Content uploaded by Nick Haslam
Author content
All content in this area was uploaded by Nick Haslam on May 07, 2019
Content may be subject to copyright.
EMOTIONAL IMPACT OF OBJECTIFICATION 1
How does it feel to be treated like an object? Direct and indirect effects of exposure to
sexual objectification on women’s emotions in daily life
Peter Koval1,2, Elise Holland1,3, Michael Zyphur1, Michelle Stratemeyer1, Jennifer Makovec
Knight2,4, Natasha H. Bailen5, Renee J. Thompson5, Tomi-Ann Roberts6 & Nick Haslam1
1 University of Melbourne, Australia
2 Australian Catholic University, Australia
3 Our Watch: End Violence against Women and their Children, Australia
4 Monash University, Australia
5 Washington University in St. Louis, USA
6 Colorado College, USA
In Press, Journal of Personality and Social Psychology
©American Psychological Association, 2019. This paper is not the copy of record and may
not exactly replicate the authoritative document published in the Journal of Personality and
Social Psychology. The final article is available at: http://dx.doi.org/10.1037/pspa0000161
Author Note
This research was supported by a Discovery Project grant from the Australian Research
Council (DP150103053) and funding from the Australian Catholic University.
Correspondence concerning this article should be addressed to Peter Koval,
Melbourne School of Psychological Sciences, The University of Melbourne, Parkville VIC
3010, Australia; email: p.koval@unimelb.edu.au
EMOTIONAL IMPACT OF OBJECTIFICATION 2
Abstract
Exposure to sexual objectification is an everyday experience for many women, yet little is
known about its emotional consequences. Fredrickson and Roberts’s (1997) objectification
theory proposed a within-person process, wherein exposure to sexual objectification causes
women to adopt a third-person perspective on their bodies, labelled self-objectification,
which has harmful downstream consequences for their emotional well-being. However,
previous studies have only tested this model at the between-person level, making them
unreliable sources of inference about the proposed intra-individual psychological
consequences of objectification. Here, we report the results of Bayesian multilevel structural
equation models that simultaneously tested Fredrickson and Roberts’s (1997) predictions
both within and between persons, using data from three ecological momentary assessment
(EMA) studies of women’s (N = 268) experiences of sexual objectification in daily life. Our
findings support the predicted within-person indirect effect of exposure to sexual
objectification on increases in negative and self-conscious emotions via self-objectification.
However, lagged analyses suggest that the emotional consequences of exposure to sexual
objectification may be relatively fleeting. Our findings advance research on sexual
objectification by providing the first comprehensive test of the within-person process
proposed by Fredrickson and Roberts’s (1997) objectification theory.
Keywords: sexual objectification, emotion, daily life, EMA / ESM, multilevel mediation
EMOTIONAL IMPACT OF OBJECTIFICATION 3
How does it feel to be treated like an object? Direct and indirect effects of exposure to
sexual objectification on women’s emotions in daily life.
Most women around the world have experienced sexual objectification first-hand,
often in the form of street harassment, such as being cat-called, ogled, wolf-whistled, or
groped in public (Hollaback!, 2016). Naturalistic studies indicate that the average college-
aged woman in the United States or Australia is targeted by sexually objectifying behaviors
roughly every one to two days (Brinkman & Rickard, 2009; Holland, Koval, Stratemeyer,
Thomson, & Haslam, 2017; Swim, Hyers, Cohen, & Ferguson, 2001). Even if a woman is not
directly targeted by sexual objectification, she is likely to witness sexually objectifying
treatment of other women regularly in her daily life, either in interpersonal encounters or in
the form of sexualized representations of women in the media (Ward, 2016). Despite the
prevalence of sexual objectification, few studies have investigated how real-world
experiences of sexual objectification impact women’s daily psychological well-being. We
address this important question in the current paper by investigating the direct and indirect
emotional impact of exposure to sexual objectification in women’s daily lives using data
from three ecological momentary assessment (EMA) studies.
Objectification Theory
The psychological consequences of sexual objectification were first
comprehensively described in two landmark articles published in the late 1990s (Fredrickson
& Roberts, 1997; McKinley & Hyde, 1996). In their objectification theory, Fredrickson and
Roberts (1997) argued that exposure to sexual objectification is both directly and indirectly
harmful to women’s mental health and well-being. First, women’s exposure to sexually
objectifying events or environments is thought to directly cause distress and increase
women’s vulnerability to eating disorders, sexual dysfunction, anxiety, and depression
(Fredrickson & Roberts, 1997; Szymanski, Moffitt, & Carr, 2011). However, Fredrickson and
Roberts (1997) also proposed that sexual objectification may lead to psychological harm by a
EMOTIONAL IMPACT OF OBJECTIFICATION 4
second “indirect and insidious route” (p. 185), namely by prompting women to adopt a third-
person objectified perspective on their own bodies, a psychic phenomenon termed self-
objectification (see Figure 1). Fredrickson and Roberts (1997) further argued that self-
objectification–involving a preoccupation with one’s appearance and sexual worth–produces
a host of harmful psychological consequences, including unpleasant and self-conscious
emotions, such as anxiety and shame, which accumulate over time to increase women’s risk
of mental illness (see also McKinley & Hyde, 1996). Fredrickson and Roberts’s landmark
(1997) paper has inspired hundreds of empirical studies testing various predictions derived
from objectification theory (for reviews, see Fredrickson et al., 2011; Roberts, Calogero &
Gervais, 2018). Yet, few studies have explicitly tested the entire process model proposed by
Fredrickson and Roberts (1997).
Figure 1. Theoretical model, in which exposure to sexual objectification is hypothesized to indirectly lead to
reduced psychological well-being (e.g., increases in negative and self-conscious emotions) via self-objectification
(see Fredrickson & Roberts; 1997; Roberts et al., 2018).
Previous Research Testing the Sexual Objectification Model
Previous research on sexual objectification has mostly been limited to testing
individual paths from Fredrickson and Roberts’s (1997) model. First, several studies have
found that exposure to sexually objectifying experiences or events leads to increased self-
objectification (e.g., Augustus-Horvath & Tylka, 2009; Karsay & Matthes, 2016). Others
have investigated the direct emotional impact of exposure to objectification (e.g., Prichard &
Exposure to
Sexual
Objectification
Self-
Objectification
Reduced
psychological
well-being
EMOTIONAL IMPACT OF OBJECTIFICATION 5
Tiggemann, 2012; Pritchard, McLachlan, Lavis, & Tiggermann, 2018; Roberts & Gettman,
2004). Finally, a number of studies have demonstrated that self-objectification predicts lower
well-being and greater negative emotions (e.g., Gapinski, Brownell & LaFrance, 2003;
Mercurio & Landry, 2008; Miner-Rubino et al., 2002). Taken together, existing research
provides support for each of the individual paths in Fredrickson and Roberts’s (1997) model
shown in Figure 1. However, studies testing the entire model are rare. Crucially, the only
exceptions have been limited to examining between-person associations, testing whether
individual differences in objectification and self-objectification are linked to individual
differences in well-being. Such studies are not diagnostic of whether Fredrickson and
Roberts’s (1997) model accurately captures how the process of sexual objectification unfolds
within individuals.
In one of the few studies to test the mediation model outlined in objectification
theory (see Figure 1), Miles-McLean et al. (2015) found that experiences of interpersonal
sexual objectification indirectly predicted higher trauma symptoms via increased self-
objectification. However, these findings were based on cross-sectional data and therefore do
not capture the within-person process of interest (Curran & Bauer, 2011; Wang & Maxwell,
2015). Specifically, Miles-McLean et al. (2015) demonstrated that women who report greater
exposure to sexually objectifying events tend to be high in trait self-objectification, which
itself is associated with greater trauma symptoms. However, such findings (see also Hebl,
King, & Lin, 2004; Kozee, Tylka, Augustus-Horvath, & Denchik, 2007) do not demonstrate
that when a woman encounters sexually objectifying behavior this predicts an increase in her
level of state self-objectification, which is subsequently harmful to her well-being. In short,
inter-individual associations assessed at one time-point do not provide evidence for a
dynamic intra-individual process (Fisher, Medaglia, & Jeronimus, 2018). Importantly, such
between-person associations are compatible with various psychological processes, including
EMOTIONAL IMPACT OF OBJECTIFICATION 6
some that conflict with objectification theory (e.g., women higher in trait self-objectification
may be more prone to report incidents of sexual objectification).
Thus, our primary aim was to test Fredrickson and Roberts’s (1997) hypothesized
mediation model at the within-person level using intensive longitudinal EMA data.
Testing a Within-Person Process Model of Sexual Objectification
The psychological process outlined in objectification theory involves a “cascade of
intra-individual psychological consequences” (p. 174) following exposure to sexually
objectifying behaviors or environments (Fredrickson & Roberts, 1997). Specifically, the
theory proposes that when a woman is confronted with sexual objectification she will become
more intensely preoccupied with her appearance and value as a sexual object (i.e., increased
self-objectification), which will, in turn, lead to reductions in her emotional well-being. How
this process unfolds over time for a given individual (i.e., at the within-person level) is
conceptually and statistically distinct from how it manifests in terms of cross-sectional
associations at the between-person level (Hamaker, 2012; Molenaar & Campbell, 2009;
Wang & Maxwell, 2015). To illustrate, consider that people who exercise more have a lower
risk of heart attack than their inactive peers (i.e., exercise and heart attack risk correlate
negatively between persons). However, an individual’s risk of having a heart attack increases
during/after exercising, meaning that exercise and heart attack risk actually correlate
positively within persons, over time (Curran & Bauer, 2011). Thus, cross-sectional surveys
are clearly not appropriate for studying within-person processes, such as the emotional
impact of exposure to sexually objectifying events.
Ecological momentary assessment. One increasingly popular method for capturing
within-person dynamics is EMA, which involves obtaining relatively frequent, momentary
self-reports from participants while they go about their usual daily activities (Bolger &
Laurenceau, 2013; Hamaker & Wichers, 2017). By combining naturalistic and real-time
assessment, EMA also overcomes two major limitations of previous objectification research:
EMOTIONAL IMPACT OF OBJECTIFICATION 7
(i) EMA ensures high ecological validity, which is often lacking in lab experiments and
undermines their generalizability to real-world functioning; and (ii) EMA eliminates or
minimizes the influence of recall biases, which are known to distort self-reports on
retrospective/trait questionnaires (Trull & Ebner-Priemer, 2014). Thus, EMA is an important
and thus far under-utilized methodology for studying social psychological processes, such as
sexual objectification, in daily life.
A handful of studies have investigated women’s experiences of sexual
objectification in daily life using EMA or diary methods (Breines, Crocker & Garcia, 2008;
Brinkman & Rickard, 2009; Holland et al., 2017; Swim, Hyers, Cohen & Ferguson, 2001).
These studies have separately tested individual paths from Fredrickson and Roberts’s (1997)
model, such as the impact of exposure to sexually objectifying events on state self-
objectification (Holland et al., 2017) or emotions (Swim et al., 2001) in daily life. However,
no previous EMA study has tested the within-person process model in its entirety.
Furthermore, not all previous EMA studies of sexual objectification have measured the
occurrence of other everyday stressors, which are known to impact emotional well-being
(Almeida, 2005). Given that daily hassles may co-occur with sexually objectifying events, it
is crucial to model the effects of both types of events to reveal the unique impact of sexually
objectifying experiences on women’s emotions.
Bayesian multilevel structural equation modeling. Although methods for testing
multilevel mediation have existed for some time (Preacher, Zyphur, & Zhang, 2010), until
recently they could not optimally handle within-person predictors with missing data
(Asparouhov & Muthèn, 2018). Given that missing data are extremely common in EMA
designs, this has represented a major challenge for obtaining unbiased estimates of within-
person indirect effects using EMA data. Here, we took advantage of recent advances in
Bayesian multilevel SEM as implemented in Mplus version 8.2 (Asparouhov & Muthèn,
EMOTIONAL IMPACT OF OBJECTIFICATION 8
2018), to test Fredrickson and Roberts’s (1997) proposed mediation model at both the within-
and between-person levels using intensive longitudinal EMA data.
The Current Study
In the current study, we report analyses of EMA data from three samples of young
women who reported on their exposure to sexually objectifying events, their state levels of
self-objectification, and their momentary experiences of positive, negative and self-conscious
emotions in daily life over several days. Drawing on Fredrickson and Roberts’s (1997)
objectification theory, we hypothesized that exposure to sexually objectifying events would
indirectly predict within-person increases in negative and self-conscious emotions over time,
via their within-person impact on heightened self-objectification. Furthermore, given that
some research has suggested that women may experience positive feelings in response to
objectifying encounters (e.g., Gervais et al., 2011; Liss, Erchull, & Ramsey, 2011), we also
examined the within-person effects of exposure to sexual objectification on women’s positive
emotions. However, given that objectification theory does not predict such positive emotional
consequences to be mediated by increases in self-objectification, we did not expect to find a
within-person indirect effect of exposure to objectifying events on positive emotions via self-
objectification. Finally, we simultaneously tested an equivalent model at the between-person
level, allowing us to determine whether previously reported between-person associations
could be replicated.
Method
Participants
We recruited three samples of women to complete an EMA study on their daily
experiences of sexual objectification. Sample 1 (collected in 2015)
1
comprises 82 women
recruited via advertisements posted online and around university campuses in Melbourne
1
We analyzed Sample 1 data for a previous manuscript (citation obscured for blind review) reporting on the
prevalence of sexual objectification in daily life and its impact on self-objectification. The current analyses are
distinct and have not been previously reported.
EMOTIONAL IMPACT OF OBJECTIFICATION 9
(Australia). One participant withdrew, leaving n = 81. Sample 2 (collected in 2016)
comprises 90 women also recruited using similar methods in Melbourne. After excluding
data from one participant due to low EMA compliance (she completed < 35% of scheduled
EMA surveys) and from two others who participated in the study the previous year and were
thus included in Sample 1, we were left with n = 87 for Sample 2. Finally, Sample 3
(collected in 2016-2017) comprises 100 women recruited in the greater St. Louis, Missouri
area (USA) using similar recruitment methods as for the other two samples. No participants
were excluded from Sample 3.
Taken together, Samples 1-3 comprise 268 women aged 18 to 46 years (M = 24.26,
SD = 5.61) who reported their ethnicity as White/Caucasian (49%), Asian (27%), South
Asian (7%), Black/African (4%), mixed (7%), or other (6%). Approximately 35% of all
participants were born in the USA, 21% in Australia, and the remaining 44% in other
countries. Most participants identified as heterosexual (82%), with the remainder identifying
as bisexual (11%), homosexual
2
(4%) or “other” (3%). Just over half (53%) of all participants
were single, 35% were in unmarried relationships, 10% were married, and 2% listed their
relationship status as “other”. Thus, our samples were relatively diverse in their demographic
composition. Detailed demographic information for each sample is provided in the
supplemental materials (see Table S1).
Materials and Procedure
Materials and procedure used in Sample 1 differed very slightly from those used in
Samples 2 and 3, which were identical. All such methodological differences are noted below.
Before commencing the main EMA component of the study, participants attended an
initial lab session in small groups (2-10 participants at a time) to receive detailed instructions
2
Although the term “homosexual” may be transitioning out of scientific usage, we use this term for consistency
with our measure of sexual orientation: participants responded to the question “what is your sexual orientation?”
by selecting “heterosexual”, “homosexual”, “bisexual” or “other”. However, researchers may consider using
alternative term(s) to measure self-reported sexual orientation in future research.
EMOTIONAL IMPACT OF OBJECTIFICATION 10
for the EMA and to complete questionnaires assessing demographics and other background
variables (not reported here). For several days after leaving the lab, participants reported on
their exposure to sexually objectifying events, their state levels of self-objectification, and
their momentary experiences of positive and negative feelings in daily life using a custom-
built EMA smartphone app called SEMA2 (Harrison, Harrison, Koval, Gleeson, & Alvarez-
Jimenez, 2017).
Momentary emotions. Participants rated their current levels of several emotions
using items in the form of “Right now, how ____ do you feel?”, with four items (“angry”,
“sad”, “anxious”, “guilty”) combined into a measure of negative emotion, two items
(“happy”, “confident”) combined to form a measure of positive emotion, and a single item
(“self-conscious”; Sample 1) or two items (“ashamed”, “embarrassed”; Samples 2 & 3)
combined to form a measure of self-conscious emotion. All emotion items were rated on
slider scales from 0 (not at all) to 100 (very much) and were presented in a random order at
the beginning of each EMA survey. We assessed emotions first to avoid any influence of
recalling objectifying events and levels of self-objectification on momentary emotions.
State self-objectification. Next, participants rated their level of state self-
objectification “since the last survey” on a scale from 0 (not at all) to 100 (very much). In
Sample 1, state self-objectification was measured with a single item (“have you been thinking
about how you look to other people?”), whereas in Samples 2 and 3, two additional items
(“have you felt self-conscious about your appearance?” and “have you been worried about
whether your clothes make you look good?”) were added to create a three-item measure of
state self-objectification. These items were adapted from the self-surveillance subscale of
McKinley and Hyde’s (1996) Objectified Body Consciousness Scale, a widely used measure
of trait self-objectification. Importantly, state self-objectification was measured before
participants reported on their exposure to sexually objectifying events to avoid event-recall
influencing ratings of self-objectification.
EMOTIONAL IMPACT OF OBJECTIFICATION 11
Exposure to sexually objectifying events. Participants reported whether they had
been targeted by one or more sexually objectifying behaviors “since the last survey”, with the
following response options adapted from Kozee et al.’s (2007) Interpersonal Sexual
Objectification Scale: (i) catcalling, wolf-whistling, or car honking; (ii) sexual remark made
about body; (iii) touched/fondled against will; (iv) body looked at sexually; (v) degrading
sexual gesture; (vi) other objectifying behavior not listed above; or (vii) none of the above.
Participants also reported whether they had witnessed one or more of the above forms of
sexual objectifying behavior directed at other women. The witnessing item included an
additional response option (media image/video) to capture exposure to sexualized depictions
of women in the media. Following Holland et al. (2017), we constructed binary target and
witness variables, for which a value of 1 indicated the occurrence of one or more sexually
objectifying events and a value of 0 indicated no objectifying event.
Exposure to other stressors/hassles. In Samples 2 and 3, the last item in the EMA
survey asked participants to report “other stressors/hassles, since the last survey” (1=yes;
0=no). As described below, responses to this item were used to control for the emotional
impact of stressors when estimating effects of exposure to objectifying events.
EMA protocol
During the initial lab session, participants downloaded the EMA app (SEMA2) onto
their personal Android or iOS smartphone. A researcher provided detailed instructions for
completing the EMA surveys and participants were given an opportunity to ask clarification
questions while completing a demo survey before leaving the lab.
Participants in Sample 1 were prompted to complete EMA surveys every 84 ± 30
minutes between 10 a.m. and midnight (i.e., approximately 10 EMA surveys daily) for seven
consecutive days. Participants in Samples 2 and 3 were prompted to complete EMA surveys
every 60 ± 30 minutes (i.e., approximately 14 EMA surveys daily) for five consecutive days.
Participants in all samples were therefore prompted to complete approximately 70 EMA
EMOTIONAL IMPACT OF OBJECTIFICATION 12
surveys over the duration of the EMA study. To prevent back-filling, EMA surveys expired
after 15 min and any incomplete items were marked as missing. An entire EMA survey was
considered missing only if no responses were recorded.
Following standard practice in EMA studies, reimbursement was partially
contingent upon completion of EMA surveys. Participants in Samples 1 and 3 received
between $30 and $50 cash (contingent upon EMA compliance). In Sample 2, all participants
received a $50 gift-card and those who completed at least 50% of scheduled EMA surveys
were entered into a raffle to win one of 10 additional $50 gift-cards, with the number of raffle
entries allocated to each participant dependent on their EMA compliance.
Overall, participants completed an average of 81.4% of scheduled EMA surveys
(Range = 38-–100%, SD = 13.2%), reflecting very good compliance. Mean EMA compliance
rates were 83.7%, 81.6%, and 79.4% for Samples 1, 2, and 3, respectively.
Data analytic strategy
We sought to maximize the statistical power and reliability of our analyses in two
ways. First, given that power is heavily influenced by the number of upper-level units (i.e.,
participants) in multilevel designs (Bolger & Laurenceau, 2013), we conducted our main
analyses using the combined data from all three samples (N = 268). This approach, referred to
as “mega-analysis” (e.g., Fleeson & Gallagher, 2009), is often preferred over traditional
meta-analysis when all raw data are available (Steinberg et al., 1997). We conducted
additional analyses to test for possible differences between the three samples, which we
report in the supplemental materials (see Tables S4-S7).
3
Second, after collecting data for
Sample 1, in which we assessed two of our central constructs (i.e., self-conscious emotion
and self-objectification) with single items, we decided to use multi-item scales to assess these
constructs in Samples 2 and 3 to increase reliability and therefore maximize statistical power.
3
Additional analyses revealed that while some parameter estimates differed between samples, the hypothesized
within-person indirect effects of objectifying events on negative and self-conscious emotions (via self-
objectification) were consistently positive (see pp. 3-8 in supplemental for further details).
EMOTIONAL IMPACT OF OBJECTIFICATION 13
For multi-item scales (negative emotion and positive emotion in all samples; self-
conscious emotion and self-objectification in Samples 2 and 3), we calculated mean scores by
averaging responses across the relevant items at each EMA survey. Multilevel reliability
coefficients for these scales are reported in Table 1. The data from all three samples were
‘stacked’ together for our main analyses and thus slightly different operationalizations of
self-conscious emotion and self-objectification were treated as equivalent for our main
analyses. However, separate analyses for each sample are also reported in the Supplemental
Materials.
To account for the hierarchical data structure (EMA surveys nested within
participants) we analyzed data using multilevel SEM in Mplus version 8.2 (Muthén &
Muthén, 1998-2017). Specifically, we followed the general multilevel mediation approach
described by Preacher et al. (2010) to estimate a series of multilevel mediation models,
including both direct and indirect (via self-objectification) effects of exposure to sexually
objectifying events (as target or witness) on momentary negative, self-conscious and positive
emotions. At the within-person level, the EMA data have a longitudinal structure, which we
accounted for by estimating lagged associations between variables, as described below and
illustrated in Figure 2. We ran separate models for each emotion (negative, self-conscious,
positive) and each type of exposure (target, witness), resulting in a total of six models.
In line with Fredrickson and Roberts’s (1997) theoretical predictions, our main focus
was on modeling direct and indirect effects at the within-person level, where the emotional
consequences of sexual objectification are predicted to unfold. Thus, it was essential to
decompose our observed variables (comprising within- and between-person variance) into
separate within- and between-person components (Kenny, Korchmaros, & Bolger, 2003;
Preacher et al., 2010). This can be done in Mplus by modeling the within- and between-
person components of an observed variable as latent variables, a technique known as latent
centering, as shown in the left panel of Figure 2 (Asparouhov & Muthèn, 2018).
EMOTIONAL IMPACT OF OBJECTIFICATION 14
Figure 2. Multilevel mediation model tested. The left panel shows the decomposition of observed variables (rectangles) into their latent within-person (grey ellipses) and
between-person (black ellipses) components. The right panels show the model estimated at the within- and between-person levels. Note that within-person paths are labeled
with the subscript i, indicating that they were estimated as random slopes and thus vary across individuals i. AR(1) paths were estimated within-persons for self-objectification
(S-Obj), and Emotion, to control for temporal carry-over and model change in these variables. All random effects were allowed to covary between-persons.
Observed
Between-Person
Within-Person
Within-Person
Between-Person
Eventti S-Objti Emotionti
EventiS-ObjiEmotioni
S-Objt
EventtEmotiont
Time
EMOTIONAL IMPACT OF OBJECTIFICATION 15
Within-person model. As shown in the bottom-right panel of Figure 2, we estimated the
within-person effects of exposure to objectification on self-objectification and emotions using
latent-centered within-person variables. Specifically, emotional intensity at time t2
(Emotiont2) was regressed onto objectifying events (Eventt1-t2) and state self-objectification
(S-Objt1-t2) reported as occurring in the interval between t1 and t2 (denoted with the subscript
t1-t2).
4
These paths represent the within-person lagged effect of exposure to sexual
objectifying events on emotions (path c’Wi) and the within-person lagged effect of state self-
objectification on emotions (path bWi), respectively. Self-objectification was also regressed
onto events to estimate the within-person effect of exposure to sexually objectifying events
on self-objectification (path aWi). To ensure that we were modeling change over time in the
two outcome variables, we simultaneously estimated autoregressive slopes for emotions and
self-objectification (EmotionAR(1)i and S-ObjAR(1)i) by including lagged versions of these
variables as predictors.
5
Within-person paths were estimated as random slopes that were
allowed to vary across individuals, as indicated by the subscript i for all within-person model
parameters shown in Figure 2. Following Preacher et al. (2010; see also Bolger &
Laurenceau, 2013), within-person indirect effects were calculated as the product of the
average within-person aWi and bWi paths, plus the covariance of the aWi and bWi paths (i.e.,
indirectW = 𝑎𝑊
̅
̅
̅
̅
× 𝑏𝑊
̅
̅
̅
̅
+ 𝑐𝑜𝑣(𝑎𝑊𝑖 , 𝑏𝑊𝑖)). Within-person total effects were calculated by
summing the direct effect (path c’W) and indirect effect (totalW = c’W + indirectW; see Bolger
and Laurenceau, 2013).
Between-person model. We estimated similar models using the latent between-
person components of each variable (see top-right panel of Figure 2). At the between-person
4
While all variables in the model were measured at the same occasion (time t2), exposure to objectifying events
(Eventt1-t2) and state self-objectification (S-Objt1-t2) were reported as occurring “since the last survey” and can
therefore be assumed to temporally precede momentary emotions, reported as “right now”.
5
Because the lagged predictors (Emotion
t1
and S-Obj
t0-t1
) were only included in the within-person model, they
were centered using observed mean-centering rather than latent centering.
EMOTIONAL IMPACT OF OBJECTIFICATION 16
level, path aB represents the effect of mean exposure to sexually objectifying events (Eventi)
on mean levels of self-objectification (S-Obji). Path bB represents the association between
mean levels of self-objectification (S-Obji) and mean levels of emotions (Emotioni). Finally,
path c’B captures the direct relationship between mean exposure to sexually objectifying
events (Eventi) and mean levels of emotion (Emotioni). As in regular single-level mediation,
the between-person indirect effect was calculated as the product of the aB and bB paths
(indirectB = 𝑎𝐵× 𝑏𝐵) and the between-person total effect was calculated by summing the direct
(path c’B) and indirect effects (totalB = c’B + indirectB).
Bayesian estimation. We used Bayesian estimation for all multilevel models
reported, which overcomes two major limitations of the more commonly used maximum
likelihood estimation. First, because our models were complex (i.e., they included multiple
random effects, categorical variables, and missing data) maximum likelihood estimation
requires the use of numerical integration, an extremely inefficient approach that produces
imprecise estimates and often results in convergence problems (Asparouhov & Muthèn,
2018). Second, Bayesian estimation does not assume that model parameters are normally
distributed. This is a particularly problematic assumption for indirect effects, which are
known to be skewed (Shrout & Bolger, 2002; Yuan & MacKinnon, 2009). Instead, Bayesian
posterior distributions of model parameters (including indirect effects) can take any form.
Importantly, we used the Mplus default uninformative priors (see Asparouhov & Muthèn
2018), resulting in parameter estimates that are driven primarily by the data and are therefore
similar to estimates obtained using maximum likelihood estimation (Yuan & MacKinnon,
2009; Zyphur & Oswald, 2015). To ensure stable parameter estimates, all models used a
minimum of 20,000 Bayesian iterations, checking to ensure models had converged by the
10,000th iteration using posterior scale reduction (PSR) values below 1.05—the first half of
iterations are dropped as a ‘burn in’ period. For each model parameter’s posterior
EMOTIONAL IMPACT OF OBJECTIFICATION 17
distribution, we report the median (point estimate), standard deviation (akin to a standard
error), and the 2.5 and 97.5 percentile values (forming 95% credibility intervals; CIs). We
consider model parameters to be meaningfully different from zero when their 95% CIs do not
cross zero.
Open practices
The data and all Mplus input and output files required to reproduce the analyses
reported in the paper and supplemental materials are available in a public repository on the
Open Science Framework (OSF), available at https://osf.io/tz9wn/.
Results
Descriptive statistics and reliability estimates for all continuous measures are shown
in Table 1 (see Table S2 in supplemental materials for separate values for each sample). We
estimated within- and between-person alpha reliability coefficients following Geldhof,
Preacher and Zyphur (2014).
6
As shown in Table 1, all scales showed adequate reliability (rs
≥ .48, ωs ≥ .62) within- and between-persons. As shown in Table 1, each measure had
substantial variance (≥ 43%) at both the within- and between-person levels. Thus, the data
were appropriate for multilevel analyses.
6
Given the limitations of alpha as an index of reliability (see e.g., Sijtsma, 2009), we also ran multilevel CFAs
to obtain estimates of omega reliability within (ωW) and between (ωB) persons (see Bolger & Laurenceau, 2013;
Geldhof et al., 2014). Multilevel omegas for Negative Emotion (ωW = 0.62, 95%CI [0.61, 0.63], ωB = 0.89, 95%
CI [0.87, 0.91]) and Self-Objectification (ωW = 0.89, 95%CI [0.89, 0.89], ωB = 0.99, 95% CI [0.99, 0.99]) were
very similar to the alpha values reported above. We estimated multilevel correlations for the two-item measures
of Positive Emotion (rW = 0.48, 95%CI [0.47, 0.49], rB = 0.83, 95% CI [0.79, 0.86]) and Self-Conscious
Emotion (rW = 0.51, 95%CI [0.50, 0.52], rB = 0.96, 95% CI [0.95, 0.97]) as multilevel CFAs would have been
under-identified.
EMOTIONAL IMPACT OF OBJECTIFICATION 18
Table 1
Descriptive statistics and reliability estimates for continuous measures
Sample size
SD (% variance)
Cronbach’s alpha
Measure
T
N
Mean
Within
Between
Within
Between
Negative Emotion
15657
268
16.09
12.27 (45%)
13.54 (55%)
0.61
0.88
[14.40, 17.70]
[12.14, 12.41]
[12.41, 14.79]
[0.60, 0.62]
[0.86, 0.90]
Positive Emotion
15634
268
59.37
17.85 (51%)
17.36 (49%)
0.65
0.90
[57.23, 61.47]
[17.65, 18.05]
[15.96, 19.06]
[0.64, 0.66]
[0.87, 0.92]
Self-Conscious Emotion
15584
268
14.36
15.14 (43%)
17.57 (57%)
0.68a
0.98a
[12.29, 16.58]
[14.97, 15.31]
[16.16, 19.24]
[0.66, 0.69]
[0.97, 0.99]
Self-Objectification
15531
268
26.01
22.04 (50%)
21.87 (50%)
0.89a
0.99a
[23.45, 28.79]
[21.79, 22.28]
[20.01, 23.84]
[0.88, 0.89]
[0.99, 0.99]
Note. T = number of occasions; N = number of participants; values in square brackets are 95% Bayesian credible intervals; all measures were assessed
on scales from 0 (not at all) to 100 (very much).
a Alphas for self-conscious emotion and self-objectification are based on Sample 2 and 3 data only because single-item measures were used in Sample 1.
Frequency data for objectifying events are shown in Table 2 (see Table S3 in the
supplemental materials for frequencies for each sample). Overall, 66% of participants
reported being targeted by sexual objectification at least once, and 85% reported witnessing
at least one sexually objectifying event during the course of the study. Given the relatively
low prevalence of sexually objectifying events (from a statistical perspective), we sought to
maximize power and obtain more reliable parameter estimates by conducting analyses using
data from all three samples.
Table 2
Descriptive statistics for exposure to sexually objectifying events reported as target and witness
Frequency of events
Proportion of EMA surveys with events
N (%) of participants
reporting 1 or more events
Range
Median
Mean (SD)
Range
Median
Mean (SD)
Target
177 (66%)
0–27
1.00
2.75 (4.28)
0.00–0.38
0.02
0.05 (0.07)
Witness
227 (85%)
0–58
4.00
7.80 (11.04)
0.00–1.00
0.07
0.13 (0.19)
Note. N (%) > 0 = number (percentage) of participants who reported being targeted by or witnessing sexually objectifying events at least once.
Frequency = number of EMA surveys on which participants reported being targeted by or witnessing sexually objectifying events.
Proportion = proportion of EMA surveys on which participants reported being targeted by or witnessing sexually objectifying events.
Multilevel mediation analyses
Estimates of direct, indirect (via self-objectification) and total effects of exposure to
objectifying events on negative, self-conscious, and positive emotions, are shown in Table 3.
EMOTIONAL IMPACT OF OBJECTIFICATION 19
Within-person effects. Estimates of aW paths in Table 3 indicate that exposure to
sexually objectifying events predicted reliable increases in state self-objectification.
Specifically, on average, state self-objectification (measured on a 0-100 scale) increased by
approximately 11 scale points after being targeted by sexually objectifying behavior, and by
approximately three points after witnessing sexual objectifying treatment of other women.
Estimates of bW paths in Table 3 show that increases in self-objectification were
associated with small, yet reliable, increases in negative and self-conscious emotions, but also
with increases in positive emotions. Specifically, on average, a six-point increase in self-
objectification was associated with a one-point increase in self-conscious emotions, whereas
an increase of 20 and 33 points in self-objectification predicted a one-point increase in
negative and positive emotions, respectively. Supporting Fredrickson and Roberts’s (1997)
predictions, exposure to sexually objectifying events had a reliable indirect effect on negative
and self-conscious emotions via self-objectification (see indirectW estimates in Table 3). In
contrast, the indirect effects of exposure to objectifying events on positive emotions via self-
objectification were not meaningfully different from zero.
Estimates of c’W paths in Table 3 indicate that exposure to sexually objectifying
events did not have reliable direct effects on any of the emotional outcomes.
Finally, estimates of within-person total effects revealed that the combined (direct
and indirect) effect of being targeted by sexually objectifying events was predicted to be
approximately 1.8 and three-point increases in negative and self-conscious emotions,
respectively. In contrast, the total effects of witnessing objectifying events on negative and
self-conscious emotions were not reliably different from zero, and neither being targeted nor
witnessing predicted overall changes in positive emotions. Separate analyses per sample
produced substantively similar findings, although with less precise and reliable parameter
estimates than the combined analyses (see Tables S5-S7 in the supplemental materials).
EMOTIONAL IMPACT OF OBJECTIFICATION 20
Given that our primary focus in the current study was to test a within-person
mediation model, wherein exposure to objectifying events indirectly influences emotions via
self-objectification, we sought to rule out two alternate explanations for the above findings:
(i) that the emotional consequences of sexually objectifying events are (partly) driven by their
co-occurrence with other daily stressors and (ii) that increases in self-objectification precede
reported exposure to sexually objectifying events rather than vice versa.
Controlling for reactivity to other stressors. Exposure to sexually objectifying
events may be likely to co-occur with other unpleasant or stressful events, which are
themselves known to reliably impact emotions (Almeida, 2005). To investigate whether the
indirect emotional impact of sexually objectifying events was independent of reactivity to
other daily stressors, we ran additional analyses using data from Samples 2 and 3, in which
self-objectification (mediator) and emotions (outcome) were simultaneously regressed onto
objectifying events and other stressful events (see Figure S1 in the supplemental materials).
These analyses revealed that indirect effects of exposure to objectifying events on negative
and self-conscious emotions were slightly weaker after controlling for reactivity to other
daily stressors. However, all previously significant indirect effects remained positive and
reliably greater than zero, except for the indirect effect of being targeted on negative
emotions (see Table S8 in the supplemental materials).
EMOTIONAL IMPACT OF OBJECTIFICATION 21
Table 3
Results of multilevel mediation models testing effects of exposure to objectifying events on emotions via self-objectification
Predictor
Target
Witness
Outcome
Parameter
Estimate (SD)
95% CI
Estimate (SD)
95% CI
Negative Emotion
Within-Person
aW
11.13 (1.29)
[8.50, 13.58]
3.15 (0.73)
[1.78, 4.61]
bW
0.05 (0.01)
[0.03, 0.06]
0.05 (0.01)
[0.03, 0.06]
c'W
1.34 (0.73)
[-0.05, 2.82]
0.15 (0.37)
[-0.59, 0.85]
cov(aW, bW)
-0.10 (0.13)
[-0.34, 0.16]
0.04 (0.07)
[-0.11, 0.18]
indirectW
0.43 (0.16)
[0.13, 0.73]
0.19 (0.08)
[0.03, 0.35]
totalW
1.76 (0.73)
[0.33, 3.19]
0.34 (0.37)
[-0.43, 1.02]
Between-Person
aB
0.88 (0.21)
[0.48, 1.31]
0.24 (0.07)
[0.10, 0.38]
bB
0.32 (0.03)
[0.25, 0.39]
0.33 (0.03)
[0.26, 0.39]
c'B
0.22 (0.12)
[-0.01, 0.45]
0.06 (0.04)
[-0.02, 0.14]
indirectB
0.28 (0.07)
[0.14, 0.43]
0.08 (0.02)
[0.03, 0.13]
totalB
0.50 (0.13)
[0.24, 0.75]
0.14 (0.04)
[0.05, 0.22]
Self-Conscious Emotion
Within-Person
aW
11.12 (1.32)
[8.51, 13.71]
3.09 (0.75)
[1.66, 4.60]
bW
0.16 (0.01)
[0.13, 0.18]
0.16 (0.01)
[0.13, 0.18]
c'W
1.28 (0.91)
[-0.48, 3.06]
0.42 (0.55)
[-0.65, 1.49]
cov(aW, bW)
0.00 (0.26)
[-0.50, 0.51]
0.03 (0.14)
[-0.25, 0.31]
indirectW
1.74 (0.36)
[1.06, 2.46]
0.52 (0.19)
[0.16, 0.89]
totalW
3.02 (0.94)
[1.23, 4.88]
0.95 (0.57)
[-0.18, 2.04]
Between-Person
aB
0.88 (0.21)
[0.48, 1.29]
0.24 (0.07)
[0.10, 0.38]
bB
0.48 (0.04)
[0.40, 0.56]
0.49 (0.04)
[0.41, 0.57]
c'B
0.24 (0.14)
[-0.04, 0.51]
0.06 (0.05)
[-0.03, 0.15]
indirectB
0.42 (0.11)
[0.22, 0.64]
0.12 (0.04)
[0.05, 0.19]
totalB
0.66 (0.17)
[0.33, 0.99]
0.18 (0.06)
[0.07, 0.29]
Positive Emotion
Within-Person
aW
11.06 (1.26)
[8.58, 13.54]
3.18 (0.74)
[1.73, 4.68]
bW
0.03 (0.01)
[0.01, 0.05]
0.03 (0.01)
[0.01, 0.05]
c'W
0.78 (0.98)
[-1.11, 2.74]
0.04 (0.50)
[-0.94, 1.03]
cov(aW, bW)
-0.22 (0.20)
[-0.64, 0.16]
-0.04 (0.12)
[-0.28, 0.19]
indirectW
0.10 (0.23)
[-0.35, 0.55]
0.06 (0.12)
[-0.18, 0.29]
totalW
0.87 (0.98)
[-1.02, 2.81]
0.09 (0.51)
[-0.91, 1.07]
Between-Person
aB
0.88 (0.21)
[0.46, 1.28]
0.24 (0.07)
[0.10, 0.38]
bB
-0.12 (0.05)
[-0.22, -0.02]
-0.10 (0.05)
[-0.20, 0.00]
c'B
0.13 (0.18)
[-0.22, 0.47]
-0.05 (0.06)
[-0.17, 0.06]
indirectB
-0.10 (0.05)
[-0.21, -0.01]
-0.02 (0.01)
[-0.05, 0.00]
totalB
0.03 (0.17)
[-0.31, 0.37]
-0.08 (0.06)
[-0.18, 0.04]
Note. CI = Bayesian credibility interval (highest posterior density); Parameters in bold have 95% CIs that do not cross zero.
Estimates of all between-person paths (except path bB) were divided by 100 so that 1 unit reflects a difference of 1% in
prevalence of objectifying events.
EMOTIONAL IMPACT OF OBJECTIFICATION 22
Temporal precedence of objectifying events. The within-person indirect effects
reported above are consistent with mediation only if we assume that exposure to objectifying
events precedes increases in self-objectification. However, since both events and self-
objectification were assessed as occurring “since the last survey” at the same measurement
occasion (i.e., EMA survey), our findings are also potentially consistent with the opposite
temporal ordering, in which increases in self-objectification predict greater reporting of
objectifying events. To rule out this alternate interpretation, we used multilevel logistic
regression models to investigate whether increases in state self-objectification predicted a
greater probability of subsequently reporting exposure to objectifying events. These analyses
revealed that the probability of reporting objectifying events (as either target or witness) at
occasion t2 was not reliably predicted by levels of state self-objectification reported either at
the previous (t1) or concurrent (t2) occasion (see Table S9 in the supplemental materials),
while controlling for objectifying events at occasion t1. Thus, we found no evidence that
increases in self-objectification predict an increased likelihood of reporting exposure to
objectifying events.
Nevertheless, to conclusively test whether objectifying events predict subsequent
increases in self-objectification, which subsequently predicts increases in negative and self-
conscious emotions, we repeated our main multilevel mediation analyses with objectifying
events measured at the previous measurement occasion to self-objectification. In these
“doubly-lagged” models (see Figure S2 in the supplemental materials) the predictor (Eventt0-
t1) was measured before the mediator (S-Objt1-t2), which is assumed to precede the outcome
(Emotiont2). As in our main analyses, we modelled the autoregressive slopes of self-
objectification and emotions. These doubly-lagged analyses revealed that the indirect effects
of exposure to objectifying events on negative and self-conscious emotions were no longer
reliable (see Table S10 in the supplemental materials). As discussed further below, the results
EMOTIONAL IMPACT OF OBJECTIFICATION 23
of these doubly-lagged results do not necessarily undermine our main mediational findings.
Instead, taken together with the previous results showing that self-objectification does not
predict increases in exposure to objectification, these results suggest an effect of exposure to
objectification on self-objectification, and subsequently on emotions, which is relatively
rapid and may quickly fade with time.
Between-person effects. Estimates of between-person effects in Table 3 indicate
that, consistent with previous cross-sectional studies, women who reported greater overall
exposure to sexually objectifying events also reported higher mean levels of self-
objectification (see estimates of aB paths in Table 3). Furthermore, women with higher mean
levels of self-objectification also tended to report higher mean levels of negative and self-
conscious emotions and lower mean levels of positive emotions across the study period (see
estimates of bB paths in Table 3). Consistent with previous cross-sectional research, mean
levels of exposure to sexually objectifying events (as target and witness) showed reliably
positive indirect effects on mean levels of negative and self-conscious emotions via mean
self-objectification, and exposure to objectifying events was indirectly (via mean self-
objectification) associated with lower mean levels of positive emotions (see indirectB
estimates in Table 3). Finally, although individual differences in mean exposure to
objectifying events were not directly associated with mean levels of emotions (see c’B paths in
Table 3), the total effects of exposure to sexually objectifying events were reliably associated
with higher mean levels of negative and self-conscious emotions (see totalB estimates in
Table 3). Total effects for positive emotions were not reliably different from zero.
Discussion
In a landmark publication that came to shape an entire area of research within
psychology, Fredrickson and Roberts (1997) proposed that exposure to sexual objectifying
experiences causes many women to adopt a third-person perspective on their bodies–
EMOTIONAL IMPACT OF OBJECTIFICATION 24
essentially sexually objectifying themselves–which, in turn, has harmful downstream
consequences for women’s well-being. Over the past two decades, research in this area has
boomed, yielding many important insights into the psychology of sexual objectification
(Loughnan & Vaes, 2017). Yet, few studies have captured the within-person psychological
consequences of sexual objectification as they unfold in daily life. In particular, the
intraindividual process model of sexual objectification proposed by Fredrickson and Roberts
(1997) has remained untested. The current findings support Fredrickson and Roberts’s (1997)
prediction that the harmful consequences of exposure to sexually objectifying behavior on
women’s daily experiences of negative and self-conscious emotions are mediated by within-
person increases in state self-objectification.
Associations between exposure to objectifying events and self-objectification
We found strong evidence that exposure to objectifying events in daily life primes a
state of self-objectification, making women more conscious of how their body appears to
others. This association between exposure to sexually objectifying events and the tendency to
self-objectify was strongest when women reported being personally targeted by objectifying
behaviour in daily life. However, just as breathing second-hand smoke is unhealthy for non-
smokers, we found that objectifying events need not be experienced first-hand to induce the
potentially harmful process of self-objectification. Witnessing sexual objectification of other
women also reliably predicted within-person increases in state self-objectification in the
current study.
7
Supporting previous correlational research (e.g., Augustus-Horvath & Tylka,
2009; Calogero & Pina, 2011), we found similar effects at the between-person level: women
who reported greater overall exposure to objectifying events also reported higher mean levels
of self-objectification across the study period. Taken together, these findings suggest that
7
These findings are consistent with our previous analyses reported in Holland et al. (2017), based exclusively
on Sample 1 data. Our previous analyses examined the effect of exposure to objectifying events on self-
objectification (path a in the model tested here) but did not simultaneously model the effects of self-
objectification or exposure to objectifying events on emotions (paths b and c' in the model tested here).
EMOTIONAL IMPACT OF OBJECTIFICATION 25
exposure to objectifying events in daily life momentarily redirects women’s attention to their
appearance, but also that this process may accumulate over time, leading to increased
habitual (i.e., trait) self-objectification in the long run. However, this between-person
association may also reflect the opposite causal process, namely that women who are higher
in trait self-objectification are more likely to notice sexually objectifying events. While we
were not able to test this alternate mechanism in the current study, future research using
traditional longitudinal designs could do so.
Associations between self-objectification and emotions
Regarding the association between self-objectification and momentary emotions, our
findings are broadly consistent with previous research (e.g., Breines et al., 2008; Mercurio &
Landry, 2008) in that greater state self-objectification predicted intraindividual increases in
negative and self-conscious emotions. These findings suggest that engaging in body
monitoring and other forms of self-objectification heightens women’s experiences of
unpleasant emotions such as anger, guilt, shame and embarrassment. Similarly, greater
habitual self-objectification was related to higher mean levels of negative and self-conscious
emotions between persons, indicating that women who are higher in chronic self-
objectification also tend to experience more unpleasant affect, on average, in their daily lives.
However, the momentary emotional consequences of self-objectification were not
exclusively negative: we also found that higher state self-objectification predicted within-
person increases in positive emotions, suggesting that attending to their physical appearance
may also increase women’s feelings of confidence and happiness (see Calogero, Herbozo, &
Thompson, 2009). However, we observed a starkly different association between self-
objectification and positive emotions at the between-person level: women who reported
higher mean levels of self-objectification experienced lower mean levels of positive emotions
in daily life. This finding is generally consistent with previous findings linking trait self-
EMOTIONAL IMPACT OF OBJECTIFICATION 26
objectification with lower self-esteem (e.g., Strelan, Mehaffey, & Tiggermann, 2003) and
psychological well-being (e.g., McKinley, 2006). Thus, whereas state self-objectification
may be accompanied by momentary spikes in positive feelings, habitually engaging in self-
objectification appears to be associated with lower tonic levels of positive emotions. These
paradoxical findings highlight the importance of examining both within- and between-person
effects, which may differ in both magnitude and direction (Fisher et al., 2018).
Indirect effects of objectifying events on negative and self-conscious emotions
The current study was the first, to our knowledge, to test the within-person indirect
effect of exposure to sexually objectifying events on women’s emotions in daily life,
originally hypothesized by Fredrickson and Roberts (1997). Our findings therefore contribute
to cumulative theoretical knowledge in the psychology of sexual objectification by providing
the first evidence to support this hypothesis. Specifically, we found that despite not having a
direct impact on women’s emotions, exposure to sexually objectifying events in daily life
reliably predicted increases in negative and self-conscious emotions via state self-
objectification. These findings suggest that being targeted by or witnessing sexual
objectification in daily life attunes women to their bodily appearance, which, in turn,
intensifies women’s experiences of negative (e.g., anger) and self-conscious (e.g., shame)
emotions. Previous studies have tested individual paths from Fredrickson and Roberts’s
(1997) model, demonstrating that exposure to objectification predicts increases in self-
objectification (e.g., Holland et al., 2017) and, separately, that state self-objectification is
associated with heightened negative emotions (Breines et al., 2008). However, this is the first
study to model these effects simultaneously and estimate the within-person indirect effect of
exposure to sexually objectifying events on emotions via state self-objectification.
The within-person indirect effect of exposure to objectifying events was strongest
for self-conscious emotions (embarrassment, shame) and when women were personally
EMOTIONAL IMPACT OF OBJECTIFICATION 27
targeted by objectifying behaviour. Similarly, at the between-person level, women who were
more frequently targeted by objectifying experiences reported higher mean levels of self-
conscious emotions, and this association was statistically accounted for by their higher mean
levels of self-objectification. Although somewhat weaker, we also found evidence for reliable
within- and between-person indirect effects of being targeted by objectifying events on
general negative affect (anger, sadness, anxiety, and guilt). The estimated within- and
between-person indirect effects of witnessing objectifying events on negative and self-
conscious emotions were substantially weaker but still reliably different from zero.
Taken together, these findings suggest that even vicarious exposure to sexual
objectification may be harmful to women, although perhaps not as acutely harmful as being
directly objectified. However, given the much higher frequency of witnessed versus targeted
objectification reported in all three samples, cumulative vicarious exposure to objectification
may eventually take an emotional toll on women. Frequently witnessing sexual
objectification of other women may serve as a reminder that such treatment is difficult to
escape if one is female. However, the high prevalence of sexual objectification in women’s
daily lives may persist because while witnessing sexually objectifying treatment of others
may evoke unpleasant emotions, it is often met with inaction (Cunningham, Miner, &
Benavides-Espinoza, 2012).
Controlling for reactivity to daily stressors. To test the robustness of our within-
person mediation findings and rule out alternate explanations, we conducted a number of
additional analyses. In particular, because objectifying events may be more likely to occur in
generally stressful contexts (e.g., being groped on a crowded train), we conducted additional
analyses to investigate whether the emotional consequences of objectifying events were
independent of reactivity to other daily stressors or hassles. When controlling for the
emotional impact of daily stressors, the within-person indirect effects of objectifying events
EMOTIONAL IMPACT OF OBJECTIFICATION 28
on negative and self-conscious emotions were slightly attenuated, but remained reliably
positive. These findings suggest that the indirect emotional effects of exposure to sexual
objectification are largely independent of reactivity to other daily hassles.
Establishing temporal precedence of objectifying events. Our main analyses were
conducted using measures of objectifying events, self-objectification, and emotions all
assessed at the same occasion (i.e., EMA survey). We used this approach because both events
and self-objectification were measured as occurring “since the last survey”, and can therefore
be assumed to precede emotions, which were reported as “right now”. However, this
approach does not guarantee that exposure to objectifying events preceded changes in self-
objectification. Thus, to test the theorised
objectifying event
→
self-objectification
→
emotion
sequence, we repeated our main analyses using a “doubly-lagged” approach (see Figure S2
and Table S10 in the supplemental materials). Specifically, objectifying events reported as
occurring between t0 and t1 (Eventt0-t1) were included as predictors of self-objectification
between t1 and t2 (S-Objt1-t2), which predicted emotions at t2 (Emotiont2). In these analyses,
within-person indirect effects unexpectedly trended in a negative direction, due to the fact
that objectifying events predicted decreases in self-objectification at the next EMA survey.
Importantly, this inverse effect was only evident when statistically controlling for self-
objectification reported in the same time-interval as the objectifying events. In contrast, when
concurrent self-objectification was not included in these analyses, objectifying events no
longer predicted decreases in self-objectification measured at the next occasion (see Figure
S2 and Table S10 in supplemental materials for more detail).
Considered together with our main analyses, which showed a positive
contemporaneous association between objectifying events and self-objectification (see aW
paths in Table 3), these findings seem to suggest that self-objectification increases at the time
of exposure to objectifying events but then decreases relatively quickly (i.e., during the next
EMOTIONAL IMPACT OF OBJECTIFICATION 29
assessment period). While an alternative explanation for this negative effect might be that
increases in self-objectification occur prior to sexually objectifying events, our additional
analyses showing that self-objectification did not predict increased reporting of objectifying
events at the next occasion (see Table S9 in supplemental materials) suggest otherwise. In
sum, our findings indicate that exposure to sexually objectifying events in daily life predicts
reliable yet transient increases in self-objectification, which lead to subsequent increases in
negative and self-conscious emotions.
The above findings have important implications for understanding the psychological
consequences of women’s exposure to sexual objectification. On one hand, given the
frequency of objectifying events in women’s daily lives, even relatively transient increases in
self-objectification—which appear to have adverse downstream emotional consequences—
may be cumulatively harmful for women’s well-being. On the other hand, the short-term
psychological impact of sexually objectifying events may reflect that many women develop
coping skills to minimize the harmful psychological consequences of objectifying events,
rendering them relatively resilient to the effects of sexual objectification (Fredrickson &
Roberts, 1997). For instance, adopting an accepting and compassionate attitude towards the
self may buffer women against the harmful psychological effects of sexual objectification
(Liss & Erchull, 2015).
Indirect effects of objectifying events on positive emotions
As predicted, and in contrast to our findings for negative and self-conscious
emotions, we found no reliable evidence of within-person indirect effects of exposure to
objectifying events (either as target or witness) on positive emotions. Thus, while state self-
objectification may predict momentary increases in both negative and positive emotions, the
indirect emotional consequences of exposure to objectifying events appear to be exclusively
negative. Furthermore, at the between-person level, exposure to objectifying events was
EMOTIONAL IMPACT OF OBJECTIFICATION 30
indirectly associated with lower mean levels of positive emotions. Thus, women who
reported more frequent exposure to objectifying events over the course of the study tended to
report lower mean levels of positive emotions, and this association was statistically accounted
for by their higher mean levels of self-objectification.
Direct effects of objectifying events on emotions
Finally, while the current study provides preliminary evidence that exposure to
objectifying events indirectly results in reduced emotional well-being, we found no reliable
evidence for the proposed direct pathway from objectifying events to reduced well-being,
either within or between persons. Although this aligns with some previous research (e.g.,
Tiggemann & Williams, 2012), it contradicts findings from other studies (Prichard &
Tiggemann, 2012).
Limitations and future directions
While the current study makes an important contribution to research on sexual
objectification by providing the first test of Fredrickson and Roberts’s (1997) hypothesized
within-person process model, we wish to acknowledge several limitations.
First, the use of EMA (a self-report methodology) to measure exposure to
objectifying events in daily life may be problematic for at least two reasons. Asking our
participants to report sexually objectifying events in their daily lives may have inadvertently
made them more vigilant or emotionally reactive to such behaviour. In addition, due to the
subjective nature of self-reports, we cannot be certain that different participants interpreted
sexually objectifying events in similar ways. Obtaining objective measures or peer-reports of
exposure to objectifying events would circumvent both the potential measurement reactivity
and inherent subjectivity of EMA. However, besides the difficulty of obtaining such data,
there may be other reasons to consider self-reports a valid method for measuring exposure to
objectifying events. First, mitigating concerns about potential measurement reactivity,
EMOTIONAL IMPACT OF OBJECTIFICATION 31
previous research demonstrates that whether or not women label events as “sexual
harassment” does not influence the resulting psychological harm (Magley, Hulin, Fitzgerald,
& DeNardo, 1999). Second, subjective appraisals of events, rather than their objective
features, are thought to largely determine emotional responding (e.g., Lazarus, 1991). Thus,
rather than abandoning self-report in favor of more objective assessment methods, future
EMA research could include open-ended event descriptions and/or ratings of events on
various appraisal dimensions (e.g., intensity, controllability) to obtain a richer qualitative
understanding of the sexually objectifying events (and other daily hassles) that women
encounter in their daily lives.
Second, consistent with early work on objectification (e.g., Fredrickson & Roberts,
1997; Kozee et al., 2007; Swim et al., 2001), we operationalised objectifying events
relatively narrowly as sexualized perception or behaviors (e.g., ogling, catcalling, unwanted
touching). More recently, researchers have begun to examine broader forms of objectification
involving the perception or treatment of others in appearance-based or instrumental, but not
necessarily sexual, ways (e.g., Loughnan, Haslam, Murnane, Vaes, Reynolds, & Suitner,
2010; Morris, Goldenberg, & Boyd, 2018). Yet, very little is known about the impact of
being targeted by (or witnessing) different forms of objectification. Thus, future research
should investigate the psychological consequences of exposure to a wider range of
objectifying events in daily life beyond the sexually objectifying behaviors examined in the
current study.
A third limitation of the current study is that our main findings supporting
Fredrickson and Roberts’s (1997) theorized within-person mediation model are based on
analyses in which the predictor, mediator and outcome variables were assessed at the same
measurement occasion. Our follow-up longitudinal mediation analyses suggest that the
interval between EMA surveys in the current study (approx. 60-90 minutes) may have been
EMOTIONAL IMPACT OF OBJECTIFICATION 32
too long to detect the short-term lagged effect of objectifying events on self-objectification.
Choosing the most appropriate sampling frequency is a major challenge in EMA studies,
which can substantially influence the obtained results (Bolger & Laurenceau, 2013; Ebner-
Priemer & Sawitzki, 2007). We suggest that future studies should seek to replicate the current
findings using more intensive EMA sampling to capture the within-person process of
objectification at a more fine-grained timescale.
A fourth limitation relates to the participant samples recruited for the current study,
which comprised women from a fairly limited range of ages (~18-40 years old) and cultural
contexts (Australia and the USA). Although our samples were relatively diverse in terms of
their ethnic composition and relationship status, it will be important to explore how the
within-person process of sexual objectification plays out among women across the lifespan
living in a variety of cultural contexts. For instance, Fredrickson and Roberts’s (1997)
objectification theory proposes that as women age they may be targeted by sexual
objectification less frequently and this may be accompanied by a reduced tendency to view
themselves in terms of their physical or sexual value to others, resulting in fewer harmful
psychological consequences. On the contrary, for women who have internalised the cultural
injunction to remain youthful and attractive as they age, the process of sexual objectification
may continue to exert harmful psychological consequences in later life (Fredrickson &
Roberts, 1997).
Finally, given our aim in the current study was to test a within-person mediation
model of sexual objectification proposed by Fredrickson and Roberts (1997), we focused
exclusively on estimating average within-person effects. However, future research should
explore potential moderators of each of the within-person paths tested in the current study.
For instance, previous research suggests that the within-person association between self-
objectification and emotional well-being may differ between individuals (Breines et al.,
EMOTIONAL IMPACT OF OBJECTIFICATION 33
2008). Thus, while we found that state self-objectification predicted increases in positive and
negative emotions, on average, across three samples of young women, these effects may be
moderated by individual differences in BMI, body dissatisfaction, feminist identification, or
enjoyment of sexualization (Liss, Erchull, & Ramsay, 2011).
Despite the limitations noted above, we believe the current study makes an
important contribution to the literature on sexual objectification by investigating the
emotional consequences of “real-world” exposure to objectifying events. Furthermore, this
study provides the first comprehensive test of Fredrickson and Roberts’s (1997) theorized
mediation model at both the within- and between-person levels. We hope this study provides
an impetus for researchers to continue studying the intraindividual dynamics of sexual
objectification in daily life and thereby to develop a richer understanding of its consequences
for women’s well-being.
EMOTIONAL IMPACT OF OBJECTIFICATION 34
References
Asparouhov, T. & Muthén, B. (2018). Latent variable centering of predictors and mediators
in multilevel and time-series models. Structural Equation Modeling: A
Multidisciplinary Journal. Advance online publication (Sep 28th, 2018). doi:
10.1080/10705511.2018.1511375.
Augustus-Horvath, C. L., & Tylka, T. L. (2009). A test and extension of objectification
theory as it predicts disordered eating: Does women's age matter? Journal of
Counseling Psychology, 56, 253-265.
Bolger, N., & Laurenceau, J. P. (2013). Intensive Longitudinal Methods: An Introduction to
Diary and Experience Sampling Research. New York, NY: Guilford.
Breines, J. G., Crocker, J., & Garcia, J. A. (2008). Self-objectification and well-being in
women's daily lives. Personality and Social Psychology Bulletin, 34, 583-598.
Brinkman, B. G., & Rickard, K. M. (2009). College students’ descriptions of everyday
gender prejudice. Sex Roles, 61, 461-475.
Calogero, R.M., Herbozo, S., & Thompson, J.K. (2009). Complimentary weightism: The
potential costs of appearance-related commentary for women’s self-objectification.
Psychology of Women Quarterly, 33, 120-132.
Calogero, R. M., & Pina, A. (2011). Body guilt: Preliminary evidence for a further subjective
experience of self-objectification. Psychology of Women Quarterly, 35, 428-440.
Cunningham, G.B., Miner, K. & Benavides-Espinoza, C. (2012). Emotional reactions to
observing misogyny: Examining the roles of gender, forecasting, political
orientation and religiosity. Sex Roles, 67, 58-68.
Curran, P. J., & Bauer, D. J. (2011). The disaggregation of within-person and between-person
effects in longitudinal models of change. Annual Review of Psychology, 62, 583-
619.
EMOTIONAL IMPACT OF OBJECTIFICATION 35
Ebner-Priemer, U. W., & Sawitzki, G. (2007). Ambulatory assessment of affective instability
in borderline personality disorder. European Journal of Psychological Assessment,
23, 238-247.
Fisher, A. J., Medaglia, J. D., & Jeronimus, B. F. (2018). Lack of group-to-individual
generalizability is a threat to human subjects research. Proceedings of the National
Academy of Sciences of the United States of America, 201711978.
Fleeson, W., & Gallagher, P. (2009). The implications of Big Five standing for the
distribution of trait manifestation in behavior: Fifteen experience-sampling studies
and a meta-analysis. Journal of Personality and Social Psychology, 97, 1097-1114.
Fredrickson, B. L., & Roberts, T. (1997). Objectification theory: Towards understanding
women’s lived experiences and mental health risks. Psychology of Women
Quarterly, 21, 173-206.
Fredrickson, B. L., Hendler, L. M., Nilsen, S., O’Barr, J. F., & Roberts, T. A. (2011).
Bringing back the body: A retrospective on the development of objectification
theory. Psychology of Women Quarterly, 35, 689-696.
Gapinski, K. D., Brownell, K. D., & Lafrance, M. (2003). Body objectification and "fat talk":
effects on emotion, motivation, and cognitive performance. Sex Roles, 48, 377-388.
Geldhof, G. J., Preacher, K. J., & Zyphur, M. J. (2014). Reliability estimation in a multilevel
confirmatory factor analysis framework. Psychological Methods, 19, 72–91.
Gervais, S. J., Vescio, T. K., & Allen, J. (2011). What you see is what you get: The
consequences of the objectifying gaze for women and men. Psychology of Women
Quarterly, 35, 5–17.
EMOTIONAL IMPACT OF OBJECTIFICATION 36
Harrison, A., Harrsion, S., Koval, P., Gleeson, J., Alvarez, M. (2017). SEMA: Smartphone
Ecological Momentary Assessment (Version 2). [Computer software]. Available
online at: https://github.com/eorygen.
Hamaker, E. L., & Wichers, M. (2017). No time like the present: Discovering the hidden
dynamics in intensive longitudinal data. Current Directions in Psychological
Science, 26, 10-15.
Hamaker, E. L. (2012). Why researchers should think “within-person”: A paradigmatic
rationale in M. R. Mehl & T. S. Conner (Eds). Handbook of Research Methods for
Studying Daily Life (pp. 43-61). New York: Guilford.
Hamaker, E. L., Asparouhov, T., Brose, A., Schmiedek, F., & Muthen, B. (2018). At the
frontiers of modeling intensive longitudinal data: Dynamic structural equation
models for the affective measurements from the COGITO study. Multivariate
Behavioural Research (advance online publication).
Hebl, M. R., King, E. B., & Lin, J. (2004). The swimsuit becomes us all: Ethnicity, gender,
and vulnerability to self-objectification. Personality and Social Psychology Bulletin,
30, 1322-1331.
Hollaback! (2016). Cornell International Survey on Street Harassment. Retrieved from
http://www.ihollaback.org/cornell-international-survey-on-street-harassment/
Holland, E., Koval, P., Stratemeyer, M., Thomson, F., & Haslam, N. (2017). Sexual
objectification in women's daily lives: A smartphone ecological momentary
assessment study. British Journal of Social Psychology, 56, 314-333.
Johnson, M., & Bennett, E. (2015). Everyday sexism: Australian women’s experiences of
street harassment. The Australia Institute, Canberra. Retrieved from
http://www.tai.org.au/sites/defualt/files/Everyday_sexism_TAIMarch2015_0.pdf
EMOTIONAL IMPACT OF OBJECTIFICATION 37
Karsay, K., & Matthes, J. (2016). Sexually objectifying pop music videos, young women’s
self objectification, and selective exposure: A moderated mediation model.
Communication Research. Advance online publication (Jul 28th, 2016). doi:
10.1177/0093650216661434
Kenny, D. A., & Judd, C. M. (2014). Power anomalies in testing mediation. Psychological
Science, 25, 334–339.
Kenny, D. A., Korchmaros, J. D., & Bolger, N. (2003). Lower level mediation in multilevel
models. Psychological Methods, 8, 115–128.
Kline, R. B. (2015). The mediation myth. Basic and Applied Social Psychology, 37, 202-213.
Kozee, H. B., Tylka, T. L., Augustus‐Horvath, C. L., & Denchik, A. (2007). Development
and psychometric evaluation of the interpersonal sexual objectification scale.
Psychology of Women Quarterly, 31, 176-189.
Liss, M., & Erchull, M. J. (2015). Not hating what you see: Self-compassion may protect
against negative mental health variables connected to self-objectification in college
women. Body Image, 14, 5-12.
Liss, M., Erchull, M. J., & Ramsey, L. R. (2011). Empowering or oppressing? Development
and exploration of the Enjoyment of Sexualization scale. Personality and Social
Psychology Bulletin, 37, 55–68.
Loughnan, S., Haslam, N., Murnane, T., Vaes, J., Reynolds, C., & Suitner, C. (2010).
Objectification leads to depersonalization: The denial of mind and moral concern to
objectified others. European Journal of Social Psychology, 40, 709-717.
Loughnan, S., & Vaes, J. (2017). Objectification: Seeing and treating people as things.
British Journal of Social Psychology, 56, 213-216.
Magley, V. J., Hulin, C. L., Fitzgerald, L. F., & DeNardo, M. (1999). Outcomes of self-
labeling sexual harassment. Journal of Applied Psychology, 84, 390-402.
EMOTIONAL IMPACT OF OBJECTIFICATION 38
Maxwell, S. E., & Cole, D. A. (2007). Bias in cross-sectional analyses of longitudinal
mediation. Psychological Methods, 12, 23-44.
McKinley, N. M. (2006). The developmental and cultural contexts of objectified body
consciousness: a longitudinal analysis of two cohorts of women. Developmental
Psychology, 42, 679-687.
Mercurio, A. E., & Landry, L. J. (2008). Self-objectification and well-being: The impact of
self-objectification on women’s overall sense of self-worth and life satisfaction. Sex
Roles, 58, 458-466.
Miles-McLean, H., Liss, M., Erchull, M. J., Robertson, C. M., Hagerman, C., Gnoleba, M.
A., & Papp, L. J. (2015). “Stop looking at me!” Interpersonal sexual objectification
as a source of insidious trauma. Psychology of Women Quarterly, 39, 363-374.
Miner-Rubino, K., Twenge, J., & Fredrickson, B. (2002). Trait self-objectification in women:
Affective and personality correlates. Journal of Research in Personality, 36, 147-
172.
Morris, K. L., Goldenberg, J., & Boyd, P. (2018). Women as animals, women as objects:
Evidence for two forms of objectification. Personality and Social Psychology
Bulletin, 44, 1302-1314.
Muthén, L. K., & Muthén, B. O. (1998-2017). Mplus User’s Guide. (8th Ed.). Los Angeles,
CA: Muthén & Muthén.
Roberts, T.-A., Calogero, R. M., & Gervais, S. J. (2018). Objectification theory: Continuing
contributions to feminist psychology. In C. B. Travis, J. W. White, A. Rutherford,
W. S. Williams, S. L. Cook, & K. F. Wyche (Eds.), APA handbook of the
psychology of women: History, theory, and battlegrounds (pp. 249-271).
Washington, DC, US: American Psychological Association.
EMOTIONAL IMPACT OF OBJECTIFICATION 39
Roberts, T-A. & Gettman, J.Y. (2004). Mere exposure: Gender differences in the negative
effects of priming a state of self-objectification. Sex Roles, 51, 17-27.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies:
new procedures and recommendations. Psychological Methods, 7, 422–445.
Sijtsma, K. (2009). On the use, the misuse, and the very limited usefulness of Cronbach’s
alpha. Psychometrika, 74, 107–120.
Steinberg, K. K., Smith, S. J., Stroup, D. F., Olkin, I., Lee, N. C., Williamson, G. D., &
Thacker, S. B. (1997). Comparison of effect estimates from a meta-analysis of
summary data from published studies and from a meta-analysis using individual
patient data for ovarian cancer studies. American Journal of Epidemiology, 145,
917-925.
Strelan, P., Mehaffey, S. J., & Tiggemann, M. (2003). Brief report: Self-objectification and
esteem in young women: The mediating role of reasons for exercise. Sex Roles, 48,
89-95.
Swim, J. K., Hyers, L. L., Cohen, L. L., & Ferguson, M. J. (2001). Everyday sexism:
Evidence for its incidence, nature, and psychological impact from three daily diary
studies. Journal of Social Issues, 57, 31-53.
Tiggemann, M., & Williams, E. (2012). The role of self-objectification in disordered eating,
depressed mood, and sexual functioning among women: A comprehensive test of
objectification theory. Psychology of Women Quarterly, 36, 66-75.
Ward, L. M. (2016). Media and sexualisation: State of empirical research, 1995–2015. The
Journal of Sex Research, 53, 560-577.
EMOTIONAL IMPACT OF OBJECTIFICATION 1
Supplemental Materials
Descriptive Statistics
Tables S1-S3 display descriptive statistics and reliabilities in each sample. Table S1
displays descriptive statistics for demographic variables in each sample.
Table S1
Demographic characteristics of Samples 1-3
Variable
Sample 1 (n = 81)a
Sample 2 (n = 87)b
Sample 3 (n = 100)c
Age
Mean (SD)
22.33 (5.47)
23.52 (4.11)
26.46 (6.12)
Range
18–46
18–35
18–40
Ethnicity, n (%)
White/Caucasian
38 (46.9%)
28 (32.2%)
65 (65.0%)
Asian
25 (30.9%)
33 (37.9%)
15 (15.0%)
South Asian
9 (11.1%)
10 (11.5%)
0 (0.0%)
Middle Eastern
2 (2.5%)
2 (2.3%)
0 (0.0%)
Indigenous
0 (0.0%)
0 (0.0%)
1 (1.0%)
Black/African
0 (0.0%)
0 (0.0%)
10 (10.0%)
Mixed ethnicity
4 (4.9%)
6 (6.9%)
8 (8.0%)
Other
3 (3.7%)
8 (9.2%)
1 (1.0%)
Country of birth, n (%)
Australia
28 (34.6%)
28 (32.2)
1 (1.0%)
Malaysia
11 (13.6%)
10 (11.5)
0 (0.0%)
Singapore
7 (8.6%)
3 (3.4%)
0 (0.0%)
United Kingdom
7 (8.6%)
1 (1.1%)
0 (0.0%)
Sri Lanka
4 (4.9%)
0 (0.0%)
0 (0.0%)
China
3 (3.7%)
5 (5.7%)
2 (2.0%)
India
2 (2.5%)
10 (11.5%)
1 (1.0%)
Indonesia
2 (2.5%)
5 (5.7%)
0 (0.0%)
USA
2 (2.5%)
3 (3.4%)
89 (89.0%)
Other
15 (18.5%)
22 (25.3%)
7 (7.0%)
Sexual orientation, n (%)
Heterosexual
74 (91.3%)
76 (87.4%)
69 (69.0%)
Homosexual
2 (2.5%)
1 (1.1%)
8 (8.0%)
Bisexual
5 (6.2%)
7 (8.0%)
18 (18.0%)
Other
0 (0.0%)
3 (3.4%)
5 (5.0%)
Relationship status, n (%)
Single
52 (64.2%)
52 (59.8%)
37 (37.0%)
In a relationship (unmarried)
26 (32.1%)
27 (31.0%)
41 (41.0%)
Married
1 (1.2%)
7 (8.0%)
19 (19.0%)
Other
2 (2.5%)
1 (1.1%)
3 (3.0%)
Note. aSample 1 data were collected in Australia in 2015.
bSample 2 data were collected in Australia in 2016.
cSample 3 data were collected in the USA in 2016-2017.
EMOTIONAL IMPACT OF OBJECTIFICATION 2
Table S2 displays descriptive statistics and reliability statistics for continuous
measures in each sample.
Table S2
Descriptive statistics and reliability estimates for continuous measures in Samples 1-3
Sample size
SD (% variance)
Cronbach’s alpha
Measure
T
N
Mean
Within
Between
Within
Between
Negative Emotion
Sample 1
4843
81
20.95
12.95 (46%)
13.98 (54%)
0.62
0.89
[17.98, 24.08]
[12.69, 13.2]
[11.83, 16.38]
[0.60, 0.64]
[0.85, 0.93]
Sample 2
5108
87
13.14
12.07 (47%)
12.81 (53%)
0.61
0.85
[10.40, 15.85]
[11.84, 12.31]
[10.99, 14.96]
[0.59, 0.62]
[0.79, 0.89]
Sample 3
5706
100
14.67
11.85 (45%)
13.21 (55%)
0.62
0.88
[12.12, 17.28]
[11.63, 12.08]
[11.48, 15.31]
[0.60, 0.63]
[0.84, 0.92]
Positive Emotion
Sample 1
4839
81
59.33
16.75 (53%)
15.88 (47%)
0.68
0.93
[55.90, 62.87]
[16.40, 17.07]
[13.45, 18.64]
[0.66, 0.7]
[0.90, 0.96]
Sample 2
5096
87
56.77
19.45 (56%)
17.30 (44%)
0.62
0.84
[53.01, 60.34]
[19.07, 19.83]
[14.78, 20.15]
[0.6, 0.64]
[0.76, 0.9]
Sample 3
5699
100
61.63
17.27 (46%)
18.80 (54%)
0.65
0.93
[58.01, 65.38]
[16.95, 17.6]
[16.24, 21.64]
[0.63, 0.67]
[0.90, 0.96]
Self-Conscious Emotion
Sample 1
4834
81
30.86
20.80 (53%)
19.77 (47%)
―
―
[26.62, 35.29]
[20.37, 21.2]
[16.87, 23.37]
Sample 2
5067
87
6.38
12.15 (60%)
10.00 (40%)
0.69
0.98
[4.23, 8.51]
[11.91, 12.39]
[8.55, 11.69]
[0.67, 0.71]
[0.97, 0.99]
Sample 3
5683
100
7.86
11.37 (53%)
10.81 (48%)
0.66
0.98
[5.79, 10.02]
[11.17, 11.59]
[9.35, 12.5]
[0.64, 0.68]
[0.97, 0.99]
Self-Objectification
Sample 1
4822
81
34.40
24.27 (57%)
20.93 (43%)
―
―
[29.85, 39.07]
[23.79, 24.77]
[17.72, 24.59]
Sample 2
5050
87
21.42
22.14 (55%)
20.17 (45%)
0.90
0.99
[17.05, 25.65]
[21.70, 22.57]
[17.23, 23.46]
[0.89, 0.90]
[0.99, 1.00]
Sample 3
5659
100
23.15
19.87 (43%)
22.82 (57%)
0.88
0.99
[18.78, 27.72]
[19.51, 20.25]
[19.61, 26.21]
[0.87, 0.88]
[0.99, 1.00]
Note. T = number of occasions; N = number of participants.
95% Bayesian credible intervals are shown in square brackets below each estimate.
Cronbach’s alphas for Self-Conscious Emotion and Self-Objectification were only calculated for Samples 2 and 3 because Sample 1
included only single-item measures.
EMOTIONAL IMPACT OF OBJECTIFICATION 3
Table S3 displays descriptive statistics for objectifying events in each sample.
Table S3
Frequencies of sexually objectifying events reported in Samples 1-3
Frequency
Proportion
N
N > 0 (%)
Range
Median
Mean (SD)
Range
Median
Mean (SD)
Target
Sample 1
81
61 (75%)
0–27
2.00
3.77 (5.41)
0.00–0.38
0.03
0.06 (0.09)
Sample 2
87
58 (67%)
0–17
1.00
2.26 (3.24)
0.00–0.30
0.02
0.04 (0.06)
Sample 3
100
58 (58%)
0–20
1.00
2.36 (3.95)
0.00–0.32
0.02
0.04 (0.07)
Witness
Sample 1
81
71 (88%)
0–58
6.00
9.43 (10.77)
0.00–0.88
0.11
0.15 (0.17)
Sample 2
87
73 (84%)
0–58
3.00
7.85 (12.88)
0.00–1.00
0.07
0.14 (0.23)
Sample 3
100
83 (83%)
0–52
3.50
6.44 (9.30)
0.00–1.00
0.06
0.12 (0.18)
Note. N > 0 = number of participants who reported being targeted or witnessing sexually objectifying events at least once.
Frequency = number of EMA surveys on which participants reported being targeted by or witnessing sexually objectifying events.
Proportion = proportion of all EMA surveys on which participants reported being targeted by or witnessing sexually objectifying
events.
Alternate Specifications of Multilevel Mediation Models
Models including sample dummy variables. Table S4 displays model fit statistics
and estimates of within-person indirect effects from our original multilevel mediation models
(for full results, see Table 3 in main text) and alternate models including dummy variables
coding for differences between the three samples. Alternate models included dummy
variables representing two out of the three samples (e.g., dummy variables for S2 and S3) and
the third sample (e.g., S1) as the reference category, in which all random intercepts and
slopes were regressed onto the dummy variables at the between-person level. For example,
the models with S2 and S3 dummies tested whether the intercepts (representing mean levels
of the predictor, mediator and outcome) and the slopes (representing the a, b and c’ paths and
autoregressive effects) differed in Samples 2 and 3 versus Sample 1. Estimates of the within-
person indirect effects in Table S4 reflect the model predicted indirect effect for the reference
category.
EMOTIONAL IMPACT OF OBJECTIFICATION 4
Table S4.
Model fit statistics and indirect effect estimates from multilevel mediation models with dummy variables representing differences across
samples
Predictor
Target
Witness
Outcome
Model
DIC
IndirectW (SD)
95% CI
DIC
IndirectW (SD)
95% CI
Negative
Emotion
Original
(Table 3 in main text)
492781.65
0.43 (0.16)
[0.13, 0.73]
503927.10
0.19 (0.08)
[0.03, 0.35]
S2 & S3 dummies
(Reference Category: S1)
492828.88
0.47 (0.21)
[0.08, 0.90]
504083.00
0.19 (0.11)
[–0.02, 0.40]
S1 & S3 dummies
(Reference Category: S2)
492828.58
0.46 (0.22)
[0.05, 0.92]
504082.59
0.29 (0.12)
[0.06, 0.54]
S1 & S2 dummies
(Reference Category: S3)
492831.24
0.32 (0.20)
[–0.07, 0.72]
504083.26
0.09 (0.10)
[–0.10, 0.28]
Self-Conscious
Emotion
Original
(Table 3 in main text)
504477.90
1.74 (0.36)
[1.06, 2.46]
515608.23
0.52 (0.19)
[0.16, 0.89]
S2 & S3 dummies
(Reference Category: S1)
504472.23
3.61 (0.73)
[2.18, 5.07]
515692.59
1.11 (0.39)
[0.36, 1.88]
S1 & S3 dummies
(Reference Category: S2)
504471.15
0.75 (0.34)
[0.14, 1.45]
515692.41
0.39 (0.18)
[0.06, 0.78]
S1 & S2 dummies
(Reference Category: S3)
504469.78
0.65 (0.32)
[0.03, 1.30]
515691.88
0.12 (0.15)
[–0.18, 0.41]
Positive
Emotion
Original
(Table 3 in main text)
515577.17
0.10 (0.23)
[–0.35, 0.55]
526542.16
0.06 (0.12)
[–0.18, 0.29]
S2 & S3 dummies
(Reference Category: S1)
515516.30
–0.01 (0.29)
[–0.59, 0.57]
526493.03
0.01 (0.14)
[–0.27, 0.29]
S1 & S3 dummies
(Reference Category: S2)
515516.20
0.31 (0.29)
[–0.25, 0.91]
526491.22
0.18 (0.16)
[–0.14, 0.50]
S1 & S2 dummies
(Reference Category: S3)
515516.28
0.09 (0.29)
[–0.48, 0.65]
526489.64
–0.02 (0.13)
[–0.28, 0.24]
Note. DIC = Deviance information criterion index of model fit; DIC values shaded in gray are the lowest (i.e. best fitting) within a set of nested models.
SD = posterior standard deviation.
CI = Bayesian credibility interval (highest posterior density).
Parameters in bold have 95% CIs that do not cross zero.
Results in Table S4 show that our original model specification (excluding dummy
variables) had the lowest DIC values (indicating the best model fit) for the models in which
Negative Emotion was predicted by Target and Witness and Self-Conscious Emotion was
predicted by Witness. This suggests that for these models, parameter estimates did not vary
substantially between samples. For the remaining three models (Self-Conscious Emotion
predicted by Target; Positive Emotion predicted by Target and Witness), one of the alternate
model specifications (including dummy variables) showed better model fit than our original
EMOTIONAL IMPACT OF OBJECTIFICATION 5
specification, as indicated by lower DIC values. This suggests that some of the model
parameters varied meaningfully between samples. However, in these models the crucial
within-person indirect effects were relatively similar across different model specifications.
Specifically, although point estimates of within-person indirect effects varied across model
specifications, their 95% CIs overlapped considerably. Furthermore, the indirect effect of
Target on Self-Conscious Emotion had 95% CIs that consistently did not cross zero, whereas
the indirect effect of Target and Witness on Positive Emotion had 95% CIs that consistently
did cross zero in all model specifications.
Separate analyses for each sample. In addition, we re-ran all multilevel mediation
analyses separately using data from each sample. Tables S5-S7 display results of these
separate multilevel mediation models using data only from Samples 1-3, respectively.
EMOTIONAL IMPACT OF OBJECTIFICATION 6
Table S5
Results of multilevel Mediation Models Using Only Sample 1 Data
Predictor
Target
Witness
Outcome
Parameter
Estimate (SD)
95% CI
Estimate (SD)
95% CI
Negative Emotion
Within-Person
aW
12.36 (2.25)
[7.85, 16.8]
3.74 (1.42)
[0.94, 6.51]
bW
0.05 (0.01)
[0.03, 0.07]
0.04 (0.01)
[0.02, 0.07]
c'W
0.35 (1.12)
[–1.88, 2.54]
0.50 (0.67)
[–0.75, 1.84]
cov(aW, bW)
–0.20 (0.25)
[–0.70, 0.29]
–0.01 (0.13)
[–0.27, 0.26]
indirectW
0.38 (0.29)
[–0.16, 0.99]
0.15 (0.16)
[–0.17, 0.47]
totalW
0.74 (1.11)
[–1.48, 2.91]
0.65 (0.69)
[–0.66, 2.06]
Between-Person
aB
–0.17 (0.30)
[–0.76, 0.43]
0.20 (0.15)
[–0.09, 0.49]
bB
0.35 (0.07)
[0.21, 0.48]
0.36 (0.07)
[0.22, 0.49]
c'B
–0.13 (0.17)
[–0.47, 0.20]
–0.05 (0.09)
[–0.23, 0.12]
indirectB
–0.06 (0.11)
[–0.28, 0.15]
0.07 (0.06)
[–0.03, 0.19]
totalB
–0.19 (0.20)
[–0.58, 0.21]
0.02 (0.10)
[–0.18, 0.22]
Self-Conscious Emotion
Within-Person
aW
12.43 (2.29)
[7.87, 16.87]
3.72 (1.41)
[1.00, 6.54]
bW
0.30 (0.03)
[0.25, 0.35]
0.30 (0.03)
[0.25, 0.35]
c'W
–0.60 (1.71)
[–4.08, 2.68]
0.30 (1.28)
[–2.25, 2.81]
cov(aW, bW)
–0.07 (0.53)
[–1.12, 0.98]
–0.17 (0.35)
[–0.87, 0.54]
indirectW
3.66 (0.99)
[1.84, 5.74]
0.94 (0.57)
[–0.20, 2.07]
totalW
3.07 (1.71)
[–0.30, 6.43]
1.24 (1.44)
[–1.46, 4.22]
Between-Person
aB
–0.18 (0.30)
[–0.76, 0.44]
0.20 (0.15)
[–0.10, 0.49]
bB
0.70 (0.08)
[0.55, 0.85]
0.71 (0.08)
[0.56, 0.86]
c'B
–0.10 (0.20)
[–0.48, 0.28]
–0.06 (0.10)
[–0.26, 0.13]
indirectB
–0.12 (0.21)
[–0.54, 0.30]
0.14 (0.11)
[–0.06, 0.36]
totalB
–0.22 (0.28)
[–0.76, 0.36]
0.08 (0.14)
[–0.20, 0.36]
Positive Emotion
Within-Person
aW
12.07 (2.27)
[7.63, 16.45]
3.69 (1.40)
[0.85, 6.38]
bW
0.01 (0.02)
[–0.02, 0.05]
0.02 (0.02)
[–0.02, 0.05]
c'W
2.37 (1.46)
[–0.58, 5.20]
–0.60 (0.84)
[–2.22, 1.05]
cov(aW, bW)
0.18 (0.37)
[–0.57, 0.90]
–0.05 (0.22)
[–0.49, 0.37]
indirectW
0.32 (0.41)
[–0.48, 1.17]
0.01 (0.23)
[–0.46, 0.46]
totalW
2.69 (1.46)
[–0.15, 5.60]
–0.58 (0.85)
[–2.25, 1.06]
Between-Person
aB
–0.16 (0.30)
[–0.74, 0.44]
0.20 (0.15)
[–0.10, 0.49]
bB
–0.06 (0.09)
[–0.23, 0.12]
–0.07 (0.09)
[–0.26, 0.10]
c'B
0.37 (0.22)
[–0.07, 0.81]
0.04 (0.12)
[–0.19, 0.27]
indirectB
0.00 (0.03)
[–0.06, 0.09]
–0.01 (0.03)
[–0.07, 0.03]
totalB
0.38 (0.23)
[–0.05, 0.83]
0.02 (0.12)
[–0.21, 0.26]
EMOTIONAL IMPACT OF OBJECTIFICATION 7
Note. CI = Bayesian credibility interval (highest posterior density); Parameters in bold have 95% CIs that do not cross zero.
Estimates of all between-person paths (except path bB) were divided by 100 so that 1 unit reflects a difference of 1% in prevalence of
objectifying events.
Table S6
Results of multilevel Mediation Models Using Only Sample 2 Data
Predictor
Target
Witness
Outcome
Parameter
Estimate (SD)
95% CI
Estimate (SD)
95% CI
Negative Emotion
Within-Person
aW
9.83 (2.49)
[5.13, 14.90]
4.56 (1.41)
[1.82, 7.36]
bW
0.06 (0.01)
[0.03, 0.09]
0.07 (0.01)
[0.04, 0.09]
c'W
0.44 (1.51)
[–2.58, 3.39]
–0.16 (0.77)
[–1.62, 1.43]
cov(aW, bW)
–0.38 (0.30)
[–0.99, 0.17]
–0.10 (0.17)
[–0.45, 0.23]
indirectW
0.21 (0.36)
[–0.49, 0.92]
0.19 (0.19)
[–0.19, 0.58]
totalW
0.64 (1.52)
[–2.48, 3.51]
0.03 (0.78)
[–1.52, 1.53]
Between-Person
aB
1.64 (0.46)
[0.75, 2.55]
0.23 (0.10)
[0.04, 0.42]
bB
0.17 (0.07)
[0.03, 0.31]
0.21 (0.07)
[0.08, 0.34]
c'B
0.79 (0.31)
[0.16, 1.37]
0.16 (0.06)
[0.04, 0.27]
indirectB
0.27 (0.14)
[0.03, 0.56]
0.05 (0.03)
[0.00, 0.10]
totalB
1.07 (0.29)
[0.50, 1.63]
0.21 (0.06)
[0.10, 0.33]
Self-Conscious Emotion
Within-Person
aW
9.87 (2.54)
[4.90, 14.91]
4.44 (1.41)
[1.68, 7.18]
bW
0.09 (0.02)
[0.06, 0.13]
0.10 (0.02)
[0.06, 0.13]
c'W
1.54 (1.43)
[–1.35, 4.28]
–0.40 (0.84)
[–2.10, 1.22]
cov(aW, bW)
–0.31 (0.48)
[–1.26, 0.61]
–0.17 (0.23)
[–0.62, 0.30]
indirectW
0.61 (0.57)
[–0.53, 1.74]
0.26 (0.27)
[–0.28, 0.78]
totalW
2.13 (1.48)
[–0.74, 5.08]
–0.15 (0.85)
[–1.81, 1.53]
Between-Person
aB
1.65 (0.46)
[0.75, 2.54]
0.24 (0.10)
[0.05, 0.43]
bB
0.22 (0.05)
[0.12, 0.31]
0.23 (0.04)
[0.14, 0.32]
c'B
0.54 (0.22)
[0.11, 0.96]
0.15 (0.04)
[0.07, 0.22]
indirectB
0.35 (0.13)
[0.12, 0.61]
0.05 (0.03)
[0.01, 0.11]
totalB
0.89 (0.22)
[0.47, 1.33]
0.20 (0.04)
[0.11, 0.29]
Positive Emotion
Within-Person
aW
9.88 (2.49)
[5.01, 14.72]
4.64 (1.42)
[1.84, 7.42]
bW
0.05 (0.02)
[0.01, 0.10]
0.05 (0.02)
[0.00, 0.10]
c'W
–3.15 (2.31)
[–7.74, 1.38]
–0.51 (1.31)
[–3.07, 2.03]
cov(aW, bW)
–0.15 (0.51)
[–1.15, 0.87]
0.02 (0.31)
[–0.61, 0.63]
indirectW
0.35 (0.57)
[–0.80, 1.47]
0.26 (0.33)
[–0.40, 0.93]
totalW
–2.79 (2.31)
[–7.36, 1.80]
–0.26 (1.33)
[–2.92, 2.30]
Between-Person
aB
1.64 (0.46)
[0.75, 2.56]
0.24 (0.10)
[0.05, 0.43]
bB
0.04 (0.11)
[–0.17, 0.26]
0.09 (0.10)
[–0.10, 0.28]
c'B
–0.03 (0.48)
[–0.95, 0.91]
–0.18 (0.09)
[–0.35, –0.01]
indirectB
0.06 (0.19)
[–0.29, 0.47]
0.02 (0.03)
[–0.03, 0.08]
EMOTIONAL IMPACT OF OBJECTIFICATION 8
totalB
0.04 (0.43)
[–0.85, 0.84]
–0.16 (0.09)
[–0.32, 0.01]
Note. CI = Bayesian credibility interval (highest posterior density); Parameters in bold have 95% CIs that do not cross zero.
Estimates of all between-person paths (except path bB) were divided by 100 so that 1 unit reflects a difference of 1% in prevalence of
objectifying events.
Table S7
Results of multilevel Mediation Models Using Only Sample 3 Data
Predictor
Target
Witness
Outcome
Parameter
Estimate (SD)
95% CI
Estimate (SD)
95% CI
Negative Emotion
Within-Person
aW
11.29 (2.51)
[6.40, 16.2]
1.46 (1.29)
[–1.14, 3.90]
bW
0.03 (0.01)
[0.00, 0.06]
0.04 (0.02)
[0.01, 0.07]
c'W
2.75 (1.48)
[–0.21, 5.64]
–0.07 (0.67)
[–1.40, 1.20]
cov(aW, bW)
0.23 (0.30)
[–0.34, 0.85]
0.14 (0.17)
[–0.20, 0.48]
indirectW
0.58 (0.34)
[–0.09, 1.25]
0.19 (0.18)
[–0.17, 0.55]
totalW
3.34 (1.53)
[0.25, 6.31]
0.13 (0.68)
[–1.22, 1.44]
Between-Person
aB
1.65 (0.36)
[0.96, 2.38]
0.24 (0.13)
[–0.03, 0.49]
bB
0.27 (0.06)
[0.17, 0.38]
0.34 (0.05)
[0.24, 0.43]
c'B
0.52 (0.21)
[0.10, 0.93]
0.03 (0.06)
[–0.11, 0.15]
indirectB
0.44 (0.13)
[0.19, 0.71]
0.08 (0.05)
[–0.01, 0.17]
totalB
0.97 (0.21)
[0.56, 1.39]
0.10 (0.08)
[–0.04, 0.26]
Self-Conscious Emotion
Within-Person
aW
10.88 (2.54)
[5.84, 15.86]
1.34 (1.29)
[–1.21, 3.84]
bW
0.06 (0.01)
[0.03, 0.09]
0.07 (0.01)
[0.04, 0.10]
c'W
2.79 (1.59)
[–0.36, 5.85]
0.93 (0.76)
[–0.54, 2.40]
cov(aW, bW)
0.14 (0.28)
[–0.40, 0.69]
0.23 (0.14)
[–0.03, 0.53]
indirectW
0.80 (0.34)
[0.17, 1.52]
0.33 (0.18)
[–0.02, 0.67]
totalW
3.60 (1.63)
[0.38, 6.75]
1.25 (0.76)
[–0.23, 2.72]
Between-Person
aB
1.64 (0.36)
[0.92, 2.33]
0.23 (0.13)
[–0.03, 0.49]
bB
0.19 (0.04)
[0.10, 0.27]
0.28 (0.04)
[0.20, 0.36]
c'B
0.75 (0.16)
[0.45, 1.06]
0.05 (0.05)
[–0.06, 0.15]
indirectB
0.29 (0.09)
[0.13, 0.48]
0.06 (0.04)
[–0.01, 0.14]
totalB
1.05 (0.15)
[0.74, 1.35]
0.11 (0.06)
[–0.01, 0.24]
Positive Emotion
Within-Person
aW
11.31 (2.42)
[6.59, 16.11]
1.40 (1.31)
[–1.20, 3.98]
bW
0.03 (0.02)
[–0.01, 0.07]
0.03 (0.02)
[–0.01, 0.07]
c'W
1.57 (1.72)
[–1.78, 4.98]
1.13 (0.92)
[–0.70, 2.88]
cov(aW, bW)
–0.74 (0.49)
[–1.76, 0.19]
–0.09 (0.24)
[–0.55, 0.4]
indirectW
–0.38 (0.52)
[–1.46, 0.61]
–0.04 (0.25)
[–0.53, 0.44]
totalW
1.20 (1.76)
[–2.24, 4.70]
1.09 (0.93)
[–0.72, 2.92]
Between-Person
aB
1.66 (0.36)
[0.95, 2.38]
0.24 (0.13)
[–0.03, 0.49]
bB
–0.23 (0.10)
[–0.42, –0.05]
–0.25 (0.08)
[–0.42, –0.09]
EMOTIONAL IMPACT OF OBJECTIFICATION 9
c'B
–0.12 (0.36)
[–0.83, 0.59]
0.04 (0.11)
[–0.17, 0.26]
indirectB
–0.37 (0.18)
[–0.75, –0.04]
–0.05 (0.04)
[–0.15, 0.01]
totalB
–0.50 (0.33)
[–1.17, 0.12]
–0.02 (0.11)
[–0.24, 0.20]
Note. CI = Bayesian credibility interval (highest posterior density); Parameters in bold have 95% CIs that do not cross zero.
Estimates of all between-person paths (except path bB) were divided by 100 so that 1 unit reflects a difference of 1% in prevalence of
objectifying events.
Controlling for the effects of other stressors. To investigate whether the indirect
emotional impact of sexually objectifying events was independent of reactivity to other daily
stressors, we ran additional analyses using data from Samples 2 and 3, in which self-
objectification (mediator) and emotions (outcome) were simultaneously regressed onto
objectifying events and other daily stressors/hassles.
Figure S1 shows the within-person model controlling for reactivity to other daily
stressors/hassles (Stress). Here, both the proposed mediator (S-Obj) and outcome (Emotion)
are regressed onto the occurrence of stressors/hassles (Stress), while also modeling the effect
of objectifying events (Event). The random slopes S1Wi and S2Wi represent the effects of other
stressors/hassles on Self-Objectification and Emotion, respectively.
EMOTIONAL IMPACT OF OBJECTIFICATION 10
Figure S1. Within-person mediation model controlling for reactivity to other daily stressors/hassles. The
double-headed arrow connecting stress and event represents a contemporaneous covariance between
objectifying events and other daily stressors/hassles reported at the same EMA survey. Covariances between all
random slopes were estimated between-persons, but are not shown above for simplicity.
Table S8 displays estimates of within-person effects from the multilevel mediation
models controlling for reactivity to other daily stressors/hassles, using data from Samples 2
and 3 (see Figure S1 above for model diagram).
Time
EMOTIONAL IMPACT OF OBJECTIFICATION 11
Table S8
Results of multilevel mediation models testing effects of exposure to sexually objectifying events on emotions via self-
objectification, controlling for other daily stressors/hassles based on combined data from Samples 2 and 3
Predictor
Target
Witness
Outcome
Parameter
Estimate (SD)
95% CI
Estimate (SD)
95% CI
Negative Emotion
Within-Person
S1W
2.06 (0.93)
[0.19, 3.85]
2.40 (0.93)
[0.63, 4.27]
S2W
10.30 (0.83)
[8.69, 11.93]
10.31 (0.82)
[8.72, 11.94]
aW
10.35 (1.69)
[7.13, 13.80]
2.97 (0.93)
[1.14, 4.76]
bW
0.04 (0.01)
[0.02, 0.05]
0.04 (0.01)
[0.02, 0.06]
c'W
1.33 (0.97)
[–0.51, 3.29]
0.30 (0.46)
[–0.58, 1.24]
cov(aW, bW)
–0.05 (0.16)
[–0.36, 0.26]
0.07 (0.09)
[–0.11, 0.26]
indirectW
0.33 (0.19)
[–0.03, 0.70]
0.19 (0.10)
[0.00, 0.40]
totalW
1.67 (0.97)
[–0.25, 3.55]
0.49 (0.46)
[–0.39, 1.43]
Self-Conscious Emotion
Within–Person
S1W
2.23 (0.97)
[0.31, 4.09]
2.67 (0.95)
[0.80, 4.52]
S2W
4.92 (0.84)
[3.27, 6.58]
5.14 (0.85)
[3.49, 6.82]
aW
9.90 (1.69)
[6.65, 13.24]
2.87 (0.91)
[1.10, 4.64]
bW
0.07 (0.01)
[0.05, 0.09]
0.07 (0.01)
[0.05, 0.09]
c'W
2.18 (1.01)
[0.19, 4.18]
0.47 (0.54)
[–0.64, 1.51]
cov(aW, bW)
–0.03 (0.21)
[–0.44, 0.38]
0.14 (0.11)
[–0.08, 0.34]
indirectW
0.63 (0.25)
[0.15, 1.14]
0.34 (0.13)
[0.09, 0.59]
totalW
2.82 (1.03)
[0.83, 4.90]
0.81 (0.54)
[–0.27, 1.85]
Positive Emotion
Within–Person
S1W
2.28 (0.97)
[0.35, 4.16]
2.68 (0.97)
[0.80, 4.61]
S2W
–10.97 (1.05)
[–13.07, –8.98]
–10.96 (1.05)
[–13.01, –8.87]
aW
10.37 (1.65)
[7.21, 13.73]
3.11 (0.94)
[1.29, 4.97]
bW
0.05 (0.02)
[0.02, 0.08]
0.05 (0.02)
[0.02, 0.08]
c'W
0.04 (1.44)
[–2.82, 2.84]
–0.06 (0.73)
[–1.50, 1.38]
cov(aW, bW)
–0.32 (0.28)
[–0.91, 0.19]
–0.03 (0.16)
[–0.37, 0.28]
indirectW
0.17 (0.31)
[–0.46, 0.76]
0.11 (0.17)
[–0.23, 0.44]
totalW
0.20 (1.45)
[–2.76, 2.98]
0.05 (0.74)
[–1.39, 1.50]
Note. CI = Bayesian credibility interval (highest posterior density).
Parameters in bold have 95% CIs that do not cross zero.
S1W and S2W represent the within-person stressor reactivity slopes; i.e., the estimated effects of stressor occurrence on
changes in state self-objectification and emotions, respectively.
EMOTIONAL IMPACT OF OBJECTIFICATION 12
Temporal precedence of objectifying events. Table S9 displays results of
multilevel logistic regression models testing the within-person effect of self-objectification
on subsequent reporting of sexually objectifying events. In each model, the probability of
reporting a sexually objectifying event (as Target or Witness) between time t1 and t2 was
predicted by level Self-Objectification during the same time interval (t1 to t2) or in the
previous time interval (t0 to t1), while controlling for the occurrence of objectifying events in
the previous interval (t0 to t1). Results in Table S9 indicate that neither the contemporaneous
nor cross-lagged effects of self-objectification on reporting of objectifying events was
meaningfully different from zero.
Table S9
Results of multilevel models testing within-person effect of self-objectification as a
predictor of exposure to sexually objectifying events
Outcome
Predictor
Estimate (SD)
95% CI
Targett1-t2
S-Objt1-t2
0.006 (0.005)
[–0.004, 0.016]
Targett0-t1
–0.268 (0.284)
[–0.850, 0.271]
Targett1-t2
S-Objt0-t1
–0.006 (0.005)
[–0.016, 0.005]
Targett0-t1
–0.130 (0.228)
[–0.612, 0.300]
Witnesst1-t2
S-Objt1-t2
–0.004 (0.005)
[–0.014, 0.005]
Witnesst0-t1
0.291 (0.086)
[0.120, 0.459]
Witnesst1-t2
S-Objt0-t1
–0.006 (0.005)
[–0.015, 0.004]
Witnesst0-t1
0.286 (0.092)
[0.101, 0.461]
“Doubly-lagged” mediation models. Although the analyses reported in Table S9
(above) do not suggest that increases in self-objectification predict subsequently increased
reporting of sexually objectifying events, we sought to conclusively test our theorized causal
sequence of objectifying events → self-objectification → emotion by repeating our main
analyses with objectifying events assessed at the previous measurement occasion to self-
objectification. Figure S2 displays the “doubly-lagged” within-person mediation model we
EMOTIONAL IMPACT OF OBJECTIFICATION 13
tested, in which objectifying events occurring between t0 and t1 were used to predict self-
objectification in the interval between t1 and t2, which in turn predicted emotions at t2.
Figure S2. Doubly-lagged within-person mediation model tested. In this model, the predictor (Eventt0-t1) was
measured at the previous occasion to the mediator (S-Objt1-t2). As in the main analyses, the outcome (Emotiont2)
and mediator (S-Objt1-t2) were measured at the same occasion (t2), but since the mediator was reported as “since
the last survey” it is assumed to precede the outcome, which was measured as “right now”. Covariances
between all random slopes were estimated between-persons, but are not shown above for simplicity.
Time
EMOTIONAL IMPACT OF OBJECTIFICATION 14
Estimates of the within-person effects from the doubly-lagged mediation models
shown in Figure S2 are displayed in Table S10, below.
Table S10
Results of “doubly-lagged” multilevel mediation models
Predictor
Target
Witness
Outcome
Parameter
Estimate (SD)
95% CI
Estimate (SD)
95% CI
Negative Emotion
Within-Person
aW
–2.58 (1.09)
[–4.74, –0.45]
–1.70 (0.80)
[–3.23, –0.10]
bW
0.05 (0.01)
[0.04, 0.06]
0.05 (0.01)
[0.03, 0.06]
c'W
–0.17 (0.69)
[–1.56, 1.14]
–0.68 (0.46)
[–1.59, 0.19]
cov(aW, bW)
0.16 (0.10)
[–0.04, 0.37]
–0.06 (0.08)
[–0.22, 0.09]
indirectW
0.03 (0.11)
[–0.19, 0.26]
–0.14 (0.09)
[–0.31, 0.04]
totalW
–0.14 (0.7)
[–1.51, 1.25]
–0.82 (0.48)
[–1.79, 0.07]
Self-Conscious Emotion
Within-Person
aW
–2.31 (1.13)
[–4.64, –0.23]
–1.49 (0.82)
[–3.13, 0.07]
bW
0.16 (0.01)
[0.14, 0.19]
0.16 (0.01)
[0.13, 0.18]
c'W
0.17 (1.10)
[–2.16, 2.15]
–0.57 (0.56)
[–1.70, 0.51]
cov(aW, bW)
0.16 (0.21)
[–0.25, 0.58]
–0.11 (0.15)
[–0.42, 0.19]
indirectW
–0.21 (0.28)
[–0.78, 0.33]
–0.34 (0.19)
[–0.70, 0.04]
totalW
–0.04 (1.09)
[–2.20, 2.07]
–0.90 (0.64)
[–2.16, 0.36]
Positive Emotion
Within-Person
aW
–2.05 (1.10)
[–4.18, 0.07]
–1.46 (0.81)
[–3.02, 0.15]
bW
0.03 (0.01)
[0.01, 0.05]
0.03 (0.01)
[0.01, 0.05]
c'W
1.53 (1.02)
[–0.41, 3.55]
0.45 (0.55)
[–0.64, 1.52]
cov(aW, bW)
–0.42 (0.17)
[–0.77, –0.12]
–0.09 (0.13)
[–0.34, 0.15]
indirectW
–0.48 (0.17)
[–0.85, –0.18]
–0.14 (0.13)
[–0.38, 0.12]
totalW
1.03 (1.02)
[–0.95, 3.04]
0.32 (0.56)
[–0.82, 1.37]
Note. CI = Bayesian credibility interval (highest posterior density); Parameters in bold have 95% CIs that do not cross zero.
Estimates of within-person indirect effects in Table S10 were not consistent with our
main analyses, and trended in a negative direction (with a 95% CI for the indirect effect of
Target on Positive Emotion not crossing zero). This is due to the fact that the aW paths
became negative when regressing self-objectification onto lagged objectifying events,
suggesting that exposure to objectifying events may predict decreases in self-objectification
EMOTIONAL IMPACT OF OBJECTIFICATION 15
at the subsequent time-interval. One plausible explanation for these findings, which go in the
opposite direction to estimates of aW paths in our main analyses (see Table 3 in main text), is
that self-objectification may increase momentarily in the period immediately following
exposure to sexually objectifying events, but may subsequently decrease due to self-
regulatory processes. Such momentary increases may be too brief to capture as lagged effects
when successive measurement occasions (i.e., EMA surveys) are approximately 60-90 min
apart, as in the current study. When controlling for self-objectification during the same time-
interval as exposure to objectifying events (see model diagram in Figure S2), objectifying
events may predict decreases in self-objectification in subsequent time-intervals as
momentary increases in self-objectification tend to “return to baseline”. In line with this
reasoning, when we repeated the analyses without controlling for self-objectification reported
between t0 and t1, exposure to events was again positively (although less strongly) associated
with self-objectification in the following time interval (t1 to t2). This is actually consistent
with the results of our main analyses (see Table 3 in main text), wherein the
contemporaneous aW path was positive and the autoregressive path for self-objectification
was also positive, implying an overall positive relationship among past objectifying events
and future self-objectification (in our main analyses this is an indirect path).
