Women's body dissatisfaction, physical appearance comparisons, and Instagram use throughout the COVID‐19 pandemic: A longitudinal study

Abstract Objective This study aimed to determine the evolution of Instagram use, body dissatisfaction and physical appearance comparisons throughout the coronavirus disease‐2019 (COVID‐19) pandemic, and to explore whether there was a relationship between the changes in Instagram use throughout the pandemic and body dissatisfaction and physical appearance comparisons. Method A total of 272 Spanish women (16–70 years old) were followed‐up across four waves of assessment between November 2019 (before the pandemic started) and July 2021. Body dissatisfaction, social appearance comparisons, and Instagram use were assessed using the Eating Disorders Inventory‐3, the Physical Appearance Comparison Scale‐Revised, and an ad hoc questionnaire for Instagram use, respectively. Results No statistically significant changes were found in the frequency of Instagram use, nor on the proportion of women following appearance‐focused accounts on Instagram, among the data collection periods. Body dissatisfaction significantly increased from T1 to T4, and physical appearance comparisons significantly increased from T1 to T2, T3, and T4. These increases were not found to be significant for those with eating disorder risk. No significant differences were found in body dissatisfaction and physical appearance comparisons depending on whether participants' frequency of Instagram use had changed or remained the same, or whether they had started/stopped/continued following appearance‐focused accounts on Instagram during the pandemic. Discussion Women's body dissatisfaction and physical appearance comparisons seem to have increased throughout the pandemic. The experiences of individuals with eating disorder risk throughout the pandemic, and the relationship between the pandemic and Instagram use, might be complex and need further research. Public Significance This study suggests that women's body dissatisfaction and physical appearance comparisons have increased throughout the COVID‐19 pandemic. However, this increase might not be as clear for those who had eating disorder risk before the pandemic. Instagram frequency of use, and the percentage of women following appearance‐focused accounts on Instagram, do not seem to have significantly increased. More research is needed to explore the impact of the pandemic.


| INTRODUCTION
The emergence of the highly infectious coronavirus disease-2019  has posed a significant threat to global health. The rapid spread of the virus across the world resulted in several governments implementing drastic health measures, including lockdowns, physical distancing, closure of nonessential services, and travel restrictions (Castex et al., 2021).
In Spain, on March 14, 2020, a state of emergency was declared, placing the entire country in lockdown. All nonessential work activity was suspended, and the population was mandated to remain at home.
After the initial 6 weeks, restrictions were gradually lifted. However, some health and social measures lasted for more than 1 year or are still in place at the time of writing this article (e.g., wearing a face mask in some indoor spaces and public transport). Six waves of coronavirus infections have been reported in Spain since the beginning of the pandemic until May 2022.
Since the outbreak of the pandemic, several researchers and clinicians have expressed concern about its impact on mental health (World Health Organization Regional Office for Europe, 2022). It has been suggested that the pandemic increased the prevalence of mental health problems such as depression, anxiety, distress, and insomnia (Jin et al., 2021;Wu et al., 2021).
Eating habits, appearance concerns, eating disorder risk, and disordered eating might have also worsened with the pandemic (e.g., Linardon et al., 2022;Robertson et al., 2021;Touyz et al., 2020;Vall-Roqué et al., 2021a;Weissman et al., 2020). According to a recent systematic review (Schneider et al., 2022), several studies show a negative influence of the pandemic on body image and disordered eating, but conversely, other studies report positive outcomes of the COVID-19 pandemic, including reduction in eating disorder symptomatology, more time to reflect on recovery and engage in self-care, greater motivation to recover, and more time to connect with family.
These positive outcomes might be related with the socioeconomic status of participants, reflecting that those with higher social privilege have incurred fewer financial pressures, and this might have facilitated engagement with self-care, recovery strategies, and social support (Schneider et al., 2022).
According to Rodgers et al. (2020), a pathway by which the current pandemic could increase eating disorders risk and symptoms is through an increased consumption of media (particularly social media) due to social distancing measures. This would happen through increased exposure to harmful eating and appearance-related content, as well as more general stressful or traumatic world events. In this regard, in a recent study, it was found that there was a significant increase in the frequency of use of social network sites (SNS) and in the number of women following appearance-focused accounts on Instagram during lockdown (Vall-Roqué et al., 2021b). Taking into consideration that the use of social networks, especially an appearance-focused use, is linked to body dissatisfaction (Fardouly & Vartanian, 2016;Holland & Tiggemann, 2016;Sherlock & Wagstaff, 2019), an increase in the (appearance-focused) use of SNS during the pandemic could have led to an increase in body dissatisfaction. Furthermore, according to the tripartite influence model (Thompson et al., 1999), there are three primary influence variables that contribute to the development of body image disturbances: peers, parents, and media. In this sense, it has been stated that the tendency to engage in physical appearance comparisons plays a mediating role in the link between SNS use and body dissatisfaction (Fioravanti et al., 2022). Therefore, an increase in SNS use throughout the pandemic could be related to an increase in (media-related) appearance comparisons. Also, considering that during the lockdown period people were presumably not engaging in socializing and could not compare themselves in-person with their peers, they might have increased their comparisons on social media, and it is known that social media comparisons tend to be focused on the beauty ideal and can be particularly harmful (Fardouly et al., 2017), which might have led to increased body dissatisfaction. However, it could also be argued that, as during lockdown restrictions it was not allowed to engage in activities that often facilitate the appearance ideal social media selfpresentation, the appearance focused use of SNS might have been tempered.
It should be noted that the vast majority of research that has been published examining the impact of the pandemic on body imagerelated variables or SNS use in community samples has used crosssectional designs, and there is limited research using experimental and longitudinal designs. The findings of the few studies that have longitudinally assessed the impact of the pandemic on body image are inconsistent: while there are studies that have found the pandemic to be associated with increased concerns about weight, shape and eating, and increased eating disorder symptomatology and screen time (e.g., Keel et al., 2020;Trott et al., 2021), others have found no differences between pre-and during/post-lockdown in eating disorder symptoms and eating disorder risk (e.g., Koenig et al., 2021;Martínezde-Quel et al., 2021). Regarding the studies that have longitudinally examined social media use over the pandemic, an increase in SNS use VALL-ROQU E ET AL. during the lockdown phase of the pandemic has been consistently reported (Marciano et al., 2022). For example, Fumagalli et al. (2021) indicated that social media use increased at the beginning of the lockdown (March-April 2020), and Arend et al. (2021) found that more than 40% of their participants increased their daily time spent using social media. However, to our knowledge, it has not been longitudinally assessed whether the pandemic has been associated with an appearance-focused use of SNS.
Furthermore, to our knowledge, the possible changes in physical appearance comparisons throughout the pandemic have not been studied, and no studies have longitudinally assessed the (appearancefocused) use of Instagram throughout the pandemic. Instagram is one of the most popular social media platforms worldwide, and recent research attention has turned specifically to this social network, as it is dedicated purely to the posting and sharing of photos, and it has been linked to body dissatisfaction (Fardouly & Vartanian, 2016;Horwitz, 2021;Tiggemann & Anderberg, 2020). When the first data collection of this study was conducted, Instagram was the most widely used photo-based social network in Spain (IAB Spain, 2019, 2020. Finally, only a few studies have focused on women samples. Taking into consideration that body image problems are more prevalent among women than men (Myers & Crowther, 2009;O'Dea & Caputi, 2001) and considering that women have been reported to be an especially vulnerable group to the negative psychosocial effects of COVID-19 (Ozamiz-Etxebarria et al., 2020;Sun et al., 2020;Wang et al., 2020), special attention should be paid to this population group.
This is a four-wave longitudinal study of data collection on Instagram use, body dissatisfaction and appearance comparisons in a Spanish women population. Participation was requested at four different times: the first one was some weeks before the pandemic outbreak, and the other ones took place approximately every 6 months throughout the pandemic. The objectives of this study were: (1) to determine the change in Instagram use, body dissatisfaction and physical appearance comparisons throughout the pandemic (both in the whole women sample and in two subgroups: those at risk of having an eating disorder and those without eating disorder risk), and (2) to explore whether there was a relationship between the changes in Instagram use throughout the pandemic and body dissatisfaction and physical appearance comparison tendency.  F I G U R E 1 Flowchart of the participation in each data collection period religious sites, hotels, and terraces started to reopen under strict safety conditions. Phase 2 allowed theaters, museums, indoor restaurants, and bars to open with limited capacity. In Phase 3, all shops were back in business (with a maximum capacity of 50%), and meetings and conferences were allowed again (La Moncloa, 2020).
In the third and fourth data collection periods (T3 and T4), no complete lockdown was in place. However, in T3, the third wave of COVID-19 was in full swing in Spain, and several restrictions were introduced: although these differed depending on the region of Spain, they included a night-time curfew, a reduced maximum occupancy for restaurants, businesses and gatherings, and mobility restrictions. These restrictions caused great controversy because they coincided with New Year's festive season and therefore interfered with family gatherings and other celebrations. In T4, restrictions were relaxed but still included a maximum capacity for gatherings and limitations in the capacity and opening hours of retail, among others.
Inclusion criteria for participating in the initial study (which was conducted before the pandemic started) included living in Spain, being between 16 and 70 years old, and being a woman. Participants were recruited through various social media platforms (Twitter, Instagram, and Facebook) and through personal contacts of the research team. All

| Sociodemographic data
Information on age, gender, place of birth, place of residence, employment situation, average gross annual income, marital status, household members, height, and weight was systematically collected.
Height and weight measures were used to calculate body mass index (BMI). Gender and educational level were only asked at T1.

| COVID-19 and lockdown-related information
In the second, third, and fourth waves, participants indicated whether themselves or a close person had contracted COVID-19, and if they had a loved one that had deceased due to COVID-19. This information was assessed using closed-ended questions.

| Eating Disorders
Inventory-3 (Garner, 2004, adapted to Spanish by Elosua et al., 2010) The Body Dissatisfaction subscale was used. This subscale consists of 10 items that assess discontentment with the overall shape and with the size of those regions of the body of extraordinary concern to those with eating disorders. Responses are rated on a 6-point Likert scale, ranging from "Never" to "Always" (scores range: 0-40). The scale has an adequate internal consistency, with reported alpha values of .90 in Spanish women (Elosua et al., 2010). In this study, the alpha value for the Body Dissatisfaction subscale was .89 at all measurement waves.
F I G U R E 2 Time periods of data collection within the epidemic curve of the pandemic in Spain. Data extracted from https://cnecovid.isciii.es/ covid19/#ccaa 2.2.4 | Physical Appearance Comparison Scale-Revised (Schaefer & Thompson, 2014;adapted to Spanish by Vall-Roqué et al., 2022) The PACS-R is an 11-item questionnaire that measures the tendency to compare one's physical appearance to others in a broad array of social settings. Responses are rated on a 5-point Likert-type scale ranging from "Never" to "Always." Higher scores indicate higher levels of general appearance comparison (scores range: 0-4). Psychometric evaluations of the Spanish version of the PACS-R indicate that the questionnaire has a single factor structure, and excellent internal consistency (Cronbach's alpha of .97) and convergent validity in Spanish women (Vall-Roqué et al., 2022). In this study, the alpha values for the PACS-R were .97, .98, .97, and .97 at baseline, T2, T3, and T4, respectively.
2.2.5 | Eating Attitudes Test-26 (Garner et al., 1982; adapted to Spanish by Castro et al., 1991) The Eating Attitudes Test (EAT-26) is a 26-item self-report questionnaire that measures disordered eating behavior and attitudes, and it is often used to identify individuals who might be at risk for an eating disorder. Items are presented in a 6-point Likert scale ranging from "Never" to "Always," and higher scores indicate higher levels of eating disturbances. It has three subscales: dieting, bulimia and food preoccupation, and oral control. The questionnaire has been reported to have adequate psychometric properties, and the alpha reliability coefficient in a Spanish sample was .93 (Castro et al., 1991). In this research, the alpha values for the .92,.93,and .94 at baseline, T2, T3, and T4, respectively.
The EAT-26 was only used to divide the sample between those at risk of having an eating disorder (total score ≥ 20) and those without eating disorder risk (total score < 20).

| Instagram use
A Likert-scale self-report item assessed the frequency of Instagram use in each time point (I do not have an Instagram account, Never/ Almost never, Less than 1 h/day, 1-2 h/day, 2-3 h/day, 3-4 h/day,

More than 4 h/day).
A multiple answer question queried which type of accounts individuals followed on Instagram (only applicable to those participants who reported having an Instagram account) in each wave of data collection. Participants who indicated that they followed fashion, clothing brands, weight-loss tips or diets, beauty, or fitness accounts were included in the "following appearance-focused accounts" group, whereas those who did not follow any of the mentioned accounts were included in the "not following appearance-focused accounts" group.
Moreover, three groups of participants were created according to whether their frequency of Instagram use had increased, remained the same or decreased during the pandemic (from T1 to T2, from T1 to T3, and from T1 to T4). Similarly, participants were categorized into three different groups depending on whether they had started following appearance-focused accounts on Instagram during the pandemic, they had stopped following appearance-focused accounts, or they had continued following or not following appearance-focused accounts. Cochran's Q tests were used to determine if there were differences on dichotomous variables between three or more related groups.

| Data analysis
Friedman tests were used to examine if there were differences between groups when the variable being measured was ordinal.
Marginal homogeneity tests were used to determine if there were differences on categorical nondichotomous variables between related groups.
For the categorical variables "frequency of Instagram use" and "following appearance-focused Instagram accounts," the Cochran's Q test was used to explore whether there were significant differences among the four time points. Significant results were followed by Bonferroni adjustment.
Student's t tests were conducted to explore whether there were significant differences in body dissatisfaction and physical appearance comparisons at baseline between individuals with and without eating disorder risk. To assess the differences among time points in the psychological variables, mixed regression models with repeated measures were used, which included the psychological variables  categorical variable to allow the model to fit every possible pattern in time and not assume linear growth, considering that a disadvantage that has been reported of modeling time as a continuous variable is that the development over time of the outcome variable Y is modeled as a linear function (Twisk, 2013).
One-way ANOVAs were carried out to assess whether the changes in the frequency of Instagram use (increase, decrease, or no changes) and the changes in following appearance-focused accounts on Instagram throughout the pandemic were significantly associated with body dissatisfaction and appearance comparison tendency. If Levene's test of homogeneity of variance was violated, the Brown-Forsyth statistic was used.

| Body dissatisfaction and physical appearance comparisons
Body dissatisfaction and physical appearance comparison scores are presented in Table 3 for each time point, both for the whole sample and for individuals with and without risk of having an eating disorder (according to their EAT-26 score). The mean baseline scores of body dissatisfaction and physical appearance comparisons for those at risk of having an eating disorder were significantly higher (with high effect sizes) than the scores of those without eating disorder risk (body dis- To assess the differences in body dissatisfaction and physical appearance comparisons among the different time points, mixed regression models with repeated measures were used. In the regression model for body dissatisfaction (Table 4), body dissatisfaction significantly increased from T1 to T4 (Β = 1.76; p < .001). The interaction between time and eating disorder risk was statistically significant at T3 (p = .028) and T4 (p = .017). BMI was significantly associated with changes in body dissatisfaction (p < .001): body dissatisfaction score increased .79 for each BMI unit. The educational level was not found to significantly predict body dissatisfaction. T A B L E 5 Differences in body dissatisfaction and physical appearance comparisons according to participants' changes in Instagram use In the subgroups analysis, it was found that the sample that was not at risk of having an eating disorder significantly increased their body dissatisfaction at T4 (Β = 1.75; p < .001) (with respect to T1), and with a trend toward significance at T2 (p = .089) and T3 (p = .075). However, no significant changes in body dissatisfaction among time points were found in the sample at risk of having an eating disorder. In both subgroups, BMI was found to be a statistically significant predictor of body dissatisfaction.
Regarding physical appearance comparisons (Table 4), they significantly increased from T1 to T2 (B = .23; p < .001), T3 (B = .17; p = .009), and T4 (Β = .24; p = .001). The interaction between time and eating disorder risk showed a trend toward significance at T3 (p = .093) and T4 (p = .053). The effect of eating disorder risk was significant, with a score in appearance comparisons 1.72 times higher in women with eating disorder risk versus those without eating disorder risk (p < .001).
In the subgroup of women without eating disorder risk, physical appearance comparisons significantly increased at T2 (Β = .23; p < .001), T3 (Β = .17; p = .009), and T4 (Β = .24; p < .001) with respect to T1. In contrast, no significant differences were found in appearance comparisons among time points in the group of women with risk of having an eating disorder.

| Relationship between the changes in Instagram use throughout the pandemic and body dissatisfaction and physical appearance comparisons
As shown in Table 5, no significant differences were found in body dissatisfaction and physical appearance comparisons at T4 depending on whether participants' Instagram frequency of use had increased, remained the same or decreased from T1 to T4. The same pattern was observed when the differences between T1-T2 and T1-T3 were assessed.
Similarly, no significant differences were found in body dissatisfaction and physical appearance comparisons at T4 depending on whether participants had started following appearance-focused accounts on Instagram during the pandemic, had stopped following appearance-focused accounts, or had continued following or not following appearance-centered Instagram accounts throughout the different data collection periods. The same was observed when the differences between T1-T2 and T1-T3 were assessed (Table 5).

| DISCUSSION
The general aim of this study was to explore the changes in body dissatisfaction, physical appearance comparisons, and Instagram use throughout the pandemic in a community sample of Spanish women, and to determine whether there was an association between the changes in Instagram use throughout the pandemic and body dissatisfaction and physical appearance comparisons.
Regarding the characteristics of this study's sample, we found that women's BMI increased throughout the pandemic, although the effect size of this significant increase was very small and should therefore be interpreted with caution. The literature is inconsistent in this regard: while some studies have reported an increase in BMI throughout the pandemic, others have found no changes or a decrease in women's BMI (Bakaloudi et al., 2021). BMI and weight trajectories during COVID-19 are probably dependent on many factors that should be considered, such as age, socioeconomic status, pandemic living and working conditions, diet, physical activity, or alcohol intake (Khan et al., 2022). Moreover, according to our regression model, a higher BMI was found to be associated with increased body dissatis-  (Fernandes et al., 2020;Vall-Roqué et al., 2021b), and this might be partly due to the longitudinal design used in this research, which included pre-pandemic data, versus the cross-sectional designs relying on retrospective reports used in the other studies. However, it should be considered that, although our results did not yield statistical significance, the percentage of women that did not have an Instagram account decreased from 16% to 8% from baseline (before the pandemic outbreak) to the last data collection period (almost a year and a half after the start of the pandemic), and the percentage of women who used Instagram more than 4 h per day raised from 1.8% to 3.8%. Similarly, in each data collection period, the percentage of women that followed appearance-focused accounts on Instagram increased. Therefore, it is possible that the pandemic is associated with changes in Instagram use, but these changes might be more complex (and less evident) than they seem in cross-sectional studies.
The pandemic seems to have had an effect on body dissatisfaction and physical appearance comparisons. Regarding body dissatisfaction, although its increase was not linear throughout the pandemic, there was a significant raise in body dissatisfaction from baseline to our last data collection period. This aligns with most of the published literature, which states the negative impacts of the COVID-19 pandemic on appearance, shape, and weight concerns (Schneider et al., 2022). Similarly, there was a significant increase in physical appearance comparisons from baseline to the different data collection periods. To our knowledge, this is the first study to longitudinally assess the changes on appearance comparisons throughout the VALL-ROQU E ET AL.
pandemic. Considering that the tendency to engage in appearance comparisons has been reported to be related to body dissatisfaction and eating disturbances (Thompson et al., 1999) According to Robertson et al. (2021) and Schneider et al. (2022), potential explanations or factors associated with changes in body image disturbances could include higher levels of stress, worry and anxiety (for example, resulting from increased caregiving responsibilities), increased levels of rumination, loneliness and depression, fear of COVID-19, exposure to increased weight stigma via public health and social media messaging regarding obesity and COVID-19, increased saliency of food and eating as a result of shopping restrictions and changes to daily routines, household arguments or family conflicts, and changes in living situation and access to usual support networks.
Another potential explanation could be related to the fact that, as the population was suddenly mandated to remain at home for longer periods of time, individuals replaced their "unfiltered" in-person comparisons with online comparisons with influencers that promoted the beauty ideal. These changes in body dissatisfaction and appearance comparisons in community women highlight the importance of implementing body image and eating disorders prevention programs and initiatives (e.g., media literacy programs).
However, an increase or decrease in Instagram use throughout the pandemic, or the fact of starting to follow or unfollow appearance-focused accounts on Instagram, was not associated with higher or lower levels of body dissatisfaction or appearance comparisons. This suggests that the changes in Instagram use during the pandemic might not be necessarily associated with body image problems. Staniewski and Awruk (2022) recently reported in their study a positive impact of Instagram on mental well-being during the pandemic, while other authors have stated that social media exposure during the pandemic may predict disordered eating symptoms (Bellapigna et al., 2021). These contradictory findings might be due to several reasons, including the different measures used in each study and the potential impact of confounding variables. Also, some authors have suggested that the specific social media activities (e.g., greater investment in "selfie" activities), rather than the total time spent using it, predicts body dissatisfaction (Cohen et al., 2018). In this line, as mentioned before, appearance comparisons on SNS (instead of changes in the frequency of use or in the types of accounts followed) might better explain increased appearance comparisons and body dissatisfaction. Hence, future research could focus on the effect of specific online activities performed during the pandemic and their impact on body image-related variables.

| LIMITATIONS AND FUTURE DIRECTIONS
To our knowledge, this is the first study that longitudinally assesses the changes in body dissatisfaction, physical appearance comparisons, and Instagram use throughout the COVID-19 pandemic using prepandemic data. Its limitations lead to several future directions. First, the sample size of this study was small, and there was a high attrition rate. Second, this study included a community sample and, although there were some women at risk of having an eating disorder, future research should examine samples with eating disorders. Third, although women are more susceptible to appearance comparisons and body dissatisfaction (Myers & Crowther, 2009), future studies should also study men and individuals who do not identify within the gender binary system. Fourth, other SNS should be assessed, such as TikTok, which is growing rapidly and has been associated with eating disorders-related content during the pandemic (Llad o Jordan et al., 2021). Fifth, we employed a survey measurement to assess participants' perceived Instagram use, which might be less accurate than insitu measurements (Naab et al., 2018). Future studies could use more objective digital trace data sources to measure SNS use. Sixth, the measure that we used to assess physical appearance comparisons poses two potential limitations: (1) although it examines comparisons in several social contexts, it does not specifically assess comparisons on social media; future research could include specific items about comparisons on social media, and (2) some items of this questionnaire refer to social situations that were prohibited during the strict phase of lockdown (e.g., eating in a restaurant or going shopping). Although the scale asks respondents to answer items thinking in what they