Can exposure to sexual objectification impact policy attitudes? Evidence from two survey experiments

Abstract

Research in social psychology has long argued that exposure to objectifying portrayals of women can lead to increasingly misogynist attitudes and behavior. We argue that such images can also impact on gendered policy attitudes. We suggest that objectifying images prime sexist attitudes and reduce perceptions of women’s agency, warmth, and competence. We argue that this may translate into decreased support for reproductive rights and other gender-salient policies. Furthermore, these effects may vary by the gender of those exposed to these images. In two survey experiments with brief exposures to objectifying images, we find mixed support for these predictions. Although we find some negative effects as predicted, we also find positive effects of objectification among women in the sample that are suggestive of a backlash effect. We discuss potential explanations for this heterogeneity. Overall, our results suggest interesting avenues to further explore the effects of objectification on political outcomes.

Research in psychology has demonstrated that the female body is objectified more than the male body (Fredrickson and Roberts, 1997). Objectification can be a form of dehumanization, and it occurs when people’s bodies or body parts are separated from their identity. Sexual objectification involves a cultural prioritization of a person’s sexual appeal over other attributes (Ward et al., 2023). Cognitive neuroscience investigations into objectification have found that it can disrupt social cognition processes typically associated with human targets (Cikara et al., 2011). Experimental social psychology studies have demonstrated that objectifying portrayals of women are linked to increased acceptance of rape myths and violence against women (Aubrey et al., 2011), subscription to masculine gender norms (Galdi et al., 2014), and increased physical aggression toward the objectified female target (Vasquez et al., 2018). Political scientists have only recently begun to explore the political implications of the dehumanization and objectification of women, immigrants, marginalized racial and ethnic groups, and political opponents (Cassese, 2021; Gothreau et al., 2023; Utych, 2018). In this article, we consider the potential for exposure to sexually objectifying and dehumanizing portrayals of women to impact gender-salient policy attitudes.

Often objectifying imagery shows women’s body parts as interchangeable with objects or disembodied entirely. Viewing sexually objectifying images of women activates cognitive processes associated with objects opposed to cognitive processes typically reserved for thinking about humans (Bernard et al., 2012). In other words, objectifying images can prime people to think of women as “things.” Feminist theorists have long discussed the way in which the objectification of women’s bodies is inherently political. However, objectification may also have political consequences in the way that positivist scholars think of the term “political.” In this research note, we present data from two experimental studies in which participants are exposed to either neutral or women-objectifying imagery. We then measure attitudes about gender-salient policy attitudes like support for abortion, birth control, policies that give women preferential hiring, Planned Parenthood, more expansive childcare policies, and free period products.

The connection between objectification and political attitudes

Objectification Theory serves as a framework for understanding the impacts of pervasive objectification in our culture, particularly in media, advertisements, pornography, and increasingly, on social media (Feltman and Szymanski, 2018). Work in philosophy has identified several different ways in which individuals can be objectified that include the denial of autonomy, fungibility (being treated as interchangeable), ignoring an individual’s thoughts and experiences, etc. (Nussbaum, 1995). Objectification can be non-sexual. For example, children, the elderly, and individuals with physical and mental disabilities are often objectified via the denial of their autonomy and agency. However, sexual objectification occurs primarily for women and adolescent girls (Stankiewicz and Rosselli, 2008). Sexual objectification can occur at the interpersonal level, such as when women receive objectifying comments or cat calls, and in the form of objectifying representations of women in film, television, and on social media (for a review of the sources and consequences of sexual objectification see Ward et al., 2023).

The prevalence of these depictions has been documented across many different types of media (Feltman and Szymanski, 2018; McDade-Montez et al., 2017). In experimental studies in which exposure to a sexually objectified target is manipulated, objectification has been associated with less moral concern for the objectified target (Loughnan et al., 2013), increased adversarial sexual beliefs, and the acceptance of interpersonal violence and rape myths (Aubrey et al., 2011), and even decreased intentions to support ethical organizations when the advertisement included sexualized images of women (Bongiorno et al., 2013). Observational evidence supports these experimental findings (Daniels and Zurbriggen, 2016; Fox and Potocki, 2016; Wright and Funk, 2014; Wright and Tokunaga, 2016). Another effect of pervasive objectification is self-objectification. When women self-objectify, they internalize objectification which can negatively impact women’s mental health, cognitive functioning, self-efficacy, and even political efficacy and interest (Gothreau, 2021; Quinn et al., 2006). Work by Calogero (2013) and Calogero et al. (2017) suggests that self-objectification can bolster support for the gender status quo and gender-specific system justification.

Research suggests that objectification is linked to perceptions of humanness and mental state attribution. Mental state attribution is our ability to use cues to infer other people’s intentions, beliefs, and desires, and it is a sign that a perceiver sees another as fully human (Waytz et al., 2010). Cikara et al. (2011) find a negative correlation between hostile sexism and activation of regions of the brain associated with mental state attribution when perceivers are viewing sexualized women. Diminished perceptions of mental state attribution have been observed in studies looking at other dehumanized groups such as homeless people and drug addicts (Harris and Fiske, 2006). ¹ Furthermore, the objectification of women is linked to reduced perceptions of women’s competence, warmth, and morality, essential humanizing traits in social cognition according to models of social judgment and social perception (Abele and Wojciszke, 2007; Fiske et al., 2002; Heflick et al., 2011). In experimental studies in which photos of girls and women were manipulated to be either sexually objectifying or not, objectified targets were evaluated as less intelligent, determined, capable, credible, and having lesser moral status (Funk and Coker, 2016; Graff et al., 2012). Overall, work on dehumanization, objectification, and social cognition all find that sexually objectified targets are denied some of the most important attributes related to perceiving others as fully human (Heflick and Goldenberg, 2009).

Do the effects of exposure to the objectification of women extend to politics? We have evidence that objectification not only impacts attitudes regarding the target of the objectification but can also impact other attitudinal outcomes (Bongiorno et al., 2013). Given the sexist, misogynistic, and gender-specific system-justifying attitudes that objectification primes, as well as the overall dehumanization of women that occurs when they are the objectified target, we posit that objectification might also impact gender-salient policy attitudes. One of the most gender-salient policy areas is reproductive rights and more specifically, access to abortion and birth control. Limits on access to these are often viewed as essential to maintaining the status of traditional gender roles (Jelen, 2015), and unsurprisingly, sexist attitudes predict less support for reproductive rights (Gothreau et al., 2022). Furthermore, exposure to objectifying imagery cues considerations that women are less moral and warm, two important humanizing traits in social cognition (Fiske et al., 2002). It could be expected that those exposed may be less likely to support women’s access to reproductive services. These extant findings lead us to our first hypothesis:

$ {H}_1: $ Exposure to objectifying portrayals of women will decrease support for women’s access to abortion and birth control, as well as support for Planned Parenthood funding.

We also explore the potential effect of objectifying imagery on other gender-salient political attitudes like hiring policies that advantage women job candidates, expanded childcare access, and government funded free period products. Though the mechanism that could potentially link objectification to these policy outcomes is less clear, we might expect that women-objectifying imagery primes the considerations mentioned above, as well as sexist attitudes (Fox and Potocki, 2016; Swami and Voracek, 2013) which we know predict decreased support for policies and public opinion issues that involve gender equality or have gendered implications (Gothreau et al., 2022). This leads to our second hypothesis:

$ {H}_2: $ Exposure to objectifying portrayals of women will decrease support for other gender-salient policies that benefit women.

Given that men are more likely to objectify women (Zurbriggen et al., 2011), we might also expect to see heterogenous treatment effects by gender. Though some prior studies on the impact of sexually objectifying portrayals of women have failed to find any moderating effects of gender (Gervais et al., 2012; Heflick et al., 2011), we expect that men will be more likely to perceive the objectified targets as less warm, moral, and competent. Men also tend to hold more sexist attitudes than women (Cowie et al., 2019) and these attitudes may be more likely to be primed during exposure to sexually objectifying imagery. Given these considerations, this leads to our final hypothesis:

$ {H}_3: $ Men who are exposed to objectifying portrayals of women will show greater decreased support for women’s access to abortion and birth control, as well as other gender-salient policies that benefit women, than women.

Experimental design

We conducted two separate but complementary survey experiments in the United States. Hypotheses were pre-registered on Open Science Framework. ² We discuss deviations from the pre-registration in the Supplementary Materials. The surveys were not identical but shared a similar structure in which respondents were invited to participate in a survey about advertisements, personality, and politics. ³ In both studies participants were randomly assigned to either a control condition where they viewed neutral imagery or a treatment condition where they viewed women-objectifying imagery. After exposure to the treatment, respondents completed a series of questions about different policy attitudes. We present the results of the two studies separately with pooled results in the Supplementary Materials.

Study 1

Participants for Study 1 were recruited among adults in the USA on Amazon’s Mechanical Turk in September 2018 (n = 500). Though our sample is whiter, more educated, and more liberal than the US population (see Supplementary Table A1 for full demographic information), MTurk samples tend to be more representative than in-person convenience samples and suitable for experimental research (Berinsky et al., 2012; Coppock, 2019). After completing pre-test measures, participants were randomly assigned to the either the treatment or control condition. At this point, participants were told they were taking part in a separate and unrelated study on marketing and advertisements. In both groups, participants viewed a series of 10 advertisements in a randomized order for 10 seconds each. After viewing each ad, they were asked to rate the ad on a scale of 1 (“strongly dislike”) to 7 (“strongly like”). In the control condition (n = 246), participants viewed neutral images such as ads for sneakers, cars, and watches. Those assigned to the treatment condition (n = 253) viewed ads that featured objectifying imagery of women. These advertisements included images in which only parts of women’s bodies were displayed, or women’s bodies were used as stand-ins for objects. These photos were pretested in a previous MTurk study to ensure they were perceived as objectifying and that they depicted female cultural appearance ideals. ⁴

The post-test that followed included a battery of questions about various policy attitudes such as attitudes regarding trade policy, education, and the death penalty. Interspersed in this battery were questions about access to contraception, support for abortion, Planned Parenthood, expanded childcare access, and support for employment policies that advantage women candidates. Responses for the abortion item ranged from 1 (“abortion is never justified”) to 11 (“abortion is always justified”). Response categories for all other items ranged from 1 (“Strongly Disagree”) to 5 (“Strongly Agree”).

Study 2

Participants for Study 2 were recruited among adults in the USA using the survey recruitment platform Forthright by Bovitz Inc. in February of 2022. Again, our online sample skewed slightly whiter, more educated, and more liberal than the US population (see Supplementary Table A2). The structure of the survey flow in Study 2 closely mirrors that of Study 1. After completing pre-test measures, participants were randomly assigned to the treatment (n = 514) or control condition (n = 503). In both conditions, participants viewed eight images for 5 seconds each. A portion of these photos were taken from the International Affective Picture System (IAPS), a widely used collection of images that are pretested for arousal and valence (Bradley and Lang, 2007). For the objectification condition, we chose four erotica images featuring women’s bodies from the database to use as stimuli photos as well as four photos from the pretested advertisements used in Study 1. In the control condition, participants viewed a series of neutral photos from the IAPS that featured household objects like lamps and furniture. After viewing the images, participants were offered an open-ended question asking what items were being advertised in the previous photos. The inclusion of this open-ended question, as well as the question after each photo in Study 1, was to ensure participants stayed engaged in the task and also to further the light deception that the study was indeed about advertisements.

After the stimuli, participants completed a battery of gender-salient policy attitudes that included the same items about abortion, birth control, Planned Parenthood, funding for childcare, and support for hiring policies that advantage women that were included in Study 1. In addition to these items, respondents also completed a question about government funded free menstrual products. Responses for the abortion item ranged from 1 (“abortion is never justified”) to 9 (“abortion is always justified”). Response categories for all other items ranged from 1 (“Strongly Disagree”) to 5 (“Strongly Agree”).

Results

All estimates are based on linear regression models using robust standard errors and controlling for gender, age, education, income, race, and party identification (Green and Aronow, 2011). Full results for all analyses are in the Supplementary Materials. Once again, we predicted that the objectifying treatment would have a negative impact on support for the gender-salient policy items. In Study 1, we find statistically significant, negative average effect of the objectification treatment on support for access to contraception (b = −0.26 on a 1–5 scale, SE = .11, t = −2.5, p = .01) as well as support for employment policies that favor women (b = −0.18 on a 1–5 scale, SE = .10, t = −1.8, p = .08). Effects on the other three gender-related policy statements, as well as on the placebo items concerning education, free trade, and the death penalty, are statistically insignificant. Figure 1 shows the average treatment effects of the objectifying treatment on policy support. Thicker bands are 90% CIs and thinner bands represent 95% CIs. These results indicate partial support for $ {H}_1 $ and $ {H}_2 $ .

Figure 1. Average treatment effects of objectification on policy support (Study 1).

To test $ {H}_3 $ , we interacted our treatment indicator variable with binary gender with men serving as the reference groups. Again, we expected men to show greater decreased support for gender-salient policies that benefit women when exposed to the objectifying treatment. We only found a significant interaction between the treatment and gender when it came to abortion attitudes, though the effect size is substantively small. More specifically, as displayed in Figure 2, we find that the treatment effect on support for abortion is significantly lower (more negative) for women than for men, which is contrary to our expectations in $ {H}_3 $ . In sum, we see evidence that accords with our prediction that objectification reduces support for gender-salient policies that support women, though we find no support for the idea that men are more likely to be impacted by the treatment.

Figure 2. Average treatment effect on abortion support by gender (Study 1).

Turning to Study 2, we find a significant positive average effect of the treatment on support for employment policies that advantage women (b = 0.16 on a 1–5 scale, SE = .07, t = −2.2, p = .03), in contrast with the negative effect uncovered in Study 1. Average effects on the other five gender-salient policies are insignificant. Figure 3 shows the average treatment effects on policy support.

Figure 3. Average treatment effects of objectification on policy support (Study 2).

Again, to test $ {H}_3 $ , we interacted our treatment indicator with binary gender. We find significant interactions in the case of support for abortion access and preferential hiring (Figure 4). The interaction model with preferential hiring reveals that the positive effect of the treatment was driven by women in the sample. There is no significant difference between the treatment and control group among men in terms of support for hiring policies that advantage women. On average, on a scale of 1–5 with higher values indicating more support, women in the control group were at about a 3.1 on the scale whereas women in the treatment group were slightly above a 3.5 on the scale. Interestingly, we do find support for $ {H}_3 $ with respect to abortion attitudes. There was no statistically significant difference between treatment and control group among women with respect to abortion attitudes. However, men exposed to the objectifying treatment were significantly less supportive of access to abortion than men in the control. Group. On average, on a scale of 1–9 with higher values indicating more support, men in the control were at about a 5.4 on the scale and men in the treatment group were at approximately a 4.8.

Figure 4. Average treatment effect on abortion support by gender (Study 2).

Limitations and conclusions

In sum, we find support for our hypotheses across these two experimental studies though in some cases, we arrive at divergent findings. In Study 1, in line with theoretical expectations, we find a negative average treatment effect of exposure to objectifying images of women on support for access to contraception and hiring policies that advantage women candidates. Treatment effects conditioned on gender revealed that women, but not men, in the treatment group displayed lower support for abortion, though the effect size was very small. In Study 2, we find a positive effect of the treatment on preferential hiring policies. Treatment effects conditioned on gender revealed that this positive effect was driven by women in the sample. Furthermore, the interaction between the treatment and gender showed, in support of $ {H}_3 $ that men exposed to the objectifying treatment were significantly less supportive of abortion than men in the control.

What could explain some of our divergent findings? We note that both studies indicate heterogeneity in effects by gender. This suggests that the effect of seeing objectified images is real but malleable in principle, and that differences in sample composition as well as study timing may explain the differences in effects. Study 1 was a convenience sample, while Study 2 was representative on key demographics. Differences in sample demographics could potentially explain divergent findings across studies. Study timing could also be a factor. Though abortion has been a salient issue in American politics for decades, the potential overturning of Roe v. Wade was a continual news story in 2022, when Study 2 took place. Given the issue’s relevance to women, women’s attitudes could have been increasingly crystallized by the news and less susceptible to the treatment. Overall, the data from Study 1, collected in 2018, should be considered with some degree of caution considering the multitude of gender-salient political events that have unfolded in the past 5 years.

One factor, other than sample composition or timing, that could explain the average negative effect of the treatment in Study 1 on preferential hiring and the positive effect in Study 2 is the subtle differences in treatment stimuli. Though there was overlap in the objectifying stimuli photos, Study 2 also contained erotica images. These photos were more explicit and sexualized. Though it is speculative, we can imagine that this might have sparked a backlash effect among women in the sample leading them to express more support for affirmative action hiring policies that advantage women.

Another significant limitation of our study is that we are unable to directly test potential mechanisms. We suggest that the extant work points to increased sexist and gender-specific system-justifying attitudes, as well as the overall dehumanization of women, as the most compelling potential mechanisms but we did not test these mediators. Furthermore, our control images did not contain images of women. One could argue that perhaps just the mere exposure to images of women shifted attitudes and it was not necessarily the presence of objectification. Though we feel confident this is not the case given our pre-testing of the images and the extant work on objectification and dehumanization, this is a limitation.

Building on decades of work in social psychology on the impacts of pervasive objectification, we posited that these effects may extend into the political realm and more specifically, impact policy attitudes. We argued that objectification, as a form of dehumanization, decreases perceptions of women’s competence and warmth, as well as primes hostile sexist attitudes. This would lead to decreased support for social policies that benefit women, such as increased access to reproductive care, government funded childcare, period products, and hiring policies that advantage women. Overall, we interpret our results as indicating that exposure to objectifying images can move policy attitudes at least in the short-term. This is relatively surprising given the brief exposure in the experiments and the fact that policy attitudes on abortion and birth control are crystallized in comparison to other types of policy attitudes. Furthermore, there seems to be some heterogeneity as a function of gender. However, the direction of both the average and gender-specific effects varied strongly across our two studies, which suggests significant heterogeneity in effects. In future work, scholars should assess and replicate our findings as well as delve further into the mechanisms that connect exposure to objectification and political attitudes.