Girls as objects, boys as humans: Young children tend to be objectified along gender lines.

Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice J Pers Soc Psychol . Author manuscript; available in PMC: 2026 Apr 14. Published in final edited form as: J Pers Soc Psychol. 2025 Apr 14;129(2):266–285. doi: 10.1037/pspa0000448 Search in PMC Search in PubMed View in NLM Catalog Add to search Girls as objects, boys as humans: Young children tend to be objectified along gender lines Rachel A Leshin Rachel A Leshin 1 Princeton University, Department of Psychology, Peretsman Scully Hall, Room 522, Princeton, NJ 08540 2 New York University, Department of Psychology, 6 Washington Place, Room 301, New York, NY 10003 Find articles by Rachel A Leshin 1, 2 , Marjorie Rhodes Marjorie Rhodes 2 New York University, Department of Psychology, 6 Washington Place, Room 301, New York, NY 10003 Find articles by Marjorie Rhodes 2 Author information Article notes Copyright and License information 1 Princeton University, Department of Psychology, Peretsman Scully Hall, Room 522, Princeton, NJ 08540 2 New York University, Department of Psychology, 6 Washington Place, Room 301, New York, NY 10003 Author Note: Research reported in this publication was supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under Award Number R01HD087672 (to Rhodes) and Award Number F31HD107965 (to Leshin). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. All data, code, and materials for the studies included in this manuscript are publicly available on the Open Science Framework: https://osf.io/89fpn/ . Pre-registrations for Study 2b, 3b, and 3c are publicly available at https://osf.io/acu5y (Study 2b), https://osf.io/6jnqp (Study 3b), and https://osf.io/w6pm7 (Study 3c). ✉ Corresponding author: [email protected] Issue date 2025 Aug. PMC Copyright notice PMCID: PMC12245595  NIHMSID: NIHMS2063855  PMID: 40232820 The publisher's version of this article is available at J Pers Soc Psychol Abstract Objectification—the psychological phenomenon of relegating people to the status of objects , denying their humanness—is associated with a host of negative consequences for those targeted, from diminished cognitive performance to heightened risk of danger. Girls and women constitute the primary targets of objectification; thus, these harms fall disproportionately on them. Despite the persistence of such gendered patterns, however, it is not clear how they arise. That is, we do not yet know whether and to what extent perceivers objectify children along gender lines (i.e., associating girls with objects and boys with humans ), thus limiting our grasp of this phenomenon both theoretically and practically. In the present studies, we addressed this gap on two fronts. First, we tested whether adults ( n =430) objectify young children based on gender. Second, we tested whether children themselves ( n =418, ages 4–10 years) display gendered patterns of objectification toward other children. We found evidence that adults objectify children based on gender: in both their categorizations and attributions, adults revealed overlap between their concepts of girls and objects and their concepts of boys and humans (although the degree to which each specific pattern manifested varied across studies). Children showed more limited evidence of this phenomenon: boys, but not girls, displayed the predicted pattern of conceptual overlap, and only in their categorizations. Together, these findings reveal that gender-differentiated patterns of objectification may take root in perceptions of young children—suggesting that the gendered consequences of this phenomenon may be larger in scope and earlier-emerging than previously assumed. Keywords: Objectification, gender, development Statement of Limitations: Our research represents the first test, to our knowledge, of the tendency for adults (and children) to objectify children along gender lines and has a few notable limitations. First, we did not assess whether and to what extent our measures of objectification toward child targets converge with prior measures of objectification (and related phenomena, such as dehumanization) toward adolescent and adult targets. Relatedly, because our tests of attribution-based objectification were anchored on physiological attributes (e.g., “gets hungry”), our grasp of how participants conceived of child targets with respect to other types of human-like and object-like attributes (e.g., those in the psychological domain, as tested in prior research) is limited. Additionally, our measure of categorization-based objectification included only White stimuli; for a more complete investigation of this process, future research will need to diversify the stimulus set used in tests of categorization-based objectification. Finally, our analyses–and, in particular, our tests of categorization-based objectification–would benefit from replication with larger and more diverse samples. Such replications would help us to further clarify the robustness of our effects and, more specifically, shed light on which forms of objectification revealed in follow-up tests are most reliable. Objectification describes the psychological phenomenon of relegating people to the status of objects , denying their humanness (e.g., Fredrickson & Roberts, 1997 ; Heflick & Goldenberg, 2014 ). This phenomenon is linked to reduced performance on cognitive tasks ( Fredrickson et al., 1998 ; Guizzo & Cadinu, 2017 ), diminished appraisals of competence ( Heflick & Goldenberg, 2009 ; Holland & Haslam, 2016 ), heightened vulnerability to danger ( Rudman & Mescher, 2012 ; Holland & Haslam, 2016 ), and psychological dysfunction ( Harrison & Fredrickson, 2003 ; Moradi & Huang, 2008 ). Women and girls constitute the primary targets of objectification, both in daily life ( Fredrickson & Roberts, 1997 ; Zurbriggen et al., 2010) and lab-based experiments ( Loughnan & Vaes, 2017 ); thus, these psychological and societal harms fall disproportionately on them, beginning as early as elementary school ( Clark & Tiggemann, 2006 ; Slater & Tiggemann, 2016 ; Starr & Ferguson, 2012 ; Starr & Zurbriggen, 2019 ; Zurbriggen et al., 2010). Given these consequences—especially for women and girls—it is critical to understand how objectification develops: that is, how do people come to associate gender with the concept of objects (vs. humans ) in the first place? To investigate this issue, we asked two distinct but related questions. First, to what extent do adults objectify young children along gender lines? And second, do children themselves display gendered patterns of objectification in perceiving other children? We administered two novel measures to probe conceptual overlap between the categories of girls/objects and boys/humans —our operationalization of objectification—among children and adults from across the U.S. Objectification as a Gendered Phenomenon The motivation to examine objectification in such a manner stems from the gendered harms that arise from it. Indeed, Fredrickson & Roberts’ canonical objectification theory (1997) first emerged as an attempt to shed light on the starkly imbalanced rates of mental health concerns in young women compared to young men. According to this theory, the third-party perspective on the self that objectification induces—most often in women and girls—contributes to gender differences in psychological disorders, including depression and anorexia (see Moradi & Huang, 2008 , for a review), effects that emerge as early as elementary school ( Starr & Zurbriggen, 2019 ). Objectification also poses a threat to broader issues of social equity, by shaping third-party appraisals of competence: for example, priming the objectification of women (but not men) political candidates led adults to perceive these candidates as less competent, and consequently, diminished their likelihood of voting for them ( Heflick & Goldenberg, 2009 ). The harms of objectification also manifest in moment-to-moment psychological processing. For example, because people visually process objects differently from humans—perceiving them as a series of parts (i.e., analytic processing ) rather than as holistic beings (i.e., configural processing )—objectification influences person perception. Indeed, observers are more likely to engage in analytic processing when viewing pictures of sexualized women compared to men ( Bernard et al., 2012 ). Objectification also has in-the-moment implications for appraisals of others’ humanness: following an appearance-based objectification prime, people rated women (but not men) as lower in traits linked to humanity (e.g., morality; Heflick et al., 2011 ; see also Heflick & Goldenberg, 2009 ). This link between objectification and perceptions of humanness is reciprocal, such that highlighting the humanity of women targets before a visual processing task reduced people’s tendency to process them analytically ( Bernard et al., 2015 ). Thus, perceivers tend to view women, more so than men , in a manner akin to objects across a diverse range of measures (e.g., Bernard et al., 2012 ; Heflick et al., 2011 ; Vaes et al., 2011 ). These findings suggest a basic conceptual overlap between people’s concepts of women / objects (and men/humans ). Less is known, however, about whether objectification manifests in a similarly gendered manner toward young children—that is, whether people associate girls with objects (and, conversely, boys with humans ). Girls as Targets of Objectification From early in childhood, young girls engage in self-objectification (as well as related phenomena like self-sexualization; Clark & Tiggemann, 2006 ; Starr & Ferguson, 2012 ; Starr & Zurbriggen, 2019 ). Indeed, a majority of 6- to 9-year-old girls select sexualized (vs. non-sexualized) dolls as reflecting their actual and ideal selves, and the tendency to do so predicts self-objectification ( Starr & Zurbriggen, 2019 ). Moreover, nearly half of mothers reported that their 5- to 8-year-old daughters used at least one beauty product and sought feedback on their appearance, and over three-quarters reported a tendency for their daughters to inspect how they look in the mirror ( Slater & Tiggemann, 2016 ). The extent to which other people perceive young girls (relative to young boys) in an objectified manner, however, is less clear. Prior work finds that adults objectify preadolescent girls more when they are dressed in sexualized (vs. non-sexualized) clothing ( Holland & Haslam, 2016 ; see also Graff et al., 2012 ), but such paradigms did not compare how much adults objectify girls relative to boys. Thus, past findings do not address whether gender itself shapes the objectification of young children—a question that can facilitate theoretical contributions into the nature of objectification more broadly and offer practical insights regarding the presence of objectification in children’s daily lives. Although not yet tested directly, objectification could manifest in a similarly gendered manner toward children due to the shared emphasis on physical appearance placed on women and girls—a core component of objectification, as well as object perception more generally. Indeed, prominent frameworks of objectification (e.g., literal objectification , Heflick & Goldenberg, 2014 ; appearance-focused objectification , Morris & Goldenberg, 2015 ) contend that an orientation toward physical appearance is a driving force of objectification; from this perspective, the disproportionate societal emphasis placed on women’s physical appearance accounts for their outsized role as targets of objectification ( Heflick & Goldenberg, 2009 ; Heflick et al., 2011 ; Morris et al., 2018 ). More broadly, classic work in cognitive psychology reveals that the classification of everyday objects hinges on what objects look like more than other characteristics (e.g., the object’s function; Malt & Johnson, 1992 ), revealing a more foundational link between object perception and orientation to physical appearance. Such gender-differentiated focus on physical appearance begins early in life. Indeed, within the first 24 hours of a child’s birth, new parents use words like “pretty” and “cute” more often to describe newborn girls than boys (controlling for relevant physical features such as weight and length; Rubin et al., 1974 ). Parents in this study also noted a resemblance between daughters and mothers more often than sons and fathers, further revealing an inflated focus on girls’ physical appearance (one that likely only increases with development, as gender-differentiated self-presentations heighten; Halim et al., 2011 ). This disproportionate emphasis on girls’ appearance mirrors broader U.S. culture: in animated cartoons, for example, female characters are depicted as physically attractive more often than male characters ( Klein & Shiffman, 2006 ), and toys marketed to young girls emphasize physical attractiveness ( Blakemore & Centers, 2005 ), often through the direct sale of beauty products ( Auster & Mansbach, 2012 ). The spotlight on girls’ physical appearance is pervasive enough that children themselves endorse it. When asked to provide descriptions of girls, for example, 3- to 10-year-old children’s responses are dominated by appearance-related stereotypes (e.g., “girls wear dresses”). Children cited these stereotypes for girls more often than stereotypes in other domains (e.g., activities), and critically, mentioned them more often for girls than boys ( Miller et al., 2009 ). In addition to seeping into children’s descriptive characterizations of girls, children’s prescriptive gender norms also reflect an association between girls and physical appearance. Indeed, young children disapprove more of boys’ choice to appear like girls (e.g., wear girl-like clothes) than girls’ choice to appear like boys ( Blakemore, 2003 ). When turned inward, this rigid orientation toward appearance can be harmful: by middle childhood, the strength of girls’ appearance schemas predicts their body dissatisfaction ( Clark & Tiggemann, 2007 ; Hargreaves & Tiggeman, 2002 ). Altogether, the above findings suggest that the cultural emphasis placed on women’s (vs. men’s) physical appearance also extends to young girls (vs. boys)—suggesting that the gendered patterns of objectification that emerge for adult targets may also apply to child targets. Given the theoretical and practical importance of this possibility, it is critical to test it empirically. Testing the Objectification of Girls One plausible reason that objectification has rarely been tested with child targets concerns the role of sex and sexuality in conceptualizations of objectification ( Kant, 1963 ; Bernard et al., 2012 ; 2015 ; Fredrickson & Roberts, 1997 ; Vaes et al., 2011 ). However, certain forms of objectification are not sexual in nature, thus providing theoretical support and methodological insights for probing this question with child targets. One such form is literal objectification —the process by which people deny others characteristics that distinguish humans from objects (e.g., Heflick & Goldenberg, 2009 ; Heflick & Goldenberg, 2014 ; Heflick et al., 2011 ). A key feature of literal objectification is its occurrence in neutral, everyday contexts: all that is required for it to occur is a focus on physical appearance. This form of objectification, alternately termed appearance-focused objectification , is clearly differentiated from other forms that do invoke sex and sexuality (i.e., sexual-focused objectification; Morris & Goldenberg, 2015 ). For example, when asked to view an image of a beauty-focused woman (i.e., a runway model), adults rated her lower on traits distinguishing humans from objects (e.g., active, curious; Morris et al., 2018 ). Similarly, philosophical theories of objectification define it, simply, as, “treating as an object […] what is, in fact, a human being” ( Nussbaum, 1995 ). Under this definition, objectification consists of different facets that may or may not travel together—some suggest sexual motives (e.g., violability), while others do not (e.g., inertness). Further evidence for such basic forms of objectification appears in work on dehumanization and mind perception. Indeed, objectification is nested within Haslam’s two-pronged model of dehumanization as mechanistic dehumanization —occurring when people draw a psychological equivalence between humans and objects ( Haslam, 2006 )—and research on mind perception links objectification to how people appraise others’ minds (and, specifically, their agency; Gray et al., 2011 ). Thus, objectification can function as a basic and non-sexual process that could be directed toward young girls and empirically tested—for example, by examining associations with basic object-like vs. human-like characteristics. The simplicity of such approaches additionally affords us the ability to test these questions with child participants who, from early on, can reason about humanity and mental life (e.g., Costello & Hodson, 2014 ; McLoughlin & Over, 2017 ; McLoughlin et al., 2018 ; Weisman et al., 2018 ; see McLoughlin & Over, 2018 for a review). Importantly, this phenomenon is distinct from gender stereotyping ( Eagly et al., 2020 ): whereas gender stereotyping reflects people’s belief that certain features (e.g., those associated with humanness or object-ness) are typical of boys vs. girls, objectification reflects the selective overlapping of people’s concepts of girls with objects and those of boys with humans . Of course, these two processes likely intersect: part of what makes pairs of concepts (e.g., objects and girls) overlap is the extent to which people perceive features to be shared across them, and these features may manifest as category stereotypes (e.g., doesn’t smell ; see Foster-Hanson & Rhodes, 2023 and Brewer et al., 1981 for links between category representations and stereotypes). To the extent that stereotyping and objectification are related, however, it is not clear whether one precedes the other (i.e., do gender stereotypes cause objectified vs. humanized representations of girls vs. boys, or do these types of representations cause the formation of gender stereotypes?) or whether they develop in parallel—an issue we return to in the General Discussion. The Present Research In the present research, we asked whether adults and children objectify young children along gender lines. In the absence of a gendered pattern, we would expect people to associate both boys and girls with humans (to the same degree), and neither with objects. We thus operationalized objectification as the extent to which people’s concepts of girls/objects (and, conversely, boys/humans ) were selectively overlapped. We interpreted more conceptual overlap between each set of categories (i.e., girls/objects , boys / humans ) as greater evidence of gender differentiation in objectification. We tested this question with adults and children (ages 4–10) from across the U.S. We selected this age-range because we reasoned that gendered patterns of objectification likely follow from cultural input emphasizing girls’ (vs. boys’) physical appearance and thus sampled children who likely already had exposure to such input (e.g., media). Additionally, as children in this age-range undergo important developmental shifts in their representations of gender (e.g., category flexibility) and other relevant phenomena (e.g., social status; Halim et al., 2011 ), this window enabled us to explore questions of developmental change. We hypothesized that, for our more deliberative measure of objectification, the predicted patterns would strengthen with age as children develop more complex notions of gender and accumulate more exposure to objectifying media (e.g., Zurbriggen et al., 2010). In our investigation, we sought to capture both the addition of object-like characteristics and the denial of human-like characteristics (with the latter invoking theories that link objectification to dehumanization; e.g., Gervais et al., 2013 ; Heflick & Goldenberg, 2014 ; Heflick et al., 2011 ; Loughnan et al., 2010 ) and thus designed measures that varied both the gender (girl, boy) and type (object-like, human-like) of stimuli. Since we would, a priori, expect both girls and boys to be associated with humans (and not with objects), we interpreted two-way interactions between these variables as evidence of objectification. In all instances of an interaction, we conducted follow-up tests to probe the particular form of objectification driving it (e.g., the denial of human-like characteristics to girls vs. boys, the addition of object-like vs. human-like characteristics to girls). Although we view each of these patterns as consistent with objectification, the distinctions between each (and the extent to which some patterns emerge and others do not) may be meaningful; thus, we interpret our findings both in reference to the overall two-way interaction and the specific form(s) of objectification revealed in follow-up tests. Procedures for all studies reported in this manuscript were approved by the University’s Institutional Review Board (IRB-FY2016–760 and IRB-FY2016–961). Study 1 In Study 1, we sought to obtain proof of concept that objectification could feasibly target young girls. To do so, we considered the precondition to objectification described above: a heightened focus on physical appearance. We examined whether such a gender-differentiated focus on appearance exists for perceptions of young children by comparing adults’ appearance-related comments for images of girls vs. boys. Method Participants Adults ( N =250) were recruited from the psychology subject pool at a large university in New York City ( n =77) and from the online crowdsourcing platform Prolific ( n =173). The sample consisted predominately of young adults ( M age = 28.70 years, SD age = 11.77, range : 18–70) and was roughly split by gender (53% women; 47% men). Participants reflected a range of racial-ethnic backgrounds: 62% identified as White, 18% identified as Asian, 8% identified as Black, 9% identified as Hispanic, and 2% identified as another racial category or as “Other.” On average, the sample leaned politically liberal with respect to social issues (from 1-Very Liberal to 7-Very Conservative, M polid = 2.74, SD polid = 1.62, range : 1–7). 1 Procedure First, we introduced participants to a fictional character, who they were told was visiting from a faraway place and wanted to learn about people here. We then showed participants a picture of a child—a girl or boy, counterbalanced across participants, playing at a park—and prompted them tell the fictional character four things about the child ( Figure 1 ). We selected images of children engaged in an activity (rather than headshots) to give participants multiple different cues to focus on beyond physical appearance. We removed all references to the child’s gender (e.g., gender pronouns) from participants’ open-ended responses and then had each response coded, by two independent research assistants, for content related to the child’s appearance (e.g., what they were wearing, whether they were smiling). Responses containing appearance-related content were coded as “1” and those not containing it were coded as “0.” Reliability was excellent ( r = .88), and discrepancies were resolved by the first author. Figure 1. Open in a new tab Stimuli used in Study 1. (Images have been blurred.) Transparency and Openness Study 1 was exploratory in nature, and thus, analyses were not pre-registered. All materials, data, code, and codebooks are publicly available on OSF. Data were analyzed using R, version 4.3.2 (R Core Team, 2020). Results Analytic Strategy We implemented generalized linear models from the lme4 package and specified a binomial distribution (for each response, 0 represented the absence of appearance-focused content and 1 represented the presence of it). We entered stimulus gender (boy, girl; dummy-coded as 0, 1) as a between-subjects predictor and modeled subject ID as a random intercept to account for the nested structure of the response data (i.e., 4 data points per participant). Findings Participants were more likely to comment on the physical appearance of the child pictured when it was a girl compared to when it was a boy ( b = .52, SE = .19, t = 2.76, p = .006; Figure 2 ). Adults commented on girls’ appearance the majority of the time ( M = .63; comparison to .50, t = 5.67, p < .001) but on boys’ appearance roughly half the time ( M = .52; comparison to .50, t = .89, p = .374). Figure 2. Open in a new tab Participants’ likelihood of making a comment about physical appearance—a theorized precondition to objectification (left)—and mental state (right) as a function of stimulus gender. Large shapes represent group means, small shapes represent individual response patterns, and error bars represent standard errors. In follow-up analyses, we compared adults’ likelihood of making appearance-focused comments with their likelihood of making those related to the child’s humanity. To do so, we had the same coders independently code responses for the presence of mental state language (e.g., what the child is thinking, how the child is feeling; following McLoughlin & Over, 2017 ); coders again achieved high reliability ( r = .91), and discrepancies were resolved by the first author. To test whether participants’ references to the mental state vs. physical appearance of the child pictured varied by stimulus gender, we ran a second model in which we interacted these two variables (i.e., type of comment x stimulus gender). This two-way interaction was significant ( b = .96, SE = .22, t = 4.32, p < .001). Parsing this interaction, we observed that—in addition to making more appearance-focused comments about girls than boys—participants made more mental-state-focused comments about boys ( M = .22) than girls ( M = .14; p = .003). Discussion Study 1 provided evidence that adults attend to physical appearance more when describing girls than boys—an important pre-condition to objectification (e.g., Fredrickson & Roberts, 1997 ; Heflick & Goldenberg, 2014 ; Heflick et al., 2011 ; Morris et al., 2018 ). We also observed that, in addition to focusing more on girls’ appearance than boys’, adults focused more on boys’ mental state than girls’. Together, these findings lay the foundation for our test of whether adults and children objectify young children along gender lines in Studies 2–3. Study 2 We next administered the first of two measures developed to probe conceptual overlap between girls/objects and boys/human s. In this measure, we evaluated how readily participants categorized girls and boys—depicted in human form (i.e., real kids) and in object form (i.e., doll-like toys)—as objects or humans using a method previously validated for children and adults ( Lei et al., 2020 ; Leshin et al., 2022 ). Study 2a Method Participants Adults ( N =77) were recruited from the psychology subject pool at a large university in New York City ( M age = 19.45 years, SD age = 1.19, range : 18–24). Participants were roughly split by gender (54% women; 46% men) and reflected a diverse range of racial-ethnic backgrounds: 39% identified as Asian, 24% identified as White, 21% identified as Hispanic, 13% identified as Black, 1% identified as Middle Eastern/North African, and 1% identified as “Other.” Our sample skewed politically liberal with respect to social issues (from 1-Very Liberal to 7-Very Conservative, M polid = 2.68, SD polid = 1.23, range : 1–6). Children ( N =66) between the ages of 4 and 8 years ( M age = 6.22 years, SD age = 1.24, range : 4.02–8.87) were recruited from a children’s museum in New York City. Our sample was evenly split between boys and girls (50% in each category) and reflected a wide range of racial-ethnic backgrounds (based on parent report, 33% multiracial, 32% White, 12% Asian, 9% Black, 14% unreported). Parents who self-reported political ideology revealed a slight skew toward liberalism but fell closer to the center of the scale than our sample of adults (from 1-Very Liberal to 7-Very Conservative, M polid = 3.40, SD polid = 1.32, range : 1–6). 2 Procedure We first provided a brief introduction to the speeded categorization task (programmed in Inquisit 5.0; Millisecond, 2016 ) in which participants were told that they would see a series of stimuli flash on a computer screen (one at a time, for a maximum of three seconds) and that the goal was to categorize each as quickly and accurately as possible. Once the computer registered audio input (i.e., a verbal categorization), it would advance to the next stimulus. Participants completed a series of practice trials to familiarize themselves with the set-up; these stimuli were either birds or toys that looked like birds , and participants were told to categorize each as a “bird” or “toy.” Next, participants saw a new set of stimuli of kids or toys that looked like kids and were instructed to categorize each as either a “kid” or “toy.” These critical trials contained 24 unique images, presented in a randomized order across two blocks; 12 were photographs of kids (6 boys and 6 girls), and 12 were pictures of toys that looked like those kids (6 boys and 6 girls; we referred to these as “toys” rather than “dolls” for the greater gender neutrality). All stimuli, sourced from Google Images, depicted White elementary-school-aged children ( Figure 3 ). In selecting stimuli, we matched each girl-gendered toy and boy-gendered toy with a respective girl-gendered kid and boy-gendered kid that looked like it, such that toy-kid pairs were roughly matched on their physical appearance. A small sample of adults on Amazon’s Mechanical Turk confirmed the physical similarity of all pairs (and, further, confirmed that similarity ratings were roughly equivalent across pairs); for more details, see Supplemental Materials . Figure 3. Open in a new tab Stimuli used in the speeded categorization task in Studies 2a and 2b. Stimuli consisted of 12 pictures of kids (top row) and 12 pictures of toys that looked like kids (bottom row). (Images have been blurred.) Videos of study sessions were coded in Datavyu ( Datavyu Team, 2014 ) for reaction time and accuracy. Due to technical issues (e.g., videos failing to record), a subset of data ( n =14 adults, n =7 children) had to be excluded from analyses; thus, we were left with data from 63 adults and 59 children. Each study video was coded by a trained research assistant, and the first author provided reliability coding on 25% of videos; interrater agreement was excellent (for adults, reaction time: r = .96, p < .001 and accuracy: 96.4% agreement; for children, reaction time: r = .98, p < .001 and accuracy: 99.5% agreement). Transparency and Openness Since Study 2a represented our first attempt to track categorization-based objectification toward children, we treated this study as exploratory and did not pre-register methods or analyses. Materials, data, and code are publicly available on OSF. Data were analyzed using R, version 4.3.2 (R Core Team, 2020). Results Analytic Strategy We ran a series of mixed generalized linear models using the lme4 package. Reaction time analyses were modeled as linear regressions and error analyses as binary logistic regressions (0 = correct response, 1 = incorrect response) 3 . In all models, we entered the gender of the stimulus (boy, girl; dummy-coded as 0 and 1, respectively) and the type of stimulus (kid, toy; dummy-coded as 0 and 1, respectively) as within-subjects predictors, allowing them to interact. We entered subject ID as a random intercept to account for the nested structure of the data. In subsequent analyses, we added participant gender (man, woman; dummy-coded as 0 and 1, respectively) as an additional between-subjects predictor; due to the relatively small size of our samples ( N =63 for adults, N =59 for children), however, these analyses should be interpreted with caution. We conducted an additional analysis with our child sample to consider the effect of participants’ (mean-centered) age. In reporting effects from our main models, we include unstandardized beta coefficients for ease of inference, and we follow up on all significant interactions with simple-effects tests from the emmeans package. Findings (Adults) In adults, we observed a two-way interaction between stimulus gender and stimulus type for our measure of reaction time ( b = −.28.64, SE = 10.56, t = −2.71, p = .007; Figure 4 ), and follow-up tests revealed two key features of objectification. First, adults were slower to correctly categorize girl-gendered kid stimuli compared to girl-gendered toy stimuli ( b = 26.81, SE = 7.43, t = 3.61, p < .001), indicating that a “toy” categorization was more accessible than a “kid” categorization for girl stimuli. Adults were also slower to correctly categorize boy-gendered toy stimuli compared to girl-gendered toy stimuli ( b = 19.62, SE = 7.50, t = 2.61, p = .009), revealing that a “toy” categorization was easier for adults to make when viewing toys depicted as girls relative to those depicted as boys. As an additional robustness check, we re-ran our main model with the mean similarity rating of each kid-toy stimulus entered as a random intercept to confirm that any differences in the perceived similarity of matched toy-kid pairs were not driving our pattern of results. In this model, the two-way interaction between stimulus type and stimulus gender remained significant ( b = −31.70, SE = 10.44, t = −3.04, p = .002). We found no moderation by participant gender in our primary model ( b = .98, SE = 21.78, t = .05, p = .964). Figure 4. Open in a new tab Adults’ reaction time on the speeded categorization task in Study 2a as a function of stimulus gender and stimulus type (note: the y-axis has been truncated slightly to highlight the interaction effect). Larger shapes represent group means, small shapes represent individual data points, and error bars represent standard errors. Adults made very few categorization errors overall (mean accuracy = 96.40%), and we did not observe the predicted interaction between stimulus gender and stimulus type on adults’ accuracy ( b = −.53, SE = .47, t = −1.11, p = .267). As with reaction time, the three-way interaction with participant gender was not significant ( b = 1.41, SE = 1.04, t = 1.35, p = .178). Findings (Children) In contrast to adults, children did not show the two-way interaction between stimulus gender and stimulus type on our measure of reaction time ( b = −19.91, SE = 23.06, t = −.86, p = .388), and there was no moderation of this effect by children’s gender ( b = 50.22, SE = 47.10, t = 1.07, p = .287) or age ( b = −7.29, SE = 20.12, t = −.36, p = .717). However, in probing children’s mis-categorizations—which they made more than twice as many of as adults—we found a significant three-way interaction among stimulus gender, stimulus type, and child gender ( b = 1.51, SE = .63, t = 2.39, p = .017; Figure 5 ), although the two-way interaction without child gender was not significant ( b = −.54, SE = .30, t = −1.80, p = .071). Unpacking this three-way interaction, we observed a two-way interaction between stimulus gender and stimulus type among boy participants ( b = −1.09, SE = .39, t = −2.76, p = .006) but not among girl participants ( b = .42, SE = .49, t = .85, p = .393). Figure 5. Open in a new tab Children’s errors on the speeded categorization task in Study 2a as a function of stimulus gender, stimulus type, and child gender (note: the y-axis has been truncated slightly to highlight the interaction effect). Larger shapes represent group means, small shapes represent individual data points, and error bars represent standard errors. Follow-up tests of the two-way interaction among boys revealed two features of objectification. First, boys made more errors categorizing boy-gendered toys (i.e., mistakenly categorizing them as “kids”) than they did categorizing boy-gendered kids (i.e., mistakenly categorizing them as “toys”; b = −.77, SE = .29, t = − 2.61, p = .009). In other words, boys tended to mistakenly “humanize” boy-gendered toys more often than they “objectified” boy-gendered kids (while not showing the same tendency for girl-gendered stimuli; b = .32, SE = .26, t = 1.22, p = .221). Second, boys made more errors categorizing girl stimuli as “kids” than they did categorizing boy stimuli as “kids” ( b = −.75, SE = .29, t = −2.60, p = .009), suggesting that boys found “kid” categorizations to be more readily accessible for images of boys than girls. When we re-ran the three-way interaction model with the mean similarity rating of each kid-toy stimulus pair entered as a random intercept to assess robustness, the three-way interaction remained significant ( b = 1.51, SE = .63, t = 2.39, p = .017). Although this interaction (and the patterns revealed in follow-up tests) are consistent with our theoretical framework, we urge caution in the interpretation of these findings given that they only appeared in half of our sample (i.e., boys). Discussion Study 2a provided initial evidence that adults objectify young children in a gendered manner at the level of their categorizations: adults were quicker to correctly categorize girls depicted as objects relative to both girls depicted as humans and boys depicted as objects, suggesting a girl-object association that is stronger than both a girl-human and a boy-object one (with the latter pattern converging with findings that women’s faces are perceived as more artificial-looking than men’s faces; Balas, 2013). We also found preliminary evidence that young boys engage in gender-differentiated objectification: complementing the patterns in adults, boys—although a small sub-sample—made fewer mistakes when categorizing boys depicted as humans relative to both boys depicted as objects and girls depicted as humans, suggesting a boy-human association that is stronger than both a boy-object and girl-human one (a pattern dovetailing with androcentrism, or the belief that people = male ; Hamilton, 1991). Our findings suggest that this phenomenon may be anchored more closely on a girl-object association in adulthood but arise (at least for boys) from a boy-human association present in childhood—a trajectory that could reflect a range of developmental and/or sociocultural processes, including children’s heightened tendency to center themselves as they navigate tasks (i.e., leading them to anchor on the more proximal human concept) and/or adults’ increased exposure to media that features girls in an objectified manner (e.g., childhood beauty pageants). Given that the patterns from our child study emerged only in (a small sample of) boys, however, we next sought to pursue a more robust test of how categorization-based objectification operates in childhood by seeking to replicate our findings with a larger developmental sample. Study 2b Method Participants We recruited children ( N =182) between the ages of 5 and 10 years via a remote and unmoderated platform for developmental research, which captures both video and survey data from participants ( Rhodes et al., 2020 ) 4 . To account for possible exclusions, we recruited 196 children; a subset was excluded from analyses due to duplicate entries ( n =7) or failure to complete more than 40% of the protocol ( n =7), resulting in a final sample of 182 children 5 . Our sample spanned the full age-range ( M age = 7.59 years, SD age = 1.60, range : 5.05 – 10.95) and was roughly split by gender (54% girls; 46% boys). The majority (>96%) of children were from the U.S., and the remainder were from Canada. Based on parent report, 55% identified as White, 20% as Asian, 15% as multi-racial, 4% as Black, and 4% as Hispanic (1% of parents declined to provide this information). Parents skewed slightly liberal with respect to social issues (from 1-Very Liberal to 7-Very Conservative , M polid = 3.63, SD polid = 1.65, range : 1, 7). Procedure Given the adaptation of the speeded categorization task to a remote format and the subsequent switch to the survey platform Qualtrics (which is integrated within our remote platform for developmental research ( Rhodes et al., 2020 ), a few adjustments to the protocol were necessary (although all stimuli remained the same; see Figure 3 ). Most importantly, since we could not reliably capture categorization speed remotely (e.g., due to connectivity differences in children’s home computers), we focused our analyses on children’s categorization errors. Additionally, as the task did not register verbal input to cue advancement, we flashed each stimulus on the screen for 750 milliseconds—a number informed by the mean and variation of response times for children in Study 2a—and asked children to make a “kid” or “toy” categorization as quickly as possible, using key presses. 6 A random subset of approximately 20% of study videos ( n =34) were coded in Datavyu ( Datavyu Team, 2014 ) by a trained research assistant to ensure that sessions were not subject to interference (e.g., from parents), and roughly half of those ( n =18) were coded by a second research assistant for reliability. Consistent with prior research on our remote platform ( Leshin et al., 2021 ), we observed extremely low rates of interference (<1% of critical trials), and coder agreement was excellent (100%). Transparency and Openness Hypotheses, methods, and analyses for Study 2b were pre-registered on OSF 7 . Materials, data, and code are also publicly available on OSF. Data were analyzed using R, version 4.3.2 (R Core Team, 2020). Results Analytic Strategy We analyzed children’s categorization errors using mixed generalized regression models from the lme4 package and specified a binomial distribution (0 = correct response, 1 = incorrect response). 8 As pre-registered, we entered stimulus gender (boy, girl; dummy-coded as 0, 1, respectively) and stimulus type (kid, toy; dummy-coded as 0, 1, respectively) as within-subject predictors, allowing them to interact, and added subject ID as a random intercept. In pre-registered exploratory analyses testing moderation by participant characteristics (e.g., child gender or age), we entered these variables as additional between-subjects predictors and allowed them to interact. We report unstandardized beta coefficients from our main models, and we follow up on all significant interactions with simple-effects tests from the emmeans package. Findings We observed a significant interaction between stimulus gender and stimulus type on children’s categorization errors—an effect that, as in Study 2a, was moderated by children’s own gender. To preview the central finding: Boys, but not girls, showed two features of objectification on this task. First, boys made more errors when categorizing girl-gendered kids than when categorizing girl-gendered toys ( b = .85, SE = .17, t = 4.90, p < .001), suggesting that—like adults in Study 2a—“toy” categorizations were more accessible than “kid” categorizations for images of girls. Second—and replicating one of the features of objectification observed among boy participants in Study 2a—boys made more errors when categorizing girl-gendered kids than when categorizing boy-gendered kids ( b = −.47, SE = .16, t = −2.94, p = .003), suggesting that “kid” categorizations were more accessible for images of boys than girls. The above patterns were reflected in a three-way interaction among stimulus gender, stimulus type, and child gender ( b = .82, SE = .37, SE = 2.21, p = .027; Figure 6 ). Model comparisons confirmed that this three-way interaction model fit the data better than the two-way interaction model that excluded child gender ( X 2 (4) = 15.43, p = .004). Parsing the 3-way interaction by children’s gender, we observed the two-way interaction among boys ( b = −.77, SE = .25, t = −3.07, p = .002) but not girls ( b = .05, SE = .27, t = .19, p = .852), as in Study 2a. Children’s age did not moderate the two-way interaction between stimulus gender and stimulus type ( b = −.16, SE = .13, t = −1.18, p = .237), nor did it moderate the three-way interaction with child gender ( b = −.005, SE = .27, t = −.02, p = .985). However, we did observe a main effect of age ( b = −.43, SE = .08, t = −5.49, p < .001), such that children’s errors decreased with age, and the lower-level interaction between stimulus gender and stimulus type was significant within the elaborated model ( b = −.51, SE = .21, t = −2.42, p = .016)—suggesting that children at the mean age of the sample (around 7.5 years) showed the key two-way interaction. Figure 6. Open in a new tab Children’s errors on the speeded categorization task in Study 2b as a function of stimulus gender, stimulus type, and children’s gender (note: the y-axis has been truncated slightly to highlight the interaction effect). Larger shapes represent group means, small shapes represent individual responses, and error bars represent standard errors. As in Study 2a, we conducted a robustness check to confirm that differences in the perceived similarity of the matched toy-kid pairs were not driving our patterns by re-running our best-fitted model (i.e., including stimulus gender, stimulus type, and children’s gender as predictors) with the mean similarity rating of each pair entered as a random intercept. In this model, the 3-way interaction among stimulus type, stimulus gender, and children’s gender remained significant ( b = .82, SE = .36, t = 2.29, p = .022), as did the key two-way interaction between stimulus type and stimulus gender ( b = −.98, SE = .25, t = −3.95, p < .001). Discussion In Study 2b, we observed that boys—like adults and boys in Study 2a—showed gender-differentiated patterns of objectification in their categorizations, making more errors when categorizing girls depicted as humans relative to girls depicted as objects (suggesting a girl-human link that is weaker than a girl-object one, similar to adults in Study 2a) and also making more errors when categorizing girls depicted as humans relative to boys depicted as humans (suggesting a girl-human link that is weaker than a boy-human one, as observed among boys in Study 2a). These findings converge to suggest a particularly weak girl-human link among young boys, complementing the strong boy-human association observed in boys in Study 2a and providing further support for the possibility that categorization-based objectification may follow from boys’ early concepts of who is or is not closely associated with the human concept. We did not observe the predicted patterns among girls despite their making equivalent rates of errors overall, as in Study 2a. It may be that, even as girls are exposed to objectified content, their tendency to perceive women as most prototypical of people ( Lei et al., 2022 ) and/or their own-gender bias (which is stronger than boys’; Dunham et al., 2016) counteracts the development of a weak girl-human association. (Alternatively, girls’ own-gender bias could actually serve to strengthen the girl-object association, to the extent that they value doll-like ideals of appearance; Halim et al., 2024.) Future research will be needed to disentangle these possibilities. Additionally, future investigations should attempt to invoke the broader object category in lieu of the toy sub-category; although doll-like toys are the clearest example of an objectified representation of a human, our investigation leaves open the possibility that children’s associations may reflect a narrower girl-toy association (as opposed to a broader girl-object one). Study 3 We next sought to examine objectification at the level of participants’ attributions, again probing conceptual overlap between concepts of girls/objects and boys/humans . To do so, we developed a list of properties characteristic of objects and a list characteristic of humans—all of which were related to objects’ or humans’ physical being ; i.e., what objects or humans look and act like—and tracked participants’ attributions to girls vs. boys. Focusing on properties linked to physical embodiment had several advantages: they provided a point of connection with classic theorizing on objectification (in particular, the lesser acknowledgement of internal bodily functions as a consequence; Fredrickson & Roberts, 1997 ), they were relatively simple for participants to reason about in the context of child targets (i.e., as compared to more sophisticated properties used in prior work; e.g., depth), and they map onto a dimension of mental life that children can reason about from a young age ( Weisman et al., 2018 ). We note that, although property attributions may intersect with stereotypes, we view this measure as distinct from a test of gender stereotyping. First, many of the properties in this paradigm lacked explicitly gender-stereotypical content (e.g., “yawns”), and—likely relatedly—most have not featured prominently within the traditional gender stereotyping literature (e.g., Eagly & Steffen, 1984 ; Fiske et al., 2002 ; Prentice & Carranza, 2002). Second, to the extent that gender stereotypes do factor into participants’ judgments on the property attribution task, we directly account for this process by statistically controlling for the gender stereotypicality of each property in our analyses. In doing so, we sought to assess objectification in participants’ attributions above and beyond the mere possession of gender stereotypes. Study 3a Method Participants Adult and child participants for Study 3a were the same as those recruited for Study 2a 9 . Procedure As a set-up to the task, participants learned about a substitute teacher who had received notes about students in their class and were asked to help determine which note applied to which child. We showed participants photographs of the children (sourced from the CAFÉ database; LoBou & Thrasher, 2015), one at a time, and paired each with an attribute from the notes (either characteristic of humans—e.g., “Gets really, really hungry before lunch”—or objects—e.g., “Stays still most of the time”). Across nine trials, participants determined whether the child pictured was or was not the child who possessed the property paired with it (the gender of the child paired with each property was counterbalanced across participants). To ensure that the human-like and object-like properties were perceived as intended (i.e., as characteristic of humans and objects, respectively), a small sample of adults from Amazon’s Mechanical Turk rated the properties prior to study administration. All human-like items received high human (and low object) scores, and all object-like items received high object (and low human) scores (we identified one exception to this pattern and tweaked its wording). The reliability of humanness ratings for human-like traits and object-ness ratings for object-like traits (omitting the reworded property) was acceptable (Cronbach’s alpha = .71–86). We also had properties rated for their alignment with gender stereotypes by a separate sample of adults from Prolific; mean responses hovered around the scale mid-point (range: 36.19–63.46), although certain responses were more clearly associated with one gender. For more details, see Supplemental Materials . Transparency and Openness As Study 3a represented our first attempt to examine attribution-based objectification as directed toward child targets, we treated this study as exploratory. Materials, data, and code are publicly available on OSF, and data were analyzed using R, version 4.3.2 (R Core Team, 2020). Results Analytic Strategy We ran mixed generalized linear models using the lme4 package and specified a binomial distribution (with 0 representing “no” responses, i.e., “This is not the child that [property],” and 1 representing “yes” responses, i.e., “This is the child that [property]”). We entered the gender of the stimulus presented (boy, girl; dummy-coded as 0 and 1, respectively) and the type of property paired with it (human-like, object-like; dummy-coded as 0 and 1, respectively) as within-subjects predictors, allowing them to interact, and added the specific property and the participant ID as random intercepts. To test whether our findings held when considering properties’ gender stereotypicality, we ran an additional model in which we entered the mean gender stereotypicality rating for each property as a covariate. In other analyses, we included participant gender (man, woman; dummy-coded as 0 and 1, respectively) in our main model as an additional between-subjects predictor; however, since both samples were relatively small ( N =77 for adults, N =66 for children), these analyses should be interpreted with caution. For our child sample, we conducted an additional analysis probing moderation by (mean-centered) age. In reporting our main effects, we include unstandardized beta coefficients; we follow up on all significant interactions with simple-effects tests from the emmeans package. Findings (Adults) We observed a significant interaction between stimulus gender and property type ( b = 1.01, SE = .32, t = 3.11, p = .002; Figure 7 ) on adults’ attributions 10 . This interaction revealed two key features of objectification: adults attributed human-like properties more to boys than girls ( b = .56, SE = .22, t = 2.60, p = .009), and they attributed more human-like properties than object-like properties to boys ( b = 1.08, SE = .44, t =2.48, p = .013) but not to girls ( b = .07, SE = .43, t = .17, p = .867). When we controlled for the level of gender stereotypicality of each property, the two-way crossover interaction remained significant ( b = 1.35, SE =.34, t = 3.92, p < .001). For plots of attribute-level variation in our patterns, see Supplemental Materials . Figure 7. Open in a new tab Adults’ responses to the property attribution task in Study 3a (A), separated by participant gender (B). Larger shapes represent group means, small shapes represent individual response patterns, and error bars represent standard errors. In subsequent analyses, we observed that participants’ own gender moderated the two-way interaction between stimulus gender and property type ( b = 1.64, SE = .66, t = 2.49, p = .013; Figure 7 ). Specifically, the two-way interaction was significant for women ( b = 1.77, SE = .45, t = 3.90, p < .001) but not men ( b = .13, SE = .48, t = .28, p = .776). Follow-up tests revealed that women showed three key features of objectification: they were more likely to assign human-like properties to boys than to girls ( b = .81, SE = .30, t = 2.68, p = .007), to assign object-like properties to girls than to boys ( b = −.96, SE = .34, t = - 2.86, p = .004), and to assign human-like than object-like properties to boys ( b = 1.58, SE = .50, t = 3.15, p = .002) but not to girls ( b = −.19, SE = .49, t = −.40, p = .690). Findings (Children) In contrast to adults, children did not reveal the predicted interaction between stimulus gender and property type ( b = .11, SE = .35, t = .32, p = .752), with no moderation by children’s gender ( b = .08, SE = .70, t = .12, p = .904) or age ( b = −.32, SE = .29, t = −1.10, p = .270). Discussion In Study 3a, we observed that adults objectified young children based on gender at the level of their attributions: adults were more likely to attribute human-like properties to boys than girls (suggesting a boy-human association that is stronger than a girl-human one) and more likely to attribute human-like than object-like properties to boys (but not girls; suggesting a boy-human association that supersedes a boy-object one, with the same contrast not applying to girls). These patterns held when controlling for the gender stereotypicality of properties, suggesting that this phenomenon is not driven purely by properties’ stereotype content. The strong boy-human link observed (similar that of boys in Study 2a) is convergent with androcentrism—a phenomenon that has been demonstrated in adults across a wide range of experimental tasks (e.g., Bailey et al., 2020; Bailey & LaFrance, 2017). Notably, the key two-way interaction was observed among women (and not men), who additionally showed the tendency to assign object-like properties more to girls than to boys. Although this finding should be interpreted cautiously due to the small sample, it is nonetheless worth considering why it may have emerged: while it was boys (and not girls) who showed patterns of objectification on the categorization task, it is plausible that for a more deliberative measure, personal experiences with—and, perhaps, heightened awareness of—objectification among women contributed to women’s tendency to perceive girls in a more objectified (and boys in a more humanized) manner, although further research will be needed to test this possibility. Importantly, we did not find evidence of attribution-based objectification in children, supporting the possibility that the exposure children receive to objectified depictions of girls/women manifests (at least in boys) in categorizations before it manifests in attributions. In Study 3b, we sought to bear upon the robustness of our significant findings in adults (and the lack thereof in children) by conducting a conceptual replication with larger samples. Study 3b Method Participants Adult participants ( N =160) living in the U.S. were recruited from Prolific. To account for possible exclusions, we recruited a total of 180 adults; a small subset did not complete the study properly ( n =2) or failed to pass attention checks ( n =18), resulting in a final sample of 160 adults. 11 Participants were roughly split by gender (53% women; 47% men) and spanned a broad age-range ( M age = 33.24 years, SD age = 11.94, range : 18–70). Most adults (79%) identified as White; the remaining identified as Asian (9%), Black (6%), Hispanic/Latino (4%), multiracial (1%), or Native American (1%). Our sample skewed politically liberal with respect to social issues (from 1-Very Liberal to 7-Very Conservative, M polid = 2.77, SD polid = 1.78, range : 1, 7). Children ( N =170) between the ages of 5 and 10 years were recruited via a remote platform for developmental research ( Rhodes et al., 2020 ). We ultimately recruited 177 children, and after excluding those with duplicate entries ( n =5) and those identified as nonverbal or non-English speakers ( n =2), arrived at a final sample of 170 children. Children spanned the full age-range ( M age = 7.54 years, SD age = 1.77, range : 5.03 – 10.92) and were roughly split by gender (54% girls; 46% boys). Most children (>98%) were from the U.S., and the remainder were from Canada; based on parent report, 64% of participants identified as White, 18% as multi-racial, 9% as Asian, 5% as Hispanic, 2% as Black (1% of parents did not report this information). Parents of participating children skewed politically liberal with respect to social issues (from 1-Very Liberal to 7-Very Conservative , M polid = 3.03, SD polid = 1.65, range : 1, 7). Procedure We made several methodological adjustments to help clarify the scope and robustness of our Study 3a findings. First, to make race less salient and remove physical cues beyond gender that may be relevant to objectification (e.g., clothing), we used silhouette figures as targets ( Figure 8 ); this change did not erase the concept of race entirely—the silhouettes may still have brought to mind White exemplars (e.g., Ghavami & Peplau, 2013 ), particularly since they were assigned Westernized names—but instead anchored the task on more abstracted concepts of gender (i.e., girls vs. boys rather than White girls vs. White boys ). Second, we allowed for more variation in responses by making the scale ordinal, asking “ How often do you think [child name] [property]?” (response options: “Never,” “Not a lot,” “Sometimes,” “Quite a lot,” or “Always”) 12 . Finally, we eliminated one of the properties used in Study 3a to shorten the protocol and equate the number of human-like and object-like properties ( Figure 8 ; with this item omitted, the reliability of the humanness ratings for the human-like properties in our task remained acceptable, Cronbach’s alpha =.79). Approximately 20% of study videos from children ( n =38) were coded in Datavyu ( Datavyu Team, 2014 ) by a trained research assistant for quality control; we again observed extremely low rates of interference (<2% of critical trials), and agreement with a second coder was high (97%). 13 Figure 8. Open in a new tab Stimuli used in the property attribution task in Study 3b. Transparency and Openness All hypotheses, methods, and analyses reported for Study 3b were pre-registered on OSF. We additionally pre-registered exploratory analyses to test for moderation by participants’ or parents’ beliefs about gender, as well as separate measures that we do not report on here; for details related to these deviations, see Supplemental Materials . Materials, data, and code are publicly available on OSF. Data were analyzed using R, version 4.3.2 (R Core Team, 2020). Results Analytic Strategy We conducted a series of mixed-effects ordinal regression models using the ordinal package to accommodate the ordinal nature of our data. As in Study 3a, our primary model included stimulus gender (boy, girl; dummy-coded as 0 and 1, respectively) and stimulus type (human-like, object-like; dummy-coded as 0 and 1, respectively) as within-subjects fixed effects, which we allowed to interact. Property and subject ID were modeled as random intercepts. We again conducted an additional (non-pre-registered) analysis in which we entered the gender stereotypicality rating of each of our properties as a covariate within our main model to assess robustness. In pre-registered exploratory analyses, we included participant gender (man, woman; dummy-coded as 0 and 1, respectively) and mean-centered child age (for our child sample) as additional between-subjects predictors. We report unstandardized beta coefficients and follow up on all significant interactions using simple-effects tests from the emmeans package. Findings (Adults) Replicating Study 3a and consistent with hypotheses, we observed a significant two-way interaction between stimulus gender and property type among adults ( b = .67, SE = .21, t = 3.13, p = .002; Figure 9 ) 14 . This interaction reflected one key feature of objectification, which we also observed in Study 3a—adults’ tendency to perceive human-like properties as more common in boys than girls ( b = .39, SE = .15, t = 2.67, p = .008)—and again remained significant when we controlled for properties’ gender stereotypicality ( b =.67, SE = .21, t = 3.14, p = .002). Unlike in Study 3a, participants’ own gender did not moderate the key two-way interaction ( b = .19, SE = .43, t = .45, p = .653), suggesting that men and women displayed similarly gendered patterns of objectification. For plots of attribute-level variation, see the Supplemental Materials . Figure 9. Open in a new tab Adults’ responses on the property attribution task in Study 3b. Larger shapes represent group means, small shapes represent individual response patterns, and error bars represent standard errors. Findings (Children) In contrast to adults—and corroborating Study 3a—children did not show the predicted two-way interaction between stimulus gender and property type ( b = .10, SE = .20, t = .50, p = .618). Descriptively, children’s average prevalence ratings were clustered around the scale midpoint ( range : 3.12 – 3.20). We did not observe moderation by children’s age ( b = .05, SE = .11, t = .45, p = .651), counter to hypotheses, nor did we observe moderation by children’s gender ( b = .61, SE = .40, t = 1.54, p = .124) or the interaction between the two ( b = .38, SE = .23, t = 1.67 p = .094). Discussion Conceptually replicating Study 3a, Study 3b revealed gendered patterns of objectification in adults’ attributions—effects that (again) held when controlling for properties’ gender stereotypicality. This overall pattern was driven by adults’ tendency to rate human-like properties as more common in boys than girls (suggesting a boy-human association that supersedes a girl-human one, as in Study 3a). Importantly, these patterns emerged in judgements of children who were gendered only by their silhouettes and names, suggesting that gender differentiation in objectification can emerge even when gender itself is represented abstractly. However, our use of abstract stimuli may have also obscured our ability to detect differentiation in the human-object distinction (e.g., the tendency to perceive a stronger boy-human than boy-object association, as in Study 3a) precisely because the stimuli themselves were far less human-like in appearance. Also replicating Study 3a, we did not find gendered patterns of objectification in children’s attributions. A key difference between Study 3b and 3a, however, was the lack of moderation by participant gender: given the larger sample size of Study 3b relative to 3a (160 vs. 77 participants), we are inclined to believe that the effect of participant gender from Study 3a was not reliable and that the key two-way interaction is consensual across both men and women. However, it may be the case that methodological differences contributed to this inconsistency: for instance, to the extent that women’s own experiences contribute to perceived objectification (e.g., in Study 3a), the abstract silhouettes in Study 3b may have minimized women’s tendency to project or incorporate their own accounts into their attributions—a possibility we bear upon further in Study 3c. In addition, Study 3c examines whether the patterns from Study 3a and 3b extend to a broader set of properties (conceived of in a bottom-up, data-driven manner) and vary by stimulus race, providing an extended conceptual replication of our attribution results. Study 3c Method Participants Participants ( N =193) were recruited from Prolific; all adult users who live in the U.S. and are fluent in English were eligible to participate. We pre-registered a sample size of 180 participants based on the key two-way interaction from Study 3b ( OR = .196). To be able to detect an interaction effect of similar size at 80% power required a sample of 138 participants ( Faul et al., 2007 ); however, because we intended to control for property gender stereotypicality in primary analyses, we increased our desired sample to 180 and recruited a total of 200 to account for exclusions. We excluded a small subset for failure to complete the study properly ( n =4) or pass an audio check ( n =3), resulting in a total of 193 participants. 15 Roughly half of our sample identified as women (49%) or men (50%); others identified as non-binary or a third gender (1%) or declined to answer (0.5%). Participants spanned a wide age-range ( M age = 39.93 years, SD age = 12.36, range : 18–72). Most identified their race as White (65%); 10% identified as multi-racial, 8% identified as Hispanic or Latino, 8% identified as Asian, 6% identified as Black, 1% identified as Native American, and 1% identified as Other. Participants varied with respect to political ideology, with a slight skew toward the liberal end of the spectrum (from 1-Very Liberal to 7-Very Conservative, M polid = 3.15, SD polid = 1.59, range : 1–7). Procedure Prior to conducting Study 3c, we administered an item generation study with a small sample of adults (also recruited from Prolific). In this study, we asked participants to brainstorm as many attributes as they could that describe the physical being of humans (but not that of objects) and then to brainstorm as many as they could that describe the physical being of objects (but not that of humans). We selected the 11 properties that appeared most frequently in descriptions of humans (frequencies > 5) and objects (frequencies > 3; we cut off selection at 22 items to keep the protocol to a reasonable length). For additional validation, we had all properties rated by a separate sample of adults for their level of humanness and object-ness, which confirmed the initial classifications and yielded good internal reliability (for human-like properties, Cronbach’s alpha = .80; for object-like properties, Cronbach’s alpha = .82). The new properties were also rated for their gender stereotypicality; mean responses again hovered around the middle of the scale (range: 37.30–71.78), with certain properties standing out as more typical of one gender than the other. For more information, see Supplemental Materials . The Study 3c protocol functionally resembled that of Studies 3a and 3b, with several key modifications. Importantly, we more explicitly probed adults’ perception of girls and boys as central to (i.e., prototypical of) the concepts of objects and human s by adapting language from classic studies of category prototypes (i.e., Mervis & Rosch, 1981 ). That is, we presented participants with a property and asked them to consider who first comes to mind or who comes to mind as the best example of it (similar primes have been used in prior work assessing androcentrism; e.g., Bailey et al., 2017; Lei et al., 2022 ). Participants were then shown an array of six children—rather than one child at a time—and asked to select the one that most closely matches the mental image that emerged for the property ( Figure 10 ). The six faces presented, which were created using a novel AI tool (see Supplemental Materials ), varied in both gender (boy, girl) and race (White, Black, and Asian). In expanding the racial diversity of stimuli, we sought to obtain a more holistic picture of the conceptual overlap that exists between girls/objects and boys/humans (beyond what the abstract stimuli in Study 3b afforded). Figure 10. Open in a new tab Left (slides and text): An example trial from the property attribution task in Study 3c. Right (list of properties): Properties used in Study 3c. Transparency and Openness Hypotheses, methods, and analyses for Study 3c were pre-registered on OSF. Materials, data, and code are also publicly available on OSF. Data were analyzed using R, version 4.3.2 (R Core Team, 2020). Results Analytic Strategy We implemented a mixed generalized linear regression model using the lme4 package 16 . In our first pre-registered analysis, we included stimulus gender (boy, girl; dummy-coded as 0 and 1, respectively) and property type (human-like, object-like; dummy-coded as 0 and 1, respectively) as within-subjects predictors, allowing them to interact, and modeled participant ID as a random intercept. In our second pre-registered analysis, we added properties’ gender stereotypicality rating as a covariate. We subsequently conducted an exploratory (non-pre-registered) analysis in which we only included properties that achieved high frequency ratings in our item generation study to determine whether the properties most strongly associated with objects or humans yielded stronger effects. In additional pre-registered exploratory analyses, we tested for moderation by stimulus race (White, Black, Asian) and participant gender (man, woman; dummy-coded as 0 and 1, respectively) by adding each, separately, as an interacting between-subjects predictor. We report unstandardized beta coefficients from our main models and follow up on significant interactions with simple-effects tests using the emmeans package. In exploratory analyses of stimulus race, involving a non-dichotomous predictor, we follow up on our model by using the Anova function from the car package and report Wald X 2 values. Findings We observed a significant two-way interaction between stimulus gender and property type on participants’ attributions ( b = .30, SE = .07, t = 4.40, p < .001; Figure 11 ). This interaction revealed three features of objectification: adults attributed more human-like properties to boys than girls ( b =.32, SE =.05, t =6.58, p <.001), more human-like than object-like properties to boys ( b =.14, SE =.05, t =3.03, p =.003), and more object-like than human-like properties to girls ( b =−.16, SE =.05, t =−3.20, p =.001). The key two-way interaction again persisted when controlling for properties’ gender stereotypicality ( b = .30, SE = .07, t = 4.41, p <.001). Stimulus race did not moderate the key two-way interaction ( p >.080; but see Supplemental Materials for lower-order interactions with this variable), nor did participant gender ( p >.065). For plots of attribute-level variation, see the Supplemental Materials . Figure 11. Open in a new tab Adults’ responses on the property attribution task in Study 3c for the full set of properties tested (A) and for the five properties in each category that were generated with the greatest frequency (B). Larger shapes represent group means, small shapes represent individual response patterns, and error bars represent standard errors. In exploratory analyses, we considered the possibility that the lack of gender differentiation for object-like properties (i.e., the one feature of objectification we did not observe) could have arisen from the sheer breadth of properties selected for the task, including those that were generated somewhat infrequently (e.g., fewer than 5 times). To investigate this possibility, we re-ran our primary analysis on a subset of data that included the five properties generated most frequently for each category (for human-like properties, frequencies > 10; for object-like properties, frequencies > 5). The key two-way interaction remained ( b= .35, SE =.10, t =3.45, p <.001, including when controlling for gender stereotypicality: b= .35, SE =.10, t =3.45, p <.001; Figure 11 ), and this time, we observed all four forms of objectification. That is, adults attributed human-like properties more to boys than girls ( b =.18, SE =.07, t =2.54, p =.011) and object-like properties more to girls than boys ( b =−.17, SE =.07, t =−2.33, p =.020); they also attributed more human-like than object-like properties to boys ( b =.17, SE =.07, t =2.44, p =.015) and more object-like than human-like properties to girls ( b =−.17, SE =.07, t =−2.44, p =.015). Discussion Study 3c revealed further evidence of gender-differentiated objectification at the level of adults’ attributions: adults attributed more human-like properties to boys than girls (as in both Study 3a and 3b), more human-like than object-like properties to boys, and more object-like than human-like properties to girls. In an exploratory (and non-pre-registered) analysis on the subset of properties generated most frequently in our item generation task, we found evidence for a fourth form of bias: in this subset, adults attributed object-like properties more to girls than boys. The presence of multiple forms of objectification observed here is more consistent with Study 3a (i.e., in which we observed two forms of objectification in the full sample and three when we subset to women only) and less consistent with Study 3b (i.e., in which we observed only one form of objectification). This observation lends support to the possibility that more concrete and realistic depictions of children (e.g., photographs in Studies 3a and 3c) may trigger more diverse forms of objectification to emerge in adults’ attributions as compared to more abstract depictions (e.g., silhouettes in Study 3a). Particularly given that the properties adults were asked to reason about were anchored on physical embodiment , it is intuitive that distinctions between object-like and human-like properties—which patterned in a gendered manner in Studies 3a and 3c but not Study 3b—would be more salient for stimuli depicting actual human children. We found no interactive effects of participant gender, corroborating Study 3b and lending support to the notion that gendered patterns of objectification in property attributions are consensual (rather than manifesting more in women, perhaps as a projection of their lived experiences). We also did not observe an interaction with stimulus race, suggesting that adults’ selective overlapping of girls with objects and boys with humans on this task extends beyond White girls and White boys (and, at least, to Asian boys and girls and Black boys and girls ). General Discussion The present studies suggest that adults—and, in some cases, young boys—objectify children in a gendered manner. Adults did so both on our measure of categorization (Study 2a) and property attribution (Study 3a, 3b, and 3c), whereas boys did so only in their categorizations (Study 2a and 2b). Together, our findings lend support to the possibility that young children may be susceptible to gendered patterns of objectification as targets —a matter that, despite the severe and well-documented consequences of objectification, has yet received little empirical attention. The particular form of objectification displayed by adults varied across measures (and, to some extent, across studies). For the categorization task (Study 2a; not pre-registered), adults provided correct object categorizations for girls more quickly than correct human categorizations for girls (suggesting a girl-object association that is stronger than a girl-human one) and correct object categorizations for boys (suggesting a girl-object association that is stronger than a boy-object one). These findings are consistent with traditional views of objectification from psychology ( Fredrickson & Roberts, 1997 ) that highlight the addition of object-like attributes to girls and women. In contrast, the patterns revealed in the attribution task were driven, largely, by the relative absence of an association between girls and humans : all three interactions observed on this measure (Studies 3a, 3b, and 3c; 3b and 3c were pre-registered) reflected adults’ tendency to associate human-like properties less with girls than with boys—suggesting a girl-human association that is weaker than a boy-human one. This finding is consistent with more recent conceptualizations of objectification that connect it to the denial of core human-like characteristics ( Heflick & Goldenberg, 2014 ; Loughnan et al., 2010 ), as well as theoretical (Bailey et al., 2019) and empirical (e.g., Bailey et al., 2020) accounts of androcentrism; thus, to the extent that gender-differentiated objectification entails a disconnect between girls and humans , it may manifest as complementary to (or intertwined with) this latter process. Notably, we observed forms of objectification beyond the girl-human vs. boy-human contrast on our property attribution tasks (e.g., in Study 3c, adults’ girl-object association superseded their girl-human one), suggesting that androcentrism alone does not account for our pattern of results. In contrast to adults, children only exhibited gendered patterns of objectification in their categorizations (Study 2a and 2b), and with moderation by child gender: boys, but not girls, displayed this pattern. In our most robust test of this question (Study 2b; pre-registered), the forms of objectification that drove the key two-way interaction overlapped, in part, with adults’ patterns on this measure: that is, boys similarly revealed a girl-object association that superseded a girl-human association. However, boys also demonstrated a boy-human association that is stronger than a girl-human one (mirroring boys in Study 2a). This latter form of objectification did not drive adults’ patterns on the categorization task, but it did drive their patterns on the attribution measure, as described above. Thus, it is possible that the boy-human vs. girl-human association takes root at the level of categorizations for boys in childhood, and then in adulthood, manifests in both men’s and women’s more explicit judgements. Whether girls’ categorizations develop in a manner similar to that of boys (i.e., anchoring on associations with humans ) before taking the form we observed in adults (i.e., anchoring on associations with objects ) remains an empirical question that should be tested in future research. Critically, children did not show the predicted patterns on our property attribution measure (Study 3a and 3b), suggesting that gender differentiation in this process emerges later in development. Importantly, the consistent pattens of gender differentiation observed among adults on our property attribution measures did not appear driven by gender stereotypes: when controlling for the gender stereotype content of each property—an analysis we pre-registered in Study 3c—the two-way interaction between child gender and property type persisted. This finding suggests that, while properties perceived as characteristic of objects (e.g., stays still) may intersect with stereotypes of women and those perceived as characteristic of humans (e.g., has a smell) may intersect with stereotypes of men, this alignment itself does not account for the conceptual overlap documented here. In other words, the phenomenon of conceptually overlapping the categories of girls with objects and boys with humans appears to be separable from—and occurring over and above—people’s mere possession of stereotypes that link girls with objects and boys with humans. This possibility is further supported by the fact that a subset of the properties used in Study 3c were not perceived as stereotypical of either gender. Nonetheless, it is possible that individual participants’ support for gender stereotypes (or endorsement of other prescriptive gender beliefs) may have accounted for variation on this measure; future research should examine which such beliefs, if any, predict variation on attribution-based objectification. Our findings offer nuance to current understandings of objectification, most importantly suggesting that this process patterns along gender lines even for child targets: that is, much as women are objectified relative to men, so too are girls objectified relative to boys. This finding challenges the notion that sexual cues or motives are a necessary component of such gender differentiation (e.g., Bernard et al., 2015 ; Fredrickson & Roberts, 1997 ; Loughnan et al., 2010 ), which prior research on the objectification of prepubescent girls has ultimately reaffirmed (i.e., finding that girls dressed in a sexualized manner—but not those dressed neutrally—elicited objectifying perceptions by adults; Graff et al., 2012 ; Holland & Haslam, 2016 ). Our findings, however, suggest that even young girls presented in a neutral, non-sexual manner—shown in everyday clothing or as bare silhouettes—elicit objectification relative to young boys. These findings are broadly consistent with a key premise of objectification: that the driving force behind gender differentiation in this domain is not a motivated sexual gaze, per se, but a more general focus on physical appearance. Indeed, we found that the heightened appearance focus directed toward women (vs. men) also extends to girls (vs. boys; Study 1; not pre-registered). Two features of our methods may have contributed to their ability to detect such gendered patterns. First, we presented a clear gender contrast across or within trials on each task. Although this design choice may seem obvious in light of our research questions, many other paradigms do not manipulate gender at all—instead focusing on distilling the effects of other relevant physical cues in women or girls alone (e.g., Bernard et al., 2015 ; Graff et al., 2012 ; Holland & Haslam, 2016 )—or manipulate both gender and these other features at the same time ( Gray et al., 2011 ; Loughnan et al., 2010 ; Vaes et al., 2011 ). In many such cases, objectification is observed most strongly toward girls or women who embody the manipulated physical cues. However, in the absence of an isolated gender manipulation, past findings risk masking the potent force that gender itself plays in objectification, instead perhaps encouraging participants to “anchor and adjust” on a certain type of woman or girl stimulus (e.g., one that is highly sexualized) and then default to lower levels of objectification for all others in comparison. Thus, in manipulating stimulus gender and no other physical cues, our measures were equipped to detect even small differences in the extent to which people objectify children along gender lines. A second feature that may have contributed to our sensitivity to detect the predicted effects was our use of properties linked to the physical being of humans vs. objects in our attribution tasks. Detecting gender differentiation in these properties may be less reliant on additional physical cues that were, intentionally, absent from our paradigm (e.g., sexualization); indeed, whereas judgements of women’s competence are sensitive to how women are dressed ( Glick et al., 2005 ), gender-differentiated intuitions about physiology appear to be rooted more firmly in the more basic distinction between males and females ( Goldenberg & Roberts, 2004 ). Thus, in anchoring this measure on properties that solicit less within-gender (and more between-gender) variation, our design may have been particularly well-suited to capture our patterns of interest. Our finding of gender differentiation in boys’ categorizations suggests that the tendency to overlap girls/objects and boys/humans has its roots early in development for boys, converging with the stable male-centered bias found in boys’ concepts of people ( Lei et al., 2022 ). Notably, girls did not show any evidence of gender-differentiated objectification, despite typically higher levels of self -objectification (e.g., Daniels et al., 2020 ). It may be that girls’ tendency to center their own gender in their concepts of people —which declines with age but still remains prevalent in this age-range ( Lei et al., 2022 )—or strong own-gender bias (Dunham et al., 2016) negated what may otherwise manifest as a girl/object association. It is also possible, however, that our failure to capture these effects in girls is an artifact of our methodology: to the extent that girls had more first-hand experience playing with dolls similar to those representing objects in our categorization task, they may have developed a sharper ability to differentiate girls depicted as humans from those depicted as objects —a differentiation boys struggled to do as successfully. In demonstrating that girls may be targets of objectification (and its consequences) earlier than previously assumed, our work also offers practical insights. For example, to the extent that children’s gender licenses inferences about their physiological properties, young girls may be subject to similar harms as women in the domain of health (e.g., the tendency for medical practitioners to ignore women’s—more so than men’s—reports of pain; Samulowitz et al., 2018). More broadly, the minimization of girls’ physiological experiences may be one (although likely not the only) component of the culture of shame and disgust that surrounds feminine biological processes emerging in adolescence (e.g., menstruation; McHugh, 2020). Moreover, that we found evidence of gender-differentiated objectification in young boys (in their categorizations) suggests that the sociocultural processes that create and sustain associations between girls and objects (e.g., media) begin to leave their mark early on—potentially rendering young girls vulnerable to certain forms of objectification not just by adults, but by other (boy) children. Although our findings offer relevant theoretical and practical insights, they represent only a first step in understanding the processes by which young children come to be objectified along gender lines. As a next step, it will be useful for future research to seek to replicate our patterns and contextualize them within prior operationalizations of objectification. For example, future investigations could adapt prior measures of non-sexual objectification for use with child stimuli (e.g., developing measures to probe perceptions of a child’s human “essence”) to determine whether child targets elicit gendered patterns of objectification across a broader array of measures than those tested here. Relatedly, it will be informative for future work to investigate how the conceptual overlap we observed here might relate to objectification targeted toward adults (e.g., the sexual objectification of women relative to men), Finally, given the theoretical overlap between objectification and dehumanization, it will be important to test whether the patterns captured here would be corroborated with methods used to probe dehumanization more generally. That is, do the patterns of conceptual overlap we documented suggest that girls are dehumanized broadly (including with respect to other non-human categories, like animals ), or does this phenomenon occur only when the contrast category invoked is objects ? Finally, an important priority for future research will be to more closely examine the intersecting influences of gender and race in the objectification of young children. First, it will be informative to examine the extent to which the patterns documented extend to samples with greater racial-ethnic diversity; while our samples included participants from a range of racial backgrounds, we did not recruit adequate numbers of participants of color to systematically examine the role of participant race in our findings. Second, it will be important for future studies to diversify the stimulus set used to test objectification. Our findings suggest that attribution-based objectification manifests similarly across racial groups—Study 3c provided the strongest test of this—but because our measure of categorization-based objectification included only White stimuli (due, in large part, to the dearth of doll-like toys depicted as non-White), the extent to which race shapes objectification at the level of participants’ categorizations remains unknown. This is an important question given the large body of research suggesting that objectification is greater for women of color ( Anderson et al., 2018 ; Mosley et al., 2023 )—and that dehumanization is greater for children of color ( Goff et al., 2014 )—as well as the wealth of findings indicating that race biases representations of gender more broadly ( Cole, 2009 ; Crenshaw, 1989 ; Goff et al., 2008 ; Johnson et al., 2012 ; Purdie-Vaughns & Eibach, 2008 ). Such questions are ripe for investigation with both adults and children, whose concepts of women begin to incorporate race information in early childhood ( Lei et al., 2020 ; Leshin et al., 2022 ). Conclusion Across five studies, we tested the extent to which adults and children tend to objectify children in a gender-differentiated manner, selectively overlapping their concepts of girls with objects and boys with humans . Adults showed evidence of the predicted conceptual overlap in their categorizations and attributions, while children showed more limited evidence of such overlap. Altogether, our research converges to suggest that the gendered nature of objectification manifests earlier than previously assumed—that is, in our perceptions of girls and boys . Supplementary Material Graphics and text NIHMS2063855-supplement-Graphics_and_text.docx (4.6MB, docx) Table 1. Summary of key limitations of this research. Table of Limitations Did not assess convergence between measures of objectification tested here and prior measures of objectification (e.g., toward adults) or measures of related phenomena (e.g., dehumanization) Attribution-based objectification measures assessed physiological attributes only, limiting our grasp of how this process manifests in other attribute domains Measures of categorization-based objectification included only White stimuli; thus, we cannot be certain of how this process manifests toward girls and boys from other racial groups Studies of categorization- and attribution-based objectification toward child targets would benefit from more diverse (and larger) participant samples to clarify robustness of effects Did not examine the extent to which aspects of participants’ prescriptive beliefs (or descriptive beliefs beyond stereotyping) may have contributed to our pattern of results Use of overlapping samples to assess multiple measures related to objectification limits the generalizability of our findings Open in a new tab Acknowledgments: We thank the families on the Princeton & NYU Discoveries in Action site for partnering in this research. Footnotes 1 The Study 1 sample overlapped with those of two other studies reported in this manuscript, and the Study 1 measure always came last; however, since our measures were quite distinct from one another and had very different response formats (and since none sought to systematically induce objectification), we reason that the possibility of problematic carryover effects is minimized. For more information, see Supplemental Materials . 2 Our sample for Study 2a overlaps with that of another study reported in this manuscript; in all instances, Study 2a was administered first. 3 Outliers were eliminated prior to analyses. For reaction time data (which contained only trials that yielded accurate categorizations), we replaced all trials below 300 milliseconds with 300 milliseconds (adult data: n =6, 0.21% of total trials; child data: n =2, .09% of total trials) and removed all trials that exceeded the mean reaction time by more than 2.5 standard deviations (adult data: n =72 trials, 2.55% of total trials; child data: n =77, 3.31% of total trials), following protocols from prior work using similar methodology; Lei et al., 2020 ; Leshin et al., 2022 ). For error data, we computed the mean number of incorrect categorizations made across stimuli and assessed whether the number of errors elicited by any one stimulus exceeded that value by more than 2.5 standard deviations. One stimulus fell above this threshold in both the adult and child datasets and was thus removed from all error analyses. 4 Study 2b was conducted during the Covid-19 pandemic, and thus in-person administration was not possible. 5 We had pre-registered a sample size of 160 children based on a power analysis of a preliminary effect obtained in Study 2a; however, because this effect was identified before we removed the outlier stimulus, we opted to conduct a post-hoc sensitivity analysis to further assess the robustness of our effects. Using a simulation-based sensitivity analysis ( mixedpower ; Kumle et al., 2021 ), we identified the sensitivity of our analysis to detect the key two-way interaction as .898—indicating that this focal effect was significant in nearly 90% of our simulations. 6 To ensure that our adapted measure would capture variation in children’s responses, we pre-registered a plan to briefly check the responses of the first 60 children we recruited. Descriptive analyses revealed adequate variation across children; thus, we proceeded to collect data from the full sample. 7 Analyses used to determine stimulus outliers and to test for the influence of study version (i.e., left-right orientation of keys used for providing a category label) were not pre-registered and are marked as such below. We additionally pre-registered analyses probing the influence of parents’ beliefs on children’s categorizations; for more details, see Supplemental Materials . 8 We tested for outliers in the same manner as in Study 2a and identified the same stimulus as an outlier; accordingly, we removed all trials containing this stimulus from analyses. We also tested for the role of the left-right orientation of category labels and found no main ( p = .849) or interactive ( p s > .260) effects; thus, we did not consider this further in our analyses. The latter two analyses were not pre-registered. 9 Study 3a and Study 2a were completed in the same experimental session, and Study 3a always followed Study 2a. 10 This two-way interaction remained significant when object-like items phrased negatively (i.e., “Never sneezes” and “Never sweats”) were re-categorized as human-like items and ratings were reverse-coded ( b = 1.09, SE = .39, t = 2.82, p = .005). 11 We had pre-registered a sample size of 160 participants for each age-group based on a power analysis of an effect from Study 1. Upon further developing our theoretical framework, however, we now view this effect as a precursor to (rather than direct manifestation of) objectification. Thus, to obtain a more precise and reliable estimate of the robustness of our effects, we conducted a post-hoc sensitivity analysis using a simulation-based tool ( mixedpower ; Kumle et al., 2021 ). This analysis revealed that the sensitivity of our main analysis to detect the key two-way interaction between stimulus gender and property type in adults is .927 (i.e., revealing that this effect was significant in more than 90% of simulated analyses). 12 As a result of this change, items from Study 3a that themselves indicated prevalence (e.g., “Never sweats,” “Never smells”) were reworded (i.e., “How often do you think [child name] sweats?”) and then reverse-coded in analyses to reflect the original association with humans or objects. 13 For both adults and children in Study 3b, the property attribution task was the first measure completed in a given study administration session. Following this task, participants completed two other measures that we do not report on in the manner we had pre-registered. For details on these measures, see the Supplemental Materials . 14 This two-way interaction remained significant when items that were reverse-coded for analyses (“Sweats” and “Smells”; Figure 7 ) were not reverse-coded and re-categorized as human-like properties ( b = .66, SE = .25, t = 2.63, p = .009). 15 Post-hoc simulation-based sensitivity analyses ( mixedpower; Kumle et al., 2021 ) indicated that the sensitivity of our analyses to detect the key two-way interaction was .988, revealing that this effect was significant in nearly 100% of simulated analyses. 16 Before doing so, we restructured our data such that each trial was constituted in six rows of data, with each row representing one of the six stimuli presented; we then created a column to reflect stimulus gender and another to reflect stimulus race (in exploratory analyses). In this dataset, our dependent measure followed a binomial distribution, with 0 representing the stimuli that were not selected in a given trial and 1 representing the stimulus that was selected (for a similar analytic approach, see Lei et al., 2022 ). References Anderson JR, Holland E, Heldreth C, & Johnson SP (2018). Revisiting the Jezebel Stereotype: The Impact of Target Race on Sexual Objectification. Psychology of Women Quarterly, 42(4), 461–476. 10.1177/0361684318791543 [ DOI ] [ Google Scholar ] Auster CJ, & Mansbach CS (2012). The gender marketing of toys: An analysis of color and type of toy on the Disney store website. Sex roles, 67, 375–388. [ Google Scholar ] Bernard P; Gervais SJ, Allen J, Campomizzi S, and Klein O (2012). Integrating Sexual Objectification with Object Versus Person Recognition: The Sexualized-Body-Inversion Hypothesis, Faculty Publications, Department of Psychology. 568. http://digitalcommons.unl.edu/psychfacpub/568 [ DOI ] [ PubMed ] [ Google Scholar ] Bernard P, Gervais S, Allen J, Delmée A, and Klein O (2015). From Sex Objects to Human Beings: Masking Sexual Body Parts and Humanization as Moderators to Women’s Objectification. Faculty Publications, Department of Psychology. 674. http://digitalcommons.unl.edu/psychfacpub/674 [ Google Scholar ] Blakemore JEO (2003). Children’s beliefs about violating gender norms: Boys shouldn’t look like girls, and girls shouldn’t act like boys. Sex roles, 48, 411–419. [ Google Scholar ] Blakemore JEO & Centers RE (2005). Characteristics of boys’ and girls’ toys. Sex Roles, 53, 619–633. [ Google Scholar ] Clark L, & Tiggemann M (2006). Appearance culture in nine‐to 12‐year‐old girls: Media and peer influences on body dissatisfaction. Social Development, 15(4), 628–643. [ Google Scholar ] Clark L, & Tiggemann M (2007). Sociocultural influences and body image in 9 to 12-year-old girls: The role of appearance schemas. Journal of clinical child and adolescent psychology, 36(1), 76–86. [ DOI ] [ PubMed ] [ Google Scholar ] Cole ER (2009). Intersectionality and research in psychology. American psychologist, 64(3), 170. [ DOI ] [ PubMed ] [ Google Scholar ] Costello K, & Hodson G (2014). Explaining dehumanization among children: the interspecies model of prejudice. The British journal of social psychology, 53(1), 175–197. 10.1111/bjso.12016 [ DOI ] [ PubMed ] [ Google Scholar ] Crenshaw K (1989). Demarginalizing the intersection of race and sex: A Black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. The University of Chicago Legal Forum, 1989, Article 8. Retrieved from https://chicagounbound.uchicago.edu/uclf/vol1989/iss1/ [ Google Scholar ] Daniels EA, Zurbriggen EL, & Monique Ward L (2020). Becoming an object: A review of self-objectification in girls. Body image, 33, 278–299. 10.1016/j.bodyim.2020.02.016 [ DOI ] [ PubMed ] [ Google Scholar ] Datavyu Team (2014). Datavyu: A Video Coding Tool. Databrary Project, New York University. URL http://datavyu.org . [ Google Scholar ] Eagly AH, Nater C, Miller DI, Kaufmann M, & Sczesny S (2020). Gender stereotypes have changed: A cross-temporal meta-analysis of U.S. public opinion polls from 1946 to 2018. American Psychologist, 75(3), 301–315 [ DOI ] [ PubMed ] [ Google Scholar ] Eagly AH, & Steffen VJ (1984). Gender stereotypes stem from the distribution of women and men into social roles. Journal of Personality and Social Psychology, 46(4), 735–754. [ Google Scholar ] Faul F, Erdfelder E, Lang AG, & Buchner A (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior research methods, 39 ( 2 ), 175–191. [ DOI ] [ PubMed ] [ Google Scholar ] Fiske ST, Cuddy AJC, Glick P, & Xu J (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82(6), 878–902. [ PubMed ] [ Google Scholar ] Fredrickson BL, & Roberts T-A (1997). Objectification Theory: Toward Understanding Women’s Lived Experiences and Mental Health Risks. Psychology of Women Quarterly, 21(2), 173–206. 10.1111/j.1471-6402.1997.tb00108.x [ DOI ] [ Google Scholar ] Fredrickson BL, Roberts T-A, Noll SM, Quinn DM, & Twenge JM (1998). That swimsuit becomes you: Sex differences in self-objectification, restrained eating, and math performance. Journal of Personality and Social Psychology, 75(1), 269–284. 10.1037/0022-3514.75.1.269 [ DOI ] [ PubMed ] [ Google Scholar ] Gervais SJ, Bernard P, Klein O, & Allen J (2013). Toward a unified theory of objectification and dehumanization. Nebraska Symposium on Motivation. Nebraska Symposium on Motivation, 60, 1–23. 10.1007/978-1-4614-6959-9_1 [ DOI ] [ PubMed ] [ Google Scholar ] Ghavami N, & Peplau LA (2013). An intersectional analysis of gender and ethnic stereotypes: Testing three hypotheses. Psychology of Women Quarterly, 37(1), 113–127. [ Google Scholar ] Glick P & Fiske ST (1996). The ambivalent sexism inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70 ( 3 ) , 491–512. [ Google Scholar ] Glick P, Larsen S, Johnson C, & Branstiter H (2005). Evaluations of Sexy Women In Low- and High-Status Jobs. Psychology of Women Quarterly, 29(4), 389–395. 10.1111/j.1471-6402.2005.00238.x [ DOI ] [ Google Scholar ] Goff PA, Jackson MC, Di Leone BAL, Culotta CM, & DiTomasso NA (2014). The essence of innocence: consequences of dehumanizing Black children. Journal of personality and social psychology, 106(4), 526. [ DOI ] [ PubMed ] [ Google Scholar ] Goff PA, Thomas MA, & Jackson MC (2008). “Ain’t I a woman?”: Towards an intersectional approach to person perception and group-based harms. Sex Roles: A Journal of Research, 59(5–6), 392–403. 10.1007/s11199-008-9505-4 [ DOI ] [ Google Scholar ] Goldenberg JL, & Roberts TA (2004). The beast within the beauty. Handbook of Existential Experimental Social Psychology, 71–85. [ Google Scholar ] Graff K, Murnen SK, & Smolak L (2012). Too sexualized to be taken seriously? Perceptions of a girl in childlike vs. sexualizing clothing. Sex Roles, 66(11), 764–775. [ Google Scholar ] Gray HM, Gray K, & Wegner DM (2007). Dimensions of mind perception. Science, 315(5812), 619–619. [ DOI ] [ PubMed ] [ Google Scholar ] Gray K, Knobe J, Sheskin M, Bloom P, & Barrett LF (2011). More than a body: mind perception and the nature of objectification. Journal of Personality and Social Psychology, 101(6), 1207–1220. 10.1037/a0025883 [ DOI ] [ PubMed ] [ Google Scholar ] Guizzo F, & Cadinu M (2017). Effects of objectifying gaze on female cognitive performance: The role of flow experience and internalization of beauty ideals. British Journal of Social Psychology, 56(2), 281–292. 10.1111/bjso.12170 [ DOI ] [ PubMed ] [ Google Scholar ] Halim ML, Ruble DN, & Amodio DM (2011). From pink frilly dresses to ‘one of the boys’: A social-cognitive analysis of gender identity development and gender bias. Social and Personality Psychology Compass, 5(11), 933–949. [ Google Scholar ] Hargreaves D, & Tiggemann M (2002). The role of appearance schematicity in the development of adolescent body dissatisfaction. Cognitive Therapy and Research, 26, 691–700. [ Google Scholar ] Harrison K, & Fredrickson BL (2003). Women’s sport media, self-objectification, and mental health in black and white adolescent females. Journal of Communication, 53(2), 216–232. 10.1111/j.1460-2466.2003.tb02587.x [ DOI ] [ Google Scholar ] Haslam N (2006). Dehumanization: An Integrative Review. Personality and Social Psychology Review, 10(3), 252–264. 10.1207/s15327957pspr1003_4 [ DOI ] [ PubMed ] [ Google Scholar ] Heflick NA, Goldenberg JL (2009). Objectifying Sarah Palin: Evidence that objectification causes women to be perceived as less competent and less fully human. Journal of Experimental Social Psychology, 45 ( 3 ), 598–601, 10.1016/j.jesp.2009.02.008 [ DOI ] [ Google Scholar ] Heflick NA, Goldenberg JL (2014). Seeing eye to body: The literal objectification of women. Current Directions in Psychological Science, 23 (3), 225–229, 10.1177/0963721414531599 [ DOI ] [ Google Scholar ] Heflick NA, Goldenberg JL, Cooper DP, Puvia E (2011). From women to objects: Appearance focus, target gender, and perceptions of warmth, morality and competence. Journal of Experimental Social Psychology, 47, 572–581 [ Google Scholar ] Holland E, & Haslam N (2016). Cute little things: The objectification of prepubescent girls. Psychology of Women Quarterly, 40(1), 108–119. 10.1177/0361684315602887 [ DOI ] [ Google Scholar ] Johnson KL, Freeman JB, & Pauker K (2012). Race is gendered: How covarying phenotypes and stereotypes bias sex categorization. Journal of Personality and Social Psychology, 102(1), 116–131. 10.1037/a0025335 [ DOI ] [ PubMed ] [ Google Scholar ] Kant Immanuel (1963). On History. Indianapolis: Bobbs-Merrill. [ Google Scholar ] Klein H, & Shiffman KS (2006). Messages about physical attractiveness in animated cartoons. Body Image, 3(4), 353–363. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Kumle L, Võ ML, & Draschkow D (2021). Estimating power in (generalized) linear mixed models: An open introduction and tutorial in R. Behavior research methods, 53(6), 2528–2543. 10.3758/s13428-021-01546-0 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Lei RF, Leshin RA, Moty K, Foster-Hanson E, & Rhodes M (2022). How race and gender shape the development of social prototypes in the United States. Journal of Experimental Psychology: General, 151 ( 8 ), 1956–1971. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Lei RF, Leshin RA, & Rhodes M (2020). The Development of Intersectional Social Prototypes. Psychological Science, 31(8): 911–926. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Leshin RA, Lei RF, Byrne M, & Rhodes M (2022). Who is a typical woman? How race biases representations of gender across development. Developmental Science, 25(2), e13175. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Leshin RA, Leslie SJ, & Rhodes M (2021). Does it matter how we speak about social kinds? A large, preregistered, online experimental study of how language shapes the development of essentialist beliefs. Child Development, 92(4), e531–e547. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Leshin R & Rhodes M (2025, January 10). Young children tend to be objectified along gender lines. Retrieved from osf.io/89fpn [ DOI ] [ PMC free article ] [ PubMed ] LoBue V, & Thrasher C (2015). The Child Affective Facial Expression (CAFE) set: Validity and reliability from untrained adults. Frontiers in Psychology, 5, 1532. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 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(5), 709–717. [ Google Scholar ] Loughnan S & Vaes J (2017). Objectification: Seeing and treating people as things. British Journal of Social Psychology, 56, 213–216. 10.1111/bjso.12205 [ DOI ] [ PubMed ] [ Google Scholar ] Malt BC & Johnson EC (1992). Do artifact concepts have cores? Journal of Memory and Language, 31(2), 195–217. [ Google Scholar ] McLoughlin N & Over H (2017). Young Children Are More Likely to Spontaneously Attribute Mental States to Members of Their Own Group. Psychological Science. 10.1177/0956797617710724 [ DOI ] [ PubMed ] [ Google Scholar ] McLoughlin N & Over H (2018). The developmental origins of dehumanization. Advances in Child Development and Behavior, 54, 153–178. [ DOI ] [ PubMed ] [ Google Scholar ] McLoughlin N, Tipper SP, & Over H (2018). Young children perceive less humanness in outgroup faces. Developmental Science, 21 : e12539. 10.1111/desc.12539 [ DOI ] [ PubMed ] [ Google Scholar ] Mervis CB & Rosch E (1981). Categorization of natural objects. Annual Review of Psychology, 32(1), 89–115. [ Google Scholar ] Miller CF, Lurye LE, Zosuls KM, & Ruble DN (2009). Accessibility of gender stereotype domains: Developmental and gender differences in children. Sex Roles, 60(11), 870–881. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Millisecond. (2016). Inquisit 5 [Computer software]. Retrieved from http://www.millisecond.com [Google Scholar] Moradi B, & Huang Y-P (2008). Objectification theory and psychology of women: A decade of advances and future directions. Psychology of Women Quarterly, 32(4), 377–398. 10.1111/j.1471-6402.2008.00452.x [ DOI ] [ Google Scholar ] Morris KL & Goldenberg J (2015). Objects become her: The role of mortality salience on men’s attraction to literally objectified women. Journal of Experimental Social Psychology, 56, 69–72. [ Google Scholar ] Morris KL, Goldenberg J, & Boyd P (2018). Women as animals, women as objects: Evidence for two forms of objectification. Personality and Social Psychology Bulletin, 44(9), 1302–1314. [ DOI ] [ PubMed ] [ Google Scholar ] Mosley AJ, Bharj N, & Biernat M (2023). Shifting standards of sexuality: An intersectional account of men’s objectification of Black and White women. Sex Roles, 89(9), 567–594. [ Google Scholar ] Nussbaum MC (1995). Objectification. Philosophy & Public Affairs, 24(4), 249–291. http://www.jstor.org/stable/2961930 [ Google Scholar ] Purdie-Vaughns V & Eibach RP (2008). Intersectional invisibility: The distinctive advantages and disadvantages of multiple subordinate-group identities. Sex Roles: A Journal of Research, 59(5–6), 377–391. 10.1007/s11199-008-9424-4 [ DOI ] [ Google Scholar ] Rhodes M, Rizzo MT, Foster-Hanson E, Moty K, Leshin RA, Wang M, Benitez J, & Ocampo JD (2020). Advancing developmental science via unmoderated remote research with children. Journal of Cognition and Development, 21(4), 477–493. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Rubin JZ, Provenzano FJ, & Luria Z (1974). The eye of the beholder: Parents’ views on sex of newborns. American Journal of Orthopsychiatry, 44(4), 512. [ DOI ] [ PubMed ] [ Google Scholar ] Rudman LA, & Mescher K (2012). Of Animals and Objects. Personality and Social Psychology Bulletin. 10.1177/0146167212436401 [ DOI ] [ PubMed ] [ Google Scholar ] Slater A & Tiggemann M (2016). Little girls in a grown up world: Exposure to sexualized media, internalization of sexualization messages, and body image in 6–9 year-old girls. Body Image, 18, 19–22. [ DOI ] [ PubMed ] [ Google Scholar ] Starr CR & Ferguson GM (2012). Sexy dolls, sexy grade-schoolers? Media & maternal influences on young girls’ self-sexualization. Sex Roles, 67, 463–476. [ Google Scholar ] Starr CR & Zurbriggen EL (2019). Self-sexualization in preadolescent girls: Associations with self-objectification, weight concerns, and parent’s academic expectations. International Journal of Behavioral Development, 43(6), 515–522. [ Google Scholar ] Thomas EL, Dovidio JF, & West TV (2014). Lost in the categorical shuffle: Evidence for the social non-prototypicality of Black women. Cultural Diversity and Ethnic Minority Psychology, 20(3), 370–376. 10.1037/a0035096 [ DOI ] [ PubMed ] [ Google Scholar ] Trifiletti E, Di Bernardo GA, Falvo R, & Capozza D (2014). Patients are not fully human: A nurse’s coping response to stress. Journal of Applied Social Psychology, 44(12), 768–777. [ Google Scholar ] Vaes J, Paladino P, & Puvia E (2011). Are sexualized women complete human beings? Why men and women dehumanize sexually objectified women. European Journal of Social Psychology, 41(6), 774–785. 10.1002/ejsp.824 [ DOI ] [ Google Scholar ] Weisman K, Dweck CS, & Markman EM (2018). Folk philosophy of mind: Changes in conceptual structure between 4–9y of age. Proceedings of the Annual Meeting of the Cognitive Science Society. [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Graphics and text NIHMS2063855-supplement-Graphics_and_text.docx (4.6MB, docx) ACTIONS View on publisher site PDF (1.9 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top