys and girls given the small size of the non-
verify the retained number of factors in the ASMC Scale, a model binary/Others group. Age was categorized into the following two
retaining one more factor was tested comparing its model fit sta­ groups: (1) 12–14 years, (2) 15–18 years, representing early adoles­
tistics with the parsimonious model. However, it was dismissed. The cence and middle adolescence. Configural invariance is assessed
CFA assessed the underlying factor structure revealed in the EFA. The with a baseline model in which factor loadings and item intercepts
maximum likelihood (ML) estimation was used for both scales given are allowed to freely vary across groups. If the model fits the data, it
the non-multivariate distribution of the data (Satorra & Bentler, implies that the factor structure is the same across groups. The
1994). In the case of the ASMC Scale, the robust ML method was metric invariance model constrains factor loadings to be equal across
employed with the Satorra–Bentler χ2 correction due to an item’s groups. Finally, the scalar invariance model implies both item
(item 7) skewness and kurtosis values slightly deviated. The fit in­ loadings and item intercepts to be similar across groups. Nested
dices analyzed were the Comparative Fit Index (CFI), the Tucker- models (metric compared to configural; scalar compared to metric)
Lewis Index (TLI), the Root Mean Square Error of Approximation were compared using CFI, RMSEA and SRMR changes. According to
(RMSEA), with its 90% confidence interval, and the Standardized Chen (2007), invariance is supported when ΔCFI < .01 supplemented
Root Mean Square Residual (SRMR) (Hu & Bentler, 1999). The ratio of by ΔRMSEA < .015 or ΔSRMR < 0.030 for metric invariance or < .015
S-B 2 / df was also analyzed, considering good values ≤ 3.0 and for scalar invariance. Chi-squared difference test was conducted as
adequate within 3–5. For the CFI and TLI, values close to .95 are well, with a significant result indicating invariance, however Δ 2
considered an optimal fit; in the case of RMSEA values below .06 statistics are highly sensitive to large sample sizes and Chen’s (2007)
indicate a good fit and between .06 to .08 an adequate fit, and for criteria is preferably (Meade et al., 2008; Cheung and Rensvold,
SRMR values ≤ .08 (Hu & Bentler, 1999; Schreiber et al., 2006). 2002). Partial scalar invariance was measured in the case scalar in­
However, the model fit was considered acceptable but moderate variance was not achieved, releasing item intercept constraints
with values above .90 for CFI and TLI and values from .08 to .10 for (Byrne et al.,1989). If measurement invariance was established,
RMSEA (Browne & Cudeck, 1993). In addition, an unbiased SRMR comparison of means scores across groups was tested using Student
index was also analyzed with its cut-off value based on commonality t-test and Cohen’s (1988) d effect size; values below .20 reflect no
following Ximénez et al. (2022) recommendations for preventing effect while values from.21 to .49, .50 to .79, and above .80 reflect
model misspecification across sample size: uSRMR/R2 .05 indicates small, medium, and large effect sizes, respectively.
close-fitting models, an uSRMR/R2 .10 adequate¯ fitting models.
¯
Factor loadings are considered adequate with a cut-off value of .40 2.4.5. Reliability
(Stevens, 1992). Data analyses were conducted using the R Software Internal consistency was calculated through the Omega coeffi­
package. cient (McDonald, 1999) for ASMC and CTMM scales and determined

405
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

Table 2
Items and factor loadings for the Spanish version of the ASMC scale.

ASMC Scale

Factor Loadings

Items EFA CFA

1. When people take pictures of me, I think hoy I will look if the pictures are posted on social media. .74 .67
Cuando la gente me hace fotos, pienso en cómo saldré si las fotos se publican en redes sociales.
2. I think about how specific parts of my body will look when people see my pictures on social media. .79 .79
Pienso en cómo se verán ciertas partes específicas de mi cuerpo cuando la gente vea mis fotos en redes sociales.
3. Even when I’m alone, I imagine how my body would look in a social media picture. .78 .78
Incluso cuando estoy a solas, me imagino cómo se vería mi cuerpo en una foto en redes sociales.
4. During the day, I spend time thinking about how attractive I might look when people see pictures of me on social media. .76 .82
Durante el día. paso tiempo pensando en cómo de atractivo/a puedo parecer cuando la gente vea fotos mías en redes sociales.
5. I try to guess how people on social media will react to my physical appearance in my pictures. .82 .87
Intento adivinar cómo la gente en redes sociales reaccionará a mi apariencia física en mis fotos.
6. My attractiveness in pictures is more important than anything else I do on social media. .74 .78
Mi atractivo en fotos es más importante que cualquier otra cosa que hago en redes sociales.
7. When I do to social events, I care more about looking attractive in pictures people might post on social media than I care about having a fun time. .60 .57
Cuando hago planes sociales, me preocupo más sobre parecer atractivo/a en las fotos que la gente pueda publicar en redes sociales que de pasármelo bien.
8. I fan unattractive picture of me is poste don social media, I feel bad about myself. .75 .72
Si se publica en redes sociales una foto en la que no salgo atractivo/a. me siento mal conmigo mismo/a.
9. I look at pictures of myself on social media again and again. .76 .76
Miro fotos mías en redes sociales una y otra vez.
10. I zoom into social media pictures to see what specific parts of my body look like. .80 .80
Amplio fotos en redes sociales para ver cómo salen partes específicas de mi cuerpo.
11. If someone takes a picture of me that might be poste don social media, I ask to look at it first to make sure I look good. .71 .66
Si alguien me hace una foto que pueda ser publicada en redes sociales, primero pido verla para asegurarme de que salgo bien.
12. Before I post pictures on social media, I crop them or apply filters to make myself look better. .65 .67
Antes de publicar fotos en redes sociales, las recorto o les aplico filtros para salir mejor.
13. If someone takes a picture of me that might be posted on social media, I pose in a particular way so that I’ll look as attractive as possible. .77 .72
Si alguien me hace una foto que pudiese ser publicada en redes sociales, poso de una forma particular de tal manera que parezca lo más atractivo/a
posible.
Average variance extracted (commonality) .55 .55

Note. Standardized item loadings are presented. All item loadings were significant at p < .001. EFA: Exploratory Factor Analysis, CFA: Confirmatory Factor Analysis, ASMC:
Appearance-related Social Media Consciousness.

by gender as well. Likewise, omega coefficient t has been also in­ and for the CTMM Scales (Skewness = 201.83; Kurtosis = 11.45
cluded for the rest of the scales. Values of ≥ .80 were considered p < .001) in the first half of the sample (EFA, N = 401). In the second
adequate for both coefficients (Nunnally, 1978). sub-sample, the relative figures were (CFA, N = 402), for ASMC scale
(Skewness=1890.54, Kurtosis = 32.41, p < .001) and for CTMM Scale
2.4.6. Convergent validity (Skewness = 237.84, Kurtosis = 9.43 p < .001). Thus, multivariate
After checking the assumptions of normality, non-parametric normality was not assumed, and univariate normality of the items
tests were applied. Spearman coefficient was used to assess bivariate was analyzed. Regarding ASMC Scale, skewness ranged between
correlations between ASMC and CTMM scales and ED related vari­ -0.26−2.02 and kurtosis between -1.46−3.88 taking into considera­
ables in the whole sample (N = 803) and separately by gender. tion both split-half subsamples. Only item 7 of the ASMC Scale was
Relevant overall scale scores or subscale scores from the following a little skewed and with slightly deviated kurtosis values but
measures were used for convergent validity analysis: eating disorder within the limit range; consequently, the item’s univariate nor­
symptoms (EDEQ-A), sociocultural attitudes towards appearance mality was considered. The CTMM scale presented adequate
(SATAQ-4), body esteem (BESAA), general self-esteem (RSES), and skewness (between -0.42 and 0.58) and kurtosis (between -1.26
general mental health (MHC-SF). Missing values were handled by and −0.80) values and univariate normality of each item was as­
pairwise deletion. Analyses were computed including and excluding sumed.
outliers and the results remained virtually the same. According to Table 1 shows descriptive statistics for demographic data for the
Cohen’s (1992) recommendations, correlations of .10 were con­ whole sample and the two split-half subsamples. No differences
sidered small, correlations of .30 were considered medium and were found between the two halves in age (t (798) = -1.79, p = .073),
correlations of .50 were considered large. Finally, Fisher’s z-trans­ gender ( 2 (2) = 3.84, p = .146), school grade ( 2 (5) = .74, p = .189),
formation was used to compare the correlations between the stu­ father education ( 2 (4) = 7.17, p = .127), mother education ( 2 (4)
died variables among boys and girls (Myers & Sirois, 2004). Positive = 3.79, p = .434), or social media exposure (t (801) = .83, p = .405). The
values depict a higher correlation between the variables in the boys’ means, standard deviations, and scale ranges of all the measures are
group compared to the correlations in the girls’ group. Conversely, reported in Table 6. Mean levels of ASMC total score for the whole
negative values indicate higher correlations in the group of girls as sample were M = 2.94, SD = 1.49, and for CTMM general score, M
compared to those obtained from the group of boys. The SPSS 24.0 = 3.28, SD = 1.32. Regarding social media daily time use, only one
software package was used to perform these analyses. participant (0.1%) reported never using social media, 6.2% using less
than 1 h per day, 43.72% between 1 and 2 h, 38.1% between 2 and 5 h,
3. Results and 11.8% more than 5 h per day.

3.1. Descriptive statistics 3.2. Factorial structure of the ASMC scale

Item-level and full-scale descriptive statistics for both sub- KMO factor adequacy for the ASMC scale was excellent. The
samples are shown in Supplement 1. The outcomes of the Mardia’s overall KMO = .95 and KMO values for each item ranged from .92
test for the ASMC (Skewness = 1939.30; Kurtosis = 33.88, p < .001) to .97. Bartlett’s test of sphericity was significant, 2 (78) = 3379.36,

406
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

p < .001. The EFA using GLS estimation, after parallel analysis, re­ 3.4. Measurement invariance of the ASMC scale
vealed a one-factor structure. Only one factor was retained with an
eigenvalue greater than the simulated one (λ = 7.60 > 1.31; Invariance testing across gender and age group was carried out
λ = 0.83 < 1.24), which explained 55% of the total variance (mean for the CFA re-specified model. The results for the ASMC Scale from
commonality ≥.50). Model fit for the single factor model was mod­ the multi-group CFA by gender and age groups are reported in
erate, 2 (65) = 326.18, p < .001; TLI = .905; RMSEA [CI 90%] = .10 Table 4. Regarding gender invariance, the configural model showed
[.09 − .11]; SRMR = .05. Item standardized loadings were higher than an adequate fit to the data. Changes in fit indices between metric
.60 (see Table 2). A 2-factor model was tested by EFA analyses (GLS and configural model were within acceptable range, supporting
estimation, rotation oblimin) and there were not any factor loadings metric invariance across gender. Full scalar invariance across gender
higher than .40 in the second factor, thus the model was dismissed. was not supported, as evidenced by a ΔCFI and ΔSRMR over the
The single-factor model was tested by CFA analyses (robust ML es­ threshold, indicating worse model fit. Item intercepts constraints
timation method). The values revealed a poor model fit with some fit were released, and a model with items 9, 4 and 11 intercepts allowed
indices marginally above the acceptable threshold: t, 2 (65) = 281.16, to freely vary across gender showed an acceptable fit and the
p < .001 (S-B 2 /df = 4.32) (Scaling correction factor 1.38); CFI changes in fit indices were within the criteria and the Δ 2 test was
= .916; TLI = .899; RMSEA [CI 90%] = .107 [.94 − .12]; SRMR= .049. The significant, hence partial scalar invariance was supported. Given
null hypothesis for the absolute fit by S-B 2 was rejected (p < .001), partial scalar invariance was supported, gender differences were
due to sample size, and the ratio of S-B 2 /df presented a value < 5. tested in mean scores. Results revealed a significant effect with
Regarding the fit indices, CFI and SRMR presented an adequate fit, higher ASMC levels for girls (M = 3.66, SD = 1.41) compared to boys
the TLI was below and the RMSEA and its confidence intervals at 90% (M = 2.23, SD = 1.21), with large effect size (t (762) = 15.05, p < .001;
were higher than the recommended criteria. However, the unbiased d = 1.09). In the case of measurement invariance tests across age
SRMR was .046, with a uSRMR/R2 = .08, showing an adequate fit for groups, configural, metric and scalar invariance was supported with
the model. Therefore, modification ¯ indexes (M.I.) for the one-factor adequate fit indices within the recommended cut-off. Only Δ 2
solution were explored showing a high correlation between items 1 statistics was not significant in the model comparison between
and 2 (M.I. = 94.41, expected parameter change (E.P.C.) = 1.04). The metric and configural invariance, but the remaining criteria were
model was adjusted adding the error covariances between items 1 appropriate. Similarly, age groups differences were tested, with
and 2, as these two items are theoretically related sharing the same higher values of ASMC in middle adolescents (M = 3.15, SD = 1.50)
content (both are related to the cognitive experience associated to compared to early adolescents (M = 2.71, SD = 1.46), with a small
the ASMC). The re-adjusted model revealed an adequate fit to the effect size (t (798) = 4.25, p < .001; d = .30).
data: 2 (64) = 215.91, p < .001 (S-B 2/ df = 3.37) (Scaling correction
factor 1.33); CFI = .943; TLI = .930; RMSEA [CI 90%] = .89 [.76 − .102];
3.5. Measurement Invariance of the CTMM Scale
SRMR= .046. CFI and TLI fit indexes were close to the values in­
dicative of good fit and were adequate, RMSEA value was slightly
Tests of measurement invariance across gender and age group
high although acceptable and SRMR was excellent. And the unbiased
were conducted for the CFA model from the CTMM Scale, results are
SRMR was .042, with a uSRMR/R2 = 0.076 , indicating an adequate
fit model. The re-adjusted model ¯ significantly improved the model shown in Table 5. All changes in model fit indices for gender in­
variance were acceptable but non-significant chi-square test results
fit, Δ 2 (1) = 24.05, p < .001. Standardized factor loadings are shown
were obtained in the comparison between configural and metric
in Table 2. All factor loadings were statistically significant (p < .001),
invariance. Configural, metric and scalar invariance across gender
and higher than.576 and error variances ranged from .24 to .69.
was considered. No differences were found in CTMM scores between
Covariance between items 1 and 2 was .49. The single factor ex­
boys (M = 3.33, SD = 1.38) and girls (M = 3.22, SD = 1.23) with null
plained 55.31% of the total item variance.
effect size (t (762) = 1.23, p = .218; d = .09). Regarding the invariance
across age groups, apart from Δ 2 statistics not significant in the
configural vs. metric models comparison, all the fit indices changes
3.3. Factorial structure of the CTMM scale
were adequate and below the acceptable thresholds, supporting
configural, metric and scalar invariance across age groups. Likewise,
For the CTMM scale, the overall KMO = .90 and KMO values for
mean scores were compared between age groups differences, re­
each item ranged from.88 to.92, showing indeed good levels.
sulting in no differences in CTMM levels between early adolescents
Bartlet’s test of sphericity was significant: 2 (15) = 1275.72, p < .001.
(M = 3.21, SD = 1.39) and middle adolescents (M = 3.33, SD = 1.24) and
With the CTMM Scale, the EFA was conducted using ML estimation,
null effect size (t (798) = -1.33, p = .184; d = -.09).
and only one factor emerged in the parallel analysis (λ = 3.92 > 1.16;
λ = 0.64 < 1.10), explaining 59% of the total item variance (mean
commonality ≥.50). The model showed a good fit to the data, 2 (9) 3.6. Internal Consistency and Convergent Validity
= 34.98, p < .001; TLI = .966; RMSEA = .085, CI 90% [.06 − .12]; SRMR
= .04. Factor loadings from the items exceeded 0.64 (see Table 3). In The internal consistency reliabilities for ASMC and CTMM scales
the CFA (ML estimation method), the model revealed a good fit to were excellent. For the ASMC scale, the omega coefficient was.94 in
the data: 2 (9) = 25.83, p = .002 (S-B 2 /df = 2.87); CFI = .989; TLI both subsamples and the full sample (boys ω = .93, girls ω = .93). For
= .982; RMSEA = .068, CI 90% [.04 − .10]; pvalue < .50 = 0.14; CTMM Scale, the omega coefficient was .89 and .91 in the first and
SRMR= .02. The null hypothesis for the absolute fit by S-B 2 was second subsamples, respectively; and in the full sample.90 (boys
rejected (p < .001), due to sample size, and the ratio of S-B 2 /df ω = .91, girls ω = .90).
presented an adequate value within 3–5. All fit indexes presented a Regarding convergent validity in the total sample, the total score
good fit, with RMSEA value and confidence intervals at 90% mod­ of the ASMC scale presented significant associations with all the
erate. The unbiased SRMR was .019, with a uSRMR/R2 = .029, in­ variables evaluated, with moderate to strong associations. The ana­
dicating a close-fit model. Standardized factor loadings ¯ are shown in
lysis revealed positive and significant correlations (p < .01) between
Table 3. The model fit supported the one-factor structure of the the ASMC scale and EDEQ-A and SATAQ-4 subscales. ASMC had ne­
original scale, explaining 64.16% of the total item variance. All items’ gative significant correlations (p < .01) with body esteem (BESAA),
standardized loadings exceeded .72 and error variances ranged mental health (MHC-SF), and general self-esteem (RSES) at moderate
from .25 to .48. All model parameters were statistically sig­ levels. There was a small positive significant correlation between
nificant (p < .001). ASMC and BESAA Attribution subscale. Moreover, it was related with

407
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

Table 3
Items and factor loadings for the Spanish version of the CTMM scale.

CTMM Scale

Factor Loadings

Items EFA CFA

1. I think about the purpose behind a message I see on television. .79 .81
Pienso acerca de la intención que hay detrás de un mensaje que veo en la televisión
2. I think about who created the message I see on the ad. .79 .79
Reflexiono sobre quiénes han creado el mensaje que veo en el anuncio.
3. I think about what the people who made the media message want me to believe. .84 .85
Reflexiono sobre qué quieren hacerme creer las personas que hicieron los mensajes de los medios de comunicación.
4. I think about the things the advertisers do to get my attention. .80 .87
Pienso acerca de las cosas que los publicistas hacen para captar mi atención.
5. I think about whether the things that advertisers want me to do are good for me. .73 .77
Reflexiono sobre si las cosas que los publicistas quieren que haga son buenas para mí .
6. I try to think about how true or false and advertisement is. .64 .72
Trato de pensar sobre cómo de verdadero o falso es un anuncio.
Average variance extracted (commonality) .59 .64
Standardized item loadings are presented. All item loadings were significant at p < .001.

Note. EFA: Exploratory Factor Analysis, CFA: Confirmatory Factor Analysis, CTMM: Critical thinking about media messages.

Table 4
Measurement invariance for the ASMC Scale.

Model Fit Indices Model Fit Comparisons

Invariance Model 2 (df) CFI RMSEA (90%CI) SRMR Model Comparisons Δ 2 (Δdf) p ΔCFI ΔRMSEA ΔSRMR
Model Fit Statistics across Gender: boys (n = 379) vs. girls (n = 385)
Configural Invariance 423.81 (128) .924 .093 (.084-.103) .047
Metric Invariance 455.08 (140) .920 .091 (.082-.101) .063 Configural vs. Metric 31.27 (12) < .001 -.004 -.002 .016
Scalar Invariance 585.46 (152) .894 .101 (.093-.110) .079 Metric vs. Scalar 130.38 (12) < .001 -.026 .010 .016
Partial Scalar Invariance 503.54 (149) .912 .093 (.084-.102) .070 Partial Scalar vs. Metric 48.46 (9) < .001 -.008 .002 -.007

Model Fit Statistics across Age Groups: 12–14 years (n = 392) vs. 15–18 years (n = 408)
Configural Invariance 415.88 (128) .94 .089 (.080-.099) .041
Metric Invariance 436.23 (140) .94 .085 (.076-.094) .049 Configural vs. Metric 20.35 (12) .309 .000 -.004 .008
Scalar Invariance 468.03 (152) .937 .083 (.075-.092) .050 Metric vs. Scalar 31.80 (12) < .001 -.003 -.002 .001
Note. 2 : chi-square; df: degrees of freedom; CFI: Comparative Fit Index; RMSEA: Root Mean Square Error of Approximation; CI: confidence interval; SRMR: Standardized Root-
Mean Square Residual; Δ: change in parameter.

moderate significance (p < .01) with social media time spent. All correlations between the scales both for the full sample, and sepa­
these correlations were also revealed in the girls group, except for rated by gender, and the Fisher’s z-comparison.
BESAA Attribution subscale. In the boys’ group, ASMC scale did
correlate with all the variables except for BESAA Weight subscale, 4. Discussion
and MHC-SF Social Wellbeing subscale. Moreover, differences be­
tween boys and girls were found in the magnitudes of the associa­ The current study is, to our knowledge, the first to realize the
tions. Higher associations were observed between ASMC and most of translation and validation into Spanish, and to assess the psycho­
the variables studied for girls compared to boys, except for the MHC- metric properties of the ASMC (Choukas-Bradley et al., 2020), and
SF subscales and the BESAA Attribution subscale. By contrast, the CTMM scales (Scull et al., 2010; McLean et al., 2016a) in an adoles­
CTMM scale did not correlate with any eating disorder-related cent sample (12–18 years old). This study aimed to assess the factor
variables and showed only a small positive significant correlation structures and convergent validity of two social media influence
with SATAQ-4 sociocultural pressure subscale (p < .01 and p < .05). measures on body image in an adolescent sample from Spain. The
When differentiating by gender, those correlations were only assessment of these constructs is relevant as it allows a deeper
maintained in the girls’ group. Table 6 shows all bivariate understanding of the underlying mechanisms of the impact of social

Table 5
Measurement invariance for the CTMM Scale.

Model Fit Indices Model Fit Comparisons

Invariance Model 2 (df) CFI RMSEA (90%CI) SRMR Model Comparisons Δ 2 (Δdf) p ΔCFI ΔRMSEA ΔSRMR
Model Fit Statistics across Gender: boys (n = 379) vs. girls (n = 385)
Configural Invariance 62.83 (18) .983 .081 (.060-.103) .024
Metric Invariance 67.90 (23) .983 .071 (.052-.091) .033 Configural vs. Metric 5.07 (5) .309 .000 -.010 .009
Scalar Invariance 88.38 (28) .977 .075 (.058-.093) .039 Metric vs. Scalar 20.88 (5) p < .001 -.006 .004 .006

Model Fit Statistics across Age Groups: 12–14 years (n = 392) vs. 15–18 years (n = 408)
Configural Invariance 66.99 (18) .982 .082 (.062-.104) .024
Metric Invariance 74.08 (23) .982 .075 (.056-.094) .037 Configural vs. Metric 7.09 (5) .168 .000 -.007 .013
Scalar Invariance 90.18 (28) .978 .075 (.058-.092) .041 Metric vs. Scalar 16.09 (5) p < .001 -.004 .001 .004
Note. 2 : chi-square; df: degrees of freedom; CFI: Comparative Fit Index; RMSEA: Root Mean Square Error of Approximation; CI: confidence interval; SRMR: Standardized Root-
Mean Square Residual; Δ: change in parameter.

408
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

Table 6
Means, standard deviations, internal consistency, bivariate correlations of variables in the whole sample (N = 803) and separately by gender, and comparison of the correlations
across gender.

Whole sample (N = 803) Boys (n = 379) Girls (n = 385) Boys vs. Girls

Variables Interval M (SD) ω ASMC CTMM ASMC CTMM ASMC CTMM ASMC CTMM Fisher’s
rho rho rho rho rho rho Fisher’s z-values
z-values
ASMC 1–7 2.94 (1.49) .94 .04 .06 .008
CTMM 1–6 3.28 (1.32) .90 .04 .06 .01
Social Media exposure 1–5 3.55 (.79) .33 * * .01 .25 * * .07 .40 * * -.05 -2.32 * *
Disordered eating
EDEQ-A: Total 0–6 1.41 (1.45) .96 .54 * * .05 .37 * * .04 .54 * * .09 -2,97 * *
EDEQ-A: R 0–6 1.13 (1.51) .88 .37 * * .03 .27 * * .003 .41 * * .07 -2.19 *
EDEQ-A: EWSC 0–6 1.49 (1.53) .96 .56 * * .05 .36 * * .03 .55 * * .10 -3.32 * *
Sociocultural appearance attitudes
SATAQ:TI 1–5 2.99 (1.05) .92 .53 * * .03 .32 * * .07 .56 * * .02 -4.15 * *
SATAQ: MI 1–5 2.74 (.95) .89 .25 * * .05 .37 * * .09 .38 * * .002 -0.16
SATAQ: FP 1–5 2.23 (.99) .82 .14 * * .11 * * .12 * .06 .13 * .16 * * -0.14 -1.39
SATAQ: PP 1–5 1.8 (.89) .89 .29 * * .08 * .24 * * .08 .33 * * .07 -1.35
SATAQ: MP 1–5 2.31 (1.28) .95 .53 * * .11 * * .25 * * .04 .49 * * .21 * * -3.86 * * -2,38 * *
SATAQ:SMP 1–5 2.42 (1.35) .96 .62 * * .08 * .42 * * .05 .55 * * .18 * * -2.35 * * -1,82 *
Body esteem
BE - Total 0–4 2.13 (.76) .87 -.36** -.02 -.16** .01 -.45 * * -.07 -4.45 * *
BE - Appearance 0–4 2.25 (1.04) .88 -.56 * * -.03 -.39 * * -.02 -.61 * * -.06 -4.09 * *
BE - Weight 0–4 2.44 (1.20) .89 -.21 * * -.01 -.06 .05 -.29 * * -.08 -3.28 * *
BE - Attribution 0–4 1.81 (.80) .74 .08 * .01 .14 * * .04 .02 -.04 1.67 *
General self-esteem
RSES 10–40 28.91 (6.34) .88 -.34 * * -.01 -.18 * * -.02 -.37 * * -.002 -2.84 * *
General mental health
MHC-SF: General 14–84 58.94 (15.15) .95 -.26 * * -.01 -.17 * * -.004 -.24 * * -.02 -1.01
MHC-SF: EW 3–18 13.02 (3.52) .89 -.28 * * .01 -.20 * * .02 -.25 * * -.02 -0.72
MHC-SF: PW 6–36 26.49 (6.87) .89 -.26 * * .02 -.19 * * .01 -.25 * * .01 -0.87
MHC-SF: SW 5–30 19.43 (6.03) .88 -.20 * * -.04 -.10 -.05 -.19 * * -.05 -1,27

Note. * p < .05. * * p < .01. ω: McDonald’s omega coefficient; EDEQ-A: Eating Disorders Examination Questionnaire for Adolescents; R: Restrain; EWSC: Eating/Weight/Shape
Concerns; SATAQ: Sociocultural Attitudes Towards Appearance Questionnaire; MI: Muscular Internalization; TI: Thin Internalization; PP: Peers pressure; FM: Family Pressure; MP:
Media Pressure; SMP: Social Media Pressure; BE: Body-Esteem; RSES: Rosenberg’s Self-esteem Scale; MHC.ST; Mental Health Continuum-Short Form; EW: Emotional wellbeing,
PW: Psychological wellbeing; SW: Social wellbeing.

media on eating disorder symptomatology and emotional well-being than boys with a large effect size, results consistent with Choukas-
during adolescence. Concretely, appearance-related social media Bradley et al. (2020). Our study also provides new and previously
consciousness and critical thinking about media messages have been untested information, supporting groups’ age invariance, allowing
postulated as possible risk (Choukas-Bradley et al., 2020) and pro­ the comparison of means across early and middle adolescence.
tective factors (McLean et al., 2016a, 2016b) for disordered eating Middle adolescents’ ASMC levels were higher than early adoles­
and negative body image, respectively. Their factorial structure, cents’ ones.
measurement invariance across gender and across age groups, in­ Regarding convergent validity, as expected, higher scores in
ternal consistency reliability, and convergent validity were tested. ASMC were positively associated with higher disordered eating
Overall, both scales exhibited adequate psychometric properties si­ symptomatology, higher perceived media pressures, and inter­
milar to the original versions with certain caveats that should be nalization ideals. Medium to large effect size magnitudes for rela­
considered in future research. tions between ASMC scores and measures of disordered eating were
First, regarding the ASMC scale, results of the EFA and CFA sup­ similar to those obtained in Choukas-Bradley et al. (2020), sup­
ported a unidimensional model, which is consistent with the ori­ porting the theoretical link between ASMC and eating disorder pa­
ginal validation (Choukas-Bradley et al., 2020). Nonetheless, model thology risk. As hypothesized, gender differences were found in the
data fit was adjusted, modifying the original one, allowing the error strength of the correlations between ASMC and these variables.
covariances between items 1 and 2 as they were theoretically re­ More specifically, the associations between ASMC and disordered
lated. The re-adjusted fit was not as excellent as the original English eating, thin ideal internalization, media, and social media pressure
version and explained less total item variance (i.e.: total item var­ were stronger in the girls’ subsample compared to the group of boys.
iance of the current study was 55.65% vs. 63.3% of the original one). However, no gender differences emerged in the relationship be­
These differences might be due to cultural disparities between these tween ASMC scale and the muscular ideal, family pressure, and peer
countries. In fact, our results are similar to the data fit found in the pressure subscales (SATAQ-4). These findings highlight the gender
Turkish validation conducted on adolescents (Çetinkaya et al., 2022), differences in the experience of social media ideals internalization
and the validation realized by Maheux et al. (2022a), which included (Rodgers et al., 2020), and in the way of exposure to specific physical
adult participants from different anglophone countries. Our results appearance content on social media (Haferkamp et al., 2012) which
also replicate findings from previous studies which found strong particularly targets adolescent girls (Choukas-Bradley et al., 2022;
internal consistency of the ASMC Scale both when tested on ado­ Rodgers et al., 2020). Despite these gender divergences in the results,
lescents (Çetinkaya et al., 2022; Choukas-Bradley et al., 2020; it must be recognized that boys are also affected by these same
Maheux et al., 2022b), and adults (Burnell et al., 2021; Maheux et al., variables. The differences that emerged can be ascribed to the EDE-
2022a). Measurement invariance across gender revealed configural, Q-A questionnaire, which primarily focuses on shape concerns and
metric and partial scalar invariance, consistent with Choukas- the desire to lose weight as motivated by the thinness ideal. Lit­
Bradley et al. (2020), showing a goodness of fit indices similar to the erature shows that physical standards in the male population mainly
original validation that allows to compare scale scores between boys refer to muscularity ideals (Schaefer et al., 2015), which inter­
and girls (Davidov et al., 2012). ASMC levels were higher for girls nalization appears to be directly related to social media use (Rodgers

409
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

et al., 2020). Recently, an increasing focus on muscularity in girls has previous work that found an association between CTMM and thin-
been observed (Rodgers et al., 2018), and that would explain the no ideal internalization, body esteem, and media exposure (McLean
differences in the correlations between ASMC and muscular ideals et al., 2016a, 2016c). In our study, we found no evidence confirming
across boys and girls. In addition, ASMC has been negatively corre­ CTMM to be related to disordered eating variables. Similarly, McLean
lated with general self-esteem and general mental health, which is et al. (2016a) did not find any significant association with dietary
consistent to prior studies that related ASMC with depressive restraint, and media literacy interventions including CTMM did not
symptoms in adolescents (Maheux et al., 2022b; Choukas-Bradley contribute to improve disordered eating symptoms (McLean et al.,
et al., 2020), adult women (Choukas-Bradley et al., 2020), and with 2016c; 2016d). Due to the role CTMM has in mediating the relation
self-esteem in young women (Choukas-Bradley et al., 2019). No between thin internalization and body dissatisfaction (McLean et al.,
gender differences were found in the association between ASMC and 2013), gender differences were expected in correlation values be­
general mental health, in line with Choukas-Bradley’s et al. (2020) tween these variables. Accordingly, we observed significant corre­
study where higher ASMC results predicted depressive symptoms in lations between CTMM and pressure subscales (SATAQ-4) only in
both boys, and girls. The higher association between ASMC and girls. Finally, our study did not find any significant association be­
general self-esteem observed in girls compared to boys can be ex­ tween general self-esteem and general mental health.
plained by gender differences in the internalization of appearance As far as we are aware, there are no other studies exploring these
ideals and body dissatisfaction, which are negatively related to self- associations. Therefore, these results could be interpreted in line
esteem (Rodgers et al., 2020). with some other contradictory findings in the social media literacy
Besides, ASMC correlated negatively with total body esteem in prevention research. McLean et al. (2016d) published a dismantling
the general sample. Differences were found regarding the correlation study about a media literacy programme in early adolescent girls
with body esteem subscales according to gender. Specifically, ASMC incorporating a critical thinking component for reducing the media
correlated negatively with BE-Weight only for girls and positively influence in body image. Unexpectedly, CTMM values were reduced
with BE-Attribution only for boys. To our knowledge, while previous after the intervention. In a similar way, McLean et al. (2017) did not
studies confirm the negative association between ASMC and general observe any changes in CTMM scores after a social media literacy
body esteem measures (Choukas-Bradley et al., 2019), this is the first intervention in adolescent girls. Similarly, conflicting results have
study to assess the association with different dimensions of this also been found with other media literacy components, such as
construct. A possible explanation for the above-mentioned gender media processing, which was associated with higher levels of dis­
differences is that in social comparison adolescent girls focus par­ ordered eating and body concerns (McLean et al., 2016c). In light of
ticularly on weight-related dimensions, whereas men’s physical at­ these findings, it could be argued that the items from the CTMM
tractiveness is more strongly related to body muscularity than to scale could be interpreted differently as they were designed simi­
BMI (Calogero & Tylka, 2010; Fisher, Dunn, & Thompson, 2002). The larly to the Similarity Scepticism subscale from the Media Attitudes
association of body weight perception and eating disorders-related Questionnaire, which was positive associated with body dis­
unhealthy weight control behaviors, may predict an effect of ASMC satisfaction (McLean et al., 2016a). Thus, greater levels of CTMM
on eating disorders only in girls (Choukas-Bradley et al., 2020). Fi­ could be interpreted as higher consciousness of the sociocultural
nally, the attribution subscale is being questioned as a strict BE di­ pressures from the media, explaining the positive relationship with
mension (Garbett et al., 2021). It is possible that by capturing self- the sociocultural pressure subscales from the SATAQ-4 in this study.
awareness of one’s opinion of appearance (i.e., "People my age like And this would be consistent with McLean et al., (2013; 2016b)
the way I look"), it is positively related to ASMC. which postulated CTMM has a protective factor of the negative ef­
Lastly, greater social media exposure was also related to higher fects of social media images that promotes aesthetic ideal in body
levels of ASMC, as hypothesized; adolescents which are more wor­ dissatisfaction, suggesting a mitigation role on body satisfaction
ried about how they look to the online audience may spend more outcomes in the media literacy prevention research, however, it did
time on social media (Choukas-Bradley et al., 2020). Indeed, it has not show an enhancing effect on eating disorders risk variables
been argued that the negative impact of social media on mental (McLean et al., 2016b). Thus, this effect might not be detected by
health has more to do with ASMC than with the time spent on it bivariate correlations in this study. Furthermore, Wade et al. (2017)
(Holland & Tiggemann, 2016; Meier & Gray, 2014). Further, ASMC suggested that the mechanism of change of media literacy in the
may have a role in the development of mental health issues, as it has development of risk eating disorder variables such as the weight and
been shown to predict disordered eating (Choukas-Bradley et al., body concerns occurs through the media internalization (thin and
2020; 2022) and depressive symptoms (Maheux et al., 2022b) when muscular ideal), and Halliwell, Easun, Harcourt (2011) adds that
controlling time spent on social media, and has been identified as a appearance comparison may be also involved. Given that, as far as
factor that may increase body image concerns (Choukas-Bradley we are aware, no other studies have yet explored these associations,
et al., 2022). In the current study, this association was also stronger we believe that the degree to which the CTMM scale is related to
for girls than boys, which could be explained by the higher levels of these psychological variables is inconsistent and should be the focus
social media related appearance concerns (Choukas-Bradley et al., of future work for its clarification.
2020; 2022; Rodgers et al., 2020).
Concerning the CTMM scale, the EFA and CFA results supported 4.1. Limitations, strengths, and future directions
the same one-factor structure as the original scale (McLean et al.,
2016a), with a total item variance explained (59% and 64.16% vs. 60%) While this study provides incremental contributions to the prior
compared to the original validations (McLean et al., 2016a; Scull literature, it is not free from limitations. First, our study contributes
et al., 2010). No studies on between-age nor between-groups in­ to the understanding of ASMC and CTMM across cultures by vali­
variance had been previously conducted. In our case, configural, dating assessment scales among the adolescent Spanish population;
metric and scalar invariance was demonstrated both for gender, and however, our community sample was recruited incidentally (non-
age groups. However, no significant differences were found neither probability sampling), thus presenting a risk of not being a re­
between boys and girls, nor across different age groups (middle and presentative sample of the population. Another limitation is its
early adolescence). The internal consistency showed good levels, as cross-sectional design, as causality cannot be explored. Additionally,
in previous studies (McLean et al., 2016a; 2016b, 2016d; 2017; Scull while the questionnaires used are validated measures in the target
et al., 2010). However, our results have not gathered evidence for population, as they are self-reported, potential biases such as social
CTMM convergent validity as hypothesized, which is in contrast to desirability may exist. Regarding the psychometric properties, test-

410
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

retest reliability has not been analyzed, being an objective of po­ prevention must be examined, as well as the accuracy of the scale
tential future research. Additionally, a measure of self-objectifica­ items when fulfilling this purpose.
tion/objectified body consciousness was not included as part of the
scale validation process even though this construct is central to 4.2. Conclusions
ASMC. Hence, future research should include the study of the as­
sociation of these two variables in the Spanish population, as ASMC ASMC and CTMM are relevant constructs as emerging risk and
may mediate the impact of social media use in self-objectification, protective factors of eating disorders and mental health. The current
contributing to development of mental health concerns (Choukas- study validated the Spanish-version of the ASMC and CTMM scales
Bradley et al., 2022). Regarding CTTM, although specific instructions among the Spanish adolescent population. The simple construct and
were included for the CTTM scale, this scale could be improved in its appropriated psychometric properties (factorial validity, gender
future research creating more precise items for body-image related and age groups invariance, strong internal consistency, and con­
context. However, another version of this measure for evaluating vergent validity) make the ASMC scale a reliable and useful tool for
specifically physical appearance media messages called “Critical social media-related construct assessment in the Spanish adolescent
Thinking about Media Messages- appearance focus” (CTMM-AF) did population. Furthermore, its use is suggested in further clinical and
not show discriminant validity from CTMM Scale (McLean et al., research settings for understanding the impact of social media on
2016a). Given the poor convergent validity shown in this study, it is adolescents’ mental health. The CTMM Scale exhibited the same
also necessary to note the absence of questionnaires that allow us to single-factor structure as the original scale, gender and age groups
really evaluate whether the CTMM is assessing the correct construct. invariance, and good reliability (internal consistency). However, its
Other alternative theoretically-related constructs (e.g. measures of convergent validity may be questioned given the absence of rela­
plausibility of media messages or credibility of the realism of the tions with the eating disorder associated variables. Therefore, some
media messages) are proposed for future CTMM convergent Spanish preliminary evidence regarding the appropriateness of this scale for
validity assessments in the context of body image research (McLean Spanish adolescents has been found (with certain caveats). It is re­
et al., 2016a). Concretely, Realism Scepticism (McLean et al., 2016a), commended to be used with caution and improved in future re­
another media literacy measure, is proposed to be also validated into search, exploring its convergent construct validity, and revising its
Spanish to assess convergent validity, as it may evaluate media lit­ items’ appropriateness.
eracy more accurately (McLean et al., 2016d, 2017).
Despite the limitations, important strengths of this study should Data Availability
be noted. Firstly, a large sample size, made up of both boys and girls,
was employed. Another strength of the study is that our method Data will be made available on request.
followed best practices for scale validation, including split-half
testing for carrying out AFE and AFC analyses. This two-step analy­ Declarations of Competing Interest
tical strategy has been defended as the most robust approach to test
adaptation and validation (Swami & Barron, 2019; Worthington & None
Whittaker, 2006). In addition, the inclusion of both genders is a
major contribution, especially regarding the CTMM, since previous Acknowledgements
studies had been limited to women samples (Scull et al., 2010;
McLean et al., 2016a, 2016b). A further strength worthy of mention is We are grateful to S. Rodríguez; A. Cataño and E. Mercado for
the study of measurement invariance, which allows to make com­ their contribution in the recruitment and data collection. We also
parisons between boys and girls, and early and middle adolescents. extend our gratitude to the headmasters and teachers from the
Moreover, this study covers a wide range of dimensions among “Real Colegio Nuestra Señora de Loreto”, “Salesianos de Atocha”,
eating disorder risk factors, extending the exploration of these “Ramiro de Maetzu”,”, and “Brains International School’’ high
constructs in eating disorder-prevention strategies. Relatedly, it adds schools of Madrid (Spain) for their collaboration in the study and
to the literature the association between ASMC and the full BESAA assistance during the data collection process. We also would like to
and the internalization of the thin and muscular aesthetic standards. express our appreciation to J. Lajara; M. Norris, and R. Rojo, who
Likewise, it is the first to explore the relationship between CTMM assisted in the forward and backward translation procedure.
and general mental health variables. Finally, we are grateful to the adolescent participants who made
To explore gender-related differences among these social media this study possible.
experiences (Choukas-Bradley et al., 2022), future research should
examine the gendered appearance and aesthetic ideal differences, Appendix A. Supporting information
and how these constructs may differentially affect adolescents with
different gender identities. Most studies have shown a greater social Supplementary data associated with this article can be found in
media impact on appearance concerns in girls than in boys the online version at doi:10.1016/j.bodyim.2023.04.004.
(Choukas-Bradley et al., 2022); however, other gender identities
should be explored or differences across ages. In addition, it must be References
mentioned how studying these aspects in a sample of early ado­
lescents might be relevant in terms of early prevention of the impact Barlett, M. S. (1950). Tests of significance in factor analysis. The British Journal of
Mathematical and Statistical Psychology, 3, 77–85. https://doi.org/10.1111/j.2044-
of social media (Burnette et al., 2017), as this represents a devel­ 8317.1950.tb00285.x
opment stage more vulnerable to its influence (Levine & Smolak, Beltrán-Garrayo, Mercado-Garrido, Román, E., Rojo, M, F. J., & Sepúlveda, A. R. (2022).
2006). Further, these instruments could be examined in a clinical Assessing body esteem in adolescents: Psychometric properties of the spanish
version of the body esteem scale (BESAA-S). Child and Care Forum. https://doi.org/
setting among adolescents with eating disorder risk or diagnosis to 10.1007/s10566-022-09705-w
assess mediational effects of ASMC and CTMM in the impact of social Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A.
media on eating disordered symptoms and negative body image. Bollen, & J. S. Long (Eds.). Testing structural equation models (pp. 136–162). Sage.
Burnell, K., George, M. J., Kurup, A. R., & Underwood, M. K. (2021). “Ur a Freakin
Finally, given the unexpectedly weak associations shown between
Goddess!”: Examining appearance commentary on Instagram. Psychology of Popular
CTMM and other eating disorder risk variables in the present study, Media, 10, 422–433. https://doi.org/10.1037/ppm0000341
further research on its clinical relevance as a protective factor in ED Burnette, C. B., Kwitowski, M. A., & Mazzeo, S. E. (2017). “I don’t need people to tell me
I’m pretty on social media:” A qualitative study of social media and body image in

411
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

early adolescent girls. Body Image. 23, 114–125. https://doi.org/10.1016/j.bodyim. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure
2017.09.001 analysis: Conventional criteria versus new alternatives. Structural Equation
Byrne, B. M., Shavelson, R. J., & Muthén, B. (1989). Testing for the equivalence off actor Modeling, 6, 1–55. https://doi.org/10.1080/10705519909540118
covariance and mean structures: The issue of partial measurement invariance. International Test Commission. (2017). The ITC Guidelines for Translating and Adapting
Psychological Bulletin, 105, 456–466. https://doi.org/10.1037/0033-2909.105.3.456 Tests (Second edition). Retrieved from 〈<underline>www.InTestCom.org</
Calogero, R. M., & Tylka, T. L. (2010). Fiction, fashion, and function: An introduction to underline>〉.
the special issue on gendered body image. Sex Roles, 63, 1–5. https://doi.org/10. Jeri-Yabar, A., Sánchez-Carbonel, A., Tito, K., Ramirez-delCastillo, J., Torres-Alcántara,
1007/s11199-010-9821-3 A., Denegri, D., & Carreazo, Y. (2019). Association between social media use
Carter, J. C., Stewart, D. A., & Fairburn, C. G. (2001). Eating disorder examination (Twitter, Instagram, Facebook) and depressive symptoms: Are Twitter users at
questionnaire: Norms for young adolescent girls. Behaviour Research and Therapy, higher risk? International Journal of Social Psychiatry, 65(1), 14–19. https://doi.org/
39(5), 625–632. https://doi.org/10.1016/S0005-7967(00)00033-4 10.1177/0020764018814270
Çetinkaya, A., Metin, A., Erbiçer, E. S., & Yavuz, E. (2022). Adaptation of appearance- Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika, 39, 31–36. https://
related social media consciousness scale into Turkish: A validity and reliability doi.org/10.1007/BF02291575
study on adolescents. International Journal of Social Sciences and Education Karsay, K., Trekels, J., Eggermont, S., & Vandenbosch, L. (2021). “I (Don’t) Respect My
Research, 8(1), 63–70. https://doi.org/10.24289/ijsser.1034318 Body”: Investigating the role of mass media use and self- objectification on
Chae, J. (2017). Virtual makeover: Selfie-taking and social media use increase selfie- adolescents’ positive body image in a cross-national study. Mass Communication
editing frequency through social comparison. Computers in Human Behavior, 66, and Society, 24(1), 57–84. https://doi.org/10.1080/15205436.2020.1827432
370–376. https://doi.org/10.1016/j.chb.2016.10.007 Keyes, C.L.M. (2009). Brief Description of the Mental Health Continuum Short-Form
Chen, F. F. (2007). Sensitivity of goodness of fit indices to lack of measurement in­ (MHC-SF). Retrieved from 〈<underline>https://peplab.web.unc.edu/wp-content/
variance. Structural Equation Modeling, 14, 464–504. https://doi.org/10.1207/ uploads/sites/18901/2018/11/MHC-SFoverview.pdf</underline>〉.
s15328007sem1203_7 Keyes, C. L. M., Wissing, M., Potgieter, J. P., Temane, M., Kruger, A., & van Rooy, S.
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes fortesting (2008). Evaluation of the Mental Health Continuum-Short Form (MHC-SF) in
measurement invariance. Structural Equation Modeling, 9, 233–255. https://doi. Setswana-speaking South Africans. Clinical Psychology and Psychotherapy, 15(3),
org/10.1207/s15328007sem0902_5 181–192. https://doi.org/10.1002/cpp.572
Choukas-Bradley, S., Nesi, J., Widman, L., & Galle, B. M. (2020). The Appearance- Levine, M. P., & Smolak, L. (2006). The prevention of eating problems and eating
Related Social Media Consciousness Scale: Development and validation with disorders: Theory, research, and practice. Lawrence EarlbaumAssociates.
adolescents. Body Image, 33, 164–174. https://doi.org/10.1016/j.bodyim.2020.02. Lin, L. Y., Sidani, J. E., Shensa, A., Radovic, A., Miller, E., Colditz, J. B., & Primack, B. A.
017 (2016). Association between social media use and depression among US young
Choukas-Bradley, S., Nesi, J., Widman, L., & Higgins, M. K. (2019). Camera-ready: adults. Depression and Anxiety, 33(4), 323–331. https://doi.org/10.1002/da.22466
Young women’s appearance-related social media consciousness. Psychology of Llorente, E., Gleaves, D. H., Warren, C. S., Pérez‐de‐Eulate, L., & Rakhkovskaya, L. (2015).
Popular Media Culture, 8(4), 473–481. https://doi.org/10.1037/ppm0000196 Translation and validation of a Spanish version of the sociocultural attitudes to­
Choukas-Bradley, S., Roberts, S. R., Maheux, A. J., & Nesi, J. (2022). The Perfect Storm: A wards appearance questionnaire‐4 (SATAQ‐4). International Journal of Eating
developmental–sociocultural framework for the role of social media in adolescent Disorders, 48(2), 170–175. https://doi.org/10.1002/eat.22263
girls’ body image concerns and mental health. Clinical Child and Family Psychology MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and de­
Review. https://doi.org/10.1007/s10567-022-00404-5 termination of sample size for covariance structural modeling. Psychological
Cohen, J. (1988). Statistical power analysis for the behavioral science (second ed.). Methods, 1, 130–149. https://doi.org/10.1037/1082-989X.1.2.130
Erlbaum. Maheux, A. J., Roberts, S. R., Nesi, J., Widman, L., & Choukas-Bradley, S. (2022a).
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155–159. https://doi. Psychometric validation and factor structure of the Appearance-Related Social
org/10.1037/0033-2909.112.1.155 Media Consciousness Scale among emerging adults. Body Image, 43, 63–74.
Curran, P. J., West, S. G., & Finch, J. F. (1996). The robustness of test statistics to https://doi.org/10.1016/j.bodyim.2022.08.002
nonnormality and specification error in confirmatory factor analysis. Psychological Maheux, A. J., Roberts, S. R., Nesi, J., Widman, L., & Choukas‐Bradley, S. (2022b).
Methods, 1, 16–19. https://doi.org/10.1037/1082-989X.1.1.16 Longitudinal associations between appearance‐related social media conscious­
Davidov, E., Dülmer, H., Schlüter, E., Schmidt, P., & Meuleman, B. (2012). Using ness and adolescents’ depressive symptoms. Journal of Adolescence, 94(2),
amultilevel structural equation modeling approach to explain cross-cultur­ 264–269. https://doi.org/10.1002/jad.12009
almeasurement noninvariance. Journal of Cross-cultural Psychology, 43, 558–575. Martín-Albo, J., Núñez, J. L., Navarro, J. G., & Grijalvo, F. (2007). The Rosenberg Self-
https://doi.org/10.1177/0022022112438397 Esteem Scale: Translation and validation in university students. The Spanish
Deighton-Smith, N., & Bell, B. T. (2018). Objectifying fitness: A content and thematic Journal of Psychology, 10(2), 458–467. https://doi.org/10.1017/S1138741600006727
analysis of #Fitspiration images on social media. Psychology of Popular Media McDonald, R. (1999). El coeficiente omega. Revista de Ciencias Sociales, 50(4), 150–155.
Culture, 7, 467–483. https://doi.org/10.1037/ppm0000143 McLean, S. A., Paxton, S. J., & Wertheim, E. H. (2013). Mediators of the relationship
Fardouly, J., Diedrichs, P. C., Vartanian, L. R., & Halliwell, E. (2015). The mediating role between media literacy and body dissatisfaction in early adolescent girls:
of appearance comparisons in the relationship between media usage and self- Implications for prevention. Body Image, 10, 282–289. https://doi.org/10.1016/j.
objectification in young women. Psychology of Women Quarterly, 39(4), 447–457. bodyim.2013.01.009
https://doi.org/10.1177/0361684315581841 McLean, S. A., Paxton, S. J., & Wertheim, E. H. (2016a). The measurement of media
Fisher, E., Dunn, M., & Thompson, J. K. (2002). Social comparison and body image: An literacy in eating disorder risk factor research: Psychometric properties of six
investigation of body comparison processes using multidimensional scaling. measures. Journal of Eating Disorders, 4, 30. https://doi.org/10.1186/s40337-016-
Journal of Social and Clinical Psychology, 21(5), 566–579. https://doi.org/10.1521/ 0116-0
jscp.21.5.566.22618 McLean, S. A., Paxton, S. J., & Wertheim, E. H. (2016b). Does media literacy mitigate
Fox, J., & Vendemia, M. A. (2016). Selective self-presentation and social comparison risk for reduced body satisfaction following exposure to thin-ideal media. Journal
through photographs on social networking sites. Cyberpsychology, Behavior, and of Youth and Adolescence, 45(8), 1678–1695. https://doi.org/10.1007/s10964-016-
Social Networking, 19(10), 593–600. https://doi.org/10.1089/cyber.2016.0248 0440-3
Fredrickson, B. L., & Roberts, T. (1997). Objectification Theory: Toward understanding McLean, S. A., Paxton, S. J., & Wertheim, E. H. (2016c). The role of media literacy in
women’s lived experiences and mental health risks. Psychology of Women body dissatisfaction and disordered eating: A systematic review. Body Image, 19,
Quarterly, 21, 173–206. https://doi.org/10.1111/j.1471-6402.2008.00452.x 9–23. https://doi.org/10.1016/j.bodyim.2016.08.002
Garbett, K. M., Lewis-Smith, H., Chaudhry, A., Uglik-Marucha, N., Vitoratou, S., Shroff, McLean, S. A., Paxton, S. J., Wertheim, E. H., & Masters, J. (2015). Photoshopping the
H., & Diedrichs, P. C. (2021). Cultural adaptation and validation of the Body Esteem selfie: Self photo editing and photo investment are associated with body dis­
Scale for Adults and Adolescents for use in English among adolescents in urban satisfaction in adolescent girls. International Journal of Eating Disorders, 48(8),
India. Body Image, 37, 246–254. https://doi.org/10.1016/j.bodyim.2021.02.012 1132–1140. https://doi.org/10.1002/eat.22449
Haferkamp, N., Eimler, S. C., Papadakis, A.-M., & Kruck, J. V. (2012). Men are fromMars, McLean, S. A., Wertheim, E. H., Marques, M. D., & Paxton, S. J. (2016d). Dismantling
women are from Venus? Examining gender differences inself-presentation on prevention: Comparison of outcomes following media literacy and appearance
social networking sites. Cyberpsychology, Behavior and Social Networking, 15(2), comparison modules in a randomized controlled trial. Journal of Health
91–98. https://doi.org/10.1089/cyber.2011.0151 Psychology, 16, 396–403. https://doi.org/10.1177/1359105316678668
Halliwell, E., Easun, A., & Harcourt, D. (2011). Body dissatisfaction: Can a short media McLean, S. A., Wertheim, E. H., Masters, J., & Paxton, S. J. (2017). A pilot evaluation of a
literacy message reduce negative media exposure effects amongst adolescent social media literacy intervention to reduce risk factors for eating disorders.
girls? British Journal of Health Psychology, 16, 396–403. https://doi.org/10.1348/ International Journal of Eating Disorders, 50(7), 847–851. https://doi.org/10.1002/
135910710×515714 eat.22708
Hobbs, R., & Frost, R. (2003). Measuring the acquisition of media‐literacy skills. Meade, A. W., Johnson, E. C., & Braddy, P. W. (2008). Power and sensitivity of alter­
Reading Research Quarterly, 38(3), 330–355. https://doi.org/10.1598/RRQ.38.3.2 native fit indices in tests of measurement invariance. The Journal of
Hogarty, K. Y., Hines, C. V., Kromrey, J. D., Ferron, J. M., & Mumford, K. R. (2005). The AppliedPsychology, 93, 568–592. https://doi.org/10.1037/0021-9010.93.3.568
quality of factor solutions in exploratory factor analysis: The influence of sample Meier, E. P., & Gray, J. (2014). Facebook photo activity associated with body image
size, communality, and over determination. Educational and Psychological disturbance in adolescent girls. Cyberpsycholy, Behavior and Social Networking,
Measurement, 65, 202–226. https://doi.org/10.1177/0013164404267287 17(4), 199–206. https://doi.org/10.1089/cyber.2013.0305
Holland, G., & Tiggemann, M. (2016). A systematic review of the impact of the use of Mendelson, B. K., Mendelson, M. J., & White, D. R. (2001). Body-esteem scale for
social networking sites on body image and disordered eating outcomes. Body adolescents and adults. Journal of Personality Assessment, 76(1), 90–106. https://
Image, 17, 100–110. https://doi.org/10.1016/j.bodyim.2016.02.008 doi.org/10.1207/S15327752JPA7601_6
Horn, J. L. (1965). A rationale and test for the number of factors in a factor analysis. Mingoia, J., Hutchinson, A. D., Wilson, C., & Gleaves, D. H. (2017). The Relationship
Psychometrika, 30, 179–185. https://doi.org/10.1007/BF02289447 between social networking site use and the internalization of a thin ideal in

412
M. Rojo, L. Beltrán-Garrayo, M.d.C. del Blanco-Barredo et al. Body Image 45 (2023) 401–413

females: A meta-analytic review. Frontiers in Psychology, 8, 1351. https://doi.org/ Analysis: Applications for Developmental Research (pp. 399–419). Sage Publications,
10.3389/fpsyg.2017.01351 Inc.
Ministerio de Interior. (2014). Encuesta sobre hábitos de uso y seguridad de Internet de Schaefer, L. M., Burke, N. L., Thompson, J. K., Dedrick, R. F., Heinberg, L. J., Calogero, R.
menores y jóvenes en España. Retrieved from 〈https://www. M., & Swami, V. (2015). Development and validation of the sociocultural attitudes
bienestaryproteccioninfantil.es/imagenes/tablaContenidos03SubSec/Encuesta towards appearance questionnaire-4 (SATAQ-4). Psychological Assessment, 27(1),
%20sobre%20h%C3%A1bitos%20de%20uso%20y%20seguridad%20de%20internet 54. https://doi.org/10.1037/a0037917
%20de%20menores%20y%20j%C3%B3venes%20en%20Espa%C3%B1a.pdf〉. Schreiber, J. B., Nora, A., Stage, F. K., Barlow, E. A., & King, J. (2006). Reporting structural
Myers, L., & Sirois, M. J. (2004). Spearman correlation coefficients, differences between. equation modeling and confirmatory factor analysis results: A review. Journal of
Encyclopedia of Statistical Sciences. https://doi.org/10.1002/9781118445112.stat02802 Educational Research, 99(6), 323–338. https://doi.org/10.3200/JOER.99.6.323-338
Nunnally, J. (1978). Psychometric Theory (second ed.). McGraw-Hill. Scull, T. M., Kupersmidt, J. B., Parket, A. E., Elmore, K. C., & Benson, J. W. (2010).
Orr, J. M., Sackett, P. R., & Dubois, C. L. (1991). Outlier detection and treatment in I/O Adolescents’ media-related cognitions and substance use in the context of par­
psychology: A survey of researcher beliefs and an empirical illustration. Personnel ental and peer influences. Journal of Youth and Adolescence, 39, 981–998. https://
Psychology, 44(3), 473–486. https://doi.org/10.1111/j.1744-6570.1991.tb02401.x doi.org/10.1007/s10964-009-9455-3
Piqueras, J. A., Vidal-Arenas, V., Falcó, R., Moreno-Amador, B., Marzo, J. C., & Keyes, C. L. Sepúlveda, A. R., Compte, E. J., Faya, M., Villaseñor, A., Gutierrez, S., Andrés, P., & Graell,
(2022). Validation of the Mental Health Continuum-Short Form (MHC-SF) for M. (2019). Spanish validation of the Eating Disorder Examination Questionnaire
multidimensional assessment of subjective well-being in Spanish adolescents. for Adolescents (EDE-QA): Confirmatory factor analyses among a clinical sample.
Psicothema, 34(2), 332–343. https://doi.org/10.7334/psicothema2021.240 Eating Disorders, 27(6), 565–576. https://doi.org/10.1080/10640266.2019.1567154
Primack, B. A., Shensa, A., Escobar-Viera, C. G., Barrett, E. L., Sidani, J. E., Colditz, J. B., & Silverblatt, A. (2001). Media literacy: Keys to interpreting media messages. Missouri:
James, A. E. (2017). Use of multiple social media platforms and symptoms of de­ Praeger.
pression and anxiety: A nationally-representative study among US young adults. Stevens, J. P. (1992). Applied multivariate statistics for the social sciences (second ed.).
Computers in Human Behavior, 69, 1–9. https://doi.org/10.1016/j.chb.2016.11.013 Hillsdale, NJ: Lawrence Erlbaum.
Qustodio. (2020). Connected More than Ever. Annual report on children’s digital habits. Swami, V., & Barron, D. (2019). Translation and validation of body image instruments:
Retrieved from 〈https://qweb.cdn.prismic.io/qweb/e59c2e0f-ef4f-4598- Challenges, good practice guidelines, and reporting recommendations for test
b330–10c430e2ec71_Qustodio+2020+Annual+Report+on+Children%27s+Digital adaptation. Body Image, 31, 204–220. https://doi.org/10.1016/j.bodyim.2018.08.
+Habits.pdf〉. 014Y
Ranzini, G., & Hoek, E. (2017). To you who (I think) are listening: Imaginary audience Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics (sixth ed.). Allyn and
and impression management on Facebook. Computers in Human Behavior, 75, Bacon.
228–235. https://doi.org/10.1016/j.chb.2017.04.047 Wade, T. D., Wilksch, S. M., Paxton, S. J., Byrne, S. M., & Austin, S. B. (2017). Do universal
Rideout, V., & Robb, M.B. (2018). Social media, social life: Teens reveal their experiences. media literacy programs have an effect on weight and shape concern by influ­
Retrieved from 〈https://www.commonsensemedia.org/sites/default/files/ encing media internalization? International Journal of Eating Disorders, 50(7),
research/report/2018-social-media-social-life-executive-summary-web.pdf〉. 731–738. https://doi.org/10.1002/eat.22689
Rodgers, R. F., Franko, D. L., Lovering, M. E., Luk, S., Pernal, W., & Matsumoto, A. (2018). Weston, R., & Gore, P. A. (2006). A brief guide to structural equation modeling. The
Development and validation of the Female Muscularity Scale. Sex Roles, 78(1–2), Counseling Psychologist, 34, 719–751. https://doi.org/10.1177/0011000006286345
18–26. https://doi.org/10.1007/s11199-017-0775-6 Worthington, R., & Whittaker, T. (2006). Scale development research: A content
Rodgers, R. F., Slater, A., Gordon, C. S., McLean, S. A., Jarman, H. K., & Paxton, S. J. analysis and recommendations for best practice. The Counseling Psychologist, 34,
(2020). A biopsychosocial model of social media use and body image concerns, 806–838. https://doi.org/10.1177/0011000006288127
disordered eating, and muscle-building behaviors among adolescent girls and Ximénez, C., Maydeu-Olivares, A., Shi, D., & Revuelta, J. (2022). Assessing cutoff values
boys. Journal of Youth and Adolescence, 49, 399–409. https://doi.org/10.1007/ of SEM fit indices: Advantages of the unbiased SRMR index and its cut-off cri­
s10964-019-01190-0 terion based on communality. Structural Equation Modeling: A Multidisciplinary
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton: Princeton Journal, 29(3), 368–380. https://doi.org/10.1080/10705511.2021.1992596
University Press. Zheng, D., Ni, X., & Luo, Y. (2019). Selfie posting on social networking sites and female
Satorra, A., & Bentler, P. M. (1994). Corrections to test statistics and standard errors in adolescents’ self-objectification: The moderating role of imaginary audience
covariance structure analysis. In A. Von Eye, & C. C. Clogg (Eds.). Latent Variables ideation. Sex Roles, 80, 325–331. https://doi.org/10.1007/s11199-018-0937-1

413