Maladaptive but Malleable: Gender-Science Stereotypes Emerge Early but are Modifiable by Language
Saved in:
| Title: | Maladaptive but Malleable: Gender-Science Stereotypes Emerge Early but are Modifiable by Language |
|---|---|
| Language: | English |
| Authors: | Michelle M. Wang (ORCID |
| Source: | Child Development. 2025 96(2):865-880. |
| Availability: | Wiley. Available from: John Wiley & Sons, Inc. 111 River Street, Hoboken, NJ 07030. Tel: 800-835-6770; e-mail: cs-journals@wiley.com; Web site: https://www.wiley.com/en-us |
| Peer Reviewed: | Y |
| Page Count: | 16 |
| Publication Date: | 2025 |
| Sponsoring Agency: | National Science Foundation (NSF), Directorate for Education and Human Resources (EHR) |
| Contract Number: | 2000617 |
| Document Type: | Journal Articles Reports - Research |
| Descriptors: | Sex Stereotypes, Scientists, Preschool Children, Sciences, Language Usage, Intervention |
| DOI: | 10.1111/cdev.14213 |
| ISSN: | 0009-3920 1467-8624 |
| Abstract: | Gender-science stereotypes emerge early in childhood, but little is known about the developmental processes by which they arise. The present study tested the hypothesis that language implying scientists are a special and distinct kind of person contributes to the development of gender-science stereotypes, even when it does not communicate stereotypic content. One cross-sectional and two longitudinal studies with racially and geographically diverse children (primarily from the United States; ages 4-5; N = 872, tested 2020-2022) revealed that gender stereotypes about science versus art (a) emerge before elementary school, (b) arise from commonplace identity-emphasizing language, especially among girls, and (c) can be durably disrupted by subtle changes to language. This study identifies a promising way to counteract stereotypes at their roots. |
| Abstractor: | As Provided |
| Notes: | https://osf.io/3hdxt/?view_only=b0bc115f94364373822429344494435b |
| Entry Date: | 2025 |
| Accession Number: | EJ1461468 |
| Database: | ERIC |
|
Full text is not displayed to guests.
Login for full access.
|
|
| FullText | Links: – Type: pdflink Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwGp-X1dQGn9sAeRN1o05i7AAAAA4jCB3wYJKoZIhvcNAQcGoIHRMIHOAgEAMIHIBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDPxorIx0sKqAZDMZ8wIBEICBmid81DDk60PE8SUkHiMY0KV3_9gtniSbVjD3dHu1UzvGkCFg4pUfLMsw9ovoya1GFxe70YOvmVg9_r9TfgMigxJhGJk_IifX3Jy7X8-S_669Qp5rJfD_wkte7dqlvpmTwZ0vZ2xgfsnfIhYMfEDZfSe2oV5dlgqdwOkPuxCtH10nLqT2jVzl3csgku-0ikVEvoVzAkEILAs25Ro= Text: Availability: 1 Value: <anid>AN0183920700;cdv01mar.25;2025Mar24.07:20;v2.2.500</anid> <title id="AN0183920700-1">Maladaptive but malleable: Gender‐science stereotypes emerge early but are modifiable by language </title> <p>Gender‐science stereotypes emerge early in childhood, but little is known about the developmental processes by which they arise. The present study tested the hypothesis that language implying scientists are a special and distinct kind of person contributes to the development of gender‐science stereotypes, even when it does not communicate stereotypic content. One cross‐sectional and two longitudinal studies with racially and geographically diverse children (primarily from the United States; ages 4–5; N = 872, tested 2020–2022) revealed that gender stereotypes about science versus art (a) emerge before elementary school, (b) arise from commonplace identity‐emphasizing language, especially among girls, and (c) can be durably disrupted by subtle changes to language. This study identifies a promising way to counteract stereotypes at their roots.</p> <p>Gender‐science stereotypes take root in early childhood and contribute to gender disparities in science (Cheryan et al., [<reflink idref="bib10" id="ref1">10</reflink>]; Master et al., [<reflink idref="bib44" id="ref2">44</reflink>]), imposing significant costs on individual women and the advancement of science and society (Bell et al., [<reflink idref="bib5" id="ref3">5</reflink>]; Ferrant &amp; Nowacka, [<reflink idref="bib18" id="ref4">18</reflink>]). The age‐related trajectory and consequences of gender stereotypes about science are well‐documented, but we know little about the processes underlying their development. We propose a social‐constructive process wherein subtle features of language, including noun labels (e.g., "Let's be <emph>scientists</emph>!") and generic descriptions (e.g., "<emph>Scientists</emph> discover things about the world!") lead children to think of <emph>scientists</emph> as a distinct kind of person, and then to search for cues to <emph>who</emph> that kind of person might be. On this account, subtle features of language contribute to the acquisition of social stereotypes and can do so even when language appears inclusive and does not include any stereotypic content. Such identity‐cuing science language is the most common way science is introduced to young children in the United States in prekindergarten classrooms and children's media (Wang et al., [<reflink idref="bib68" id="ref5">68</reflink>]). Thus, here we tested whether a prekindergarten‐level science curriculum replacing this common language with more action‐oriented descriptions of science (describing science as something that people <emph>do</emph> rather than as a category of person that one must <emph>be</emph>) can disrupt the development of gender‐science stereotypes in early childhood.</p> <p>Gender stereotypes about science and related academic domains begin to develop in early childhood and strengthen across the first several years of school (Master, [<reflink idref="bib42" id="ref6">42</reflink>]; Master et al., [<reflink idref="bib44" id="ref7">44</reflink>]; Miller et al., [<reflink idref="bib45" id="ref8">45</reflink>]; Starr &amp; Simpkins, [<reflink idref="bib64" id="ref9">64</reflink>]; Zhao et al., [<reflink idref="bib70" id="ref10">70</reflink>]). By at least age six, children begin to view scientists as male (Miller et al., [<reflink idref="bib45" id="ref11">45</reflink>]), believe boys are more interested than girls in engineering and computer science (Master et al., [<reflink idref="bib44" id="ref12">44</reflink>]), endorse the belief that math is for boys (Cvencek et al., [<reflink idref="bib14" id="ref13">14</reflink>]), and associate high‐level intellectual ability with men more often than women (Bian et al., [<reflink idref="bib6" id="ref14">6</reflink>]). From elementary to middle school, children become increasingly aware of gender‐science stereotypes; at these ages, children's explicit awareness correlates with their own endorsement of stereotypes (Cvencek et al., [<reflink idref="bib15" id="ref15">15</reflink>]; Kurtz‐Costes et al., [<reflink idref="bib29" id="ref16">29</reflink>]; Master, [<reflink idref="bib42" id="ref17">42</reflink>]). By adolescence, children's stereotypes match those held by adults, favoring boys and men across a range of scientific subfields and related domains and abilities (Kurtz‐Costes et al., [<reflink idref="bib29" id="ref18">29</reflink>]; Starr &amp; Simpkins, [<reflink idref="bib64" id="ref19">64</reflink>]).</p> <p>Gender stereotypes have immediate (e.g., Bian et al., [<reflink idref="bib6" id="ref20">6</reflink>]; Master et al., [<reflink idref="bib44" id="ref21">44</reflink>]) and long‐term problematic consequences on girls' and women's educational and career trajectories (Leibham et al., [<reflink idref="bib35" id="ref22">35</reflink>]; Newton &amp; Newton, [<reflink idref="bib49" id="ref23">49</reflink>]; Sax, [<reflink idref="bib62" id="ref24">62</reflink>]). Girls who endorse gender‐science stereotypes have more negative self‐perceptions in science (Plante et al., [<reflink idref="bib54" id="ref25">54</reflink>]), feel a lower sense of belonging in science (Master et al., [<reflink idref="bib44" id="ref26">44</reflink>]), indicate lower science interest (Master et al., [<reflink idref="bib44" id="ref27">44</reflink>]; Plante et al., [<reflink idref="bib55" id="ref28">55</reflink>]), disengage more from scientific activities (Steffens et al., [<reflink idref="bib66" id="ref29">66</reflink>]), and are less likely to pursue science as a major or career (Cundiff et al., [<reflink idref="bib13" id="ref30">13</reflink>]; Lane et al., [<reflink idref="bib30" id="ref31">30</reflink>]; Nosek &amp; Smyth, [<reflink idref="bib51" id="ref32">51</reflink>]). Therefore, gender‐science stereotypes contribute to gender disparities in science interest, engagement, and attainment (Cundiff et al., [<reflink idref="bib13" id="ref33">13</reflink>]; Lane et al., [<reflink idref="bib30" id="ref34">30</reflink>]; Master et al., [<reflink idref="bib44" id="ref35">44</reflink>]).</p> <p>Indeed, like gender stereotypes, gender gaps in science engagement emerge early in childhood and widen across development (Baram‐Tsabari et al., [<reflink idref="bib3" id="ref36">3</reflink>]; Leibham et al., [<reflink idref="bib35" id="ref37">35</reflink>]; Rhodes et al., [<reflink idref="bib59" id="ref38">59</reflink>]). Starting from as early as age four, girls engage in science‐related activities less often and display lower levels of science interest than boys (Leibham et al., [<reflink idref="bib35" id="ref39">35</reflink>]). Early disparities in science interest and engagement become more pronounced in high school and college, when adolescents choose future majors and careers (Desy et al., [<reflink idref="bib16" id="ref40">16</reflink>]; Miller et al., [<reflink idref="bib46" id="ref41">46</reflink>]; National Science Foundation, [<reflink idref="bib48" id="ref42">48</reflink>]; Sadler et al., [<reflink idref="bib61" id="ref43">61</reflink>]; Sax, [<reflink idref="bib62" id="ref44">62</reflink>]), even though girls' academic performance in science does not differ from that of boys at these ages (O'Dea et al., [<reflink idref="bib53" id="ref45">53</reflink>]). Disparities in science engagement have lifelong educational and economic implications—children who show early disengagement from science are less likely to do well in science, major or consider majoring in science, pursue graduate degrees in science, or make high levels of income (Ainley &amp; Ainley, [<reflink idref="bib2" id="ref46">2</reflink>]; Hall et al., [<reflink idref="bib24" id="ref47">24</reflink>]; Kang et al., [<reflink idref="bib27" id="ref48">27</reflink>]; Leibham et al., [<reflink idref="bib35" id="ref49">35</reflink>]).</p> <p>Due to the cumulative nature of education, the earlier gender‐science stereotypes emerge, the greater their impact may be (Agars, [<reflink idref="bib1" id="ref50">1</reflink>]; Bian et al., [<reflink idref="bib6" id="ref51">6</reflink>]; Leibham et al., [<reflink idref="bib35" id="ref52">35</reflink>]; Master et al., [<reflink idref="bib44" id="ref53">44</reflink>])—and correspondingly, the earlier they can be disrupted, the less they may interfere with children's educational trajectories. Therefore, it is crucial to identify <emph>when</emph> and <emph>how</emph> these stereotypes first develop to prevent them at their initial emergence. Although the emergence and trajectory of gender‐science stereotypes—as well as gender stereotypes in other related domains (e.g., math, computer science, engineering; Cvencek et al., [<reflink idref="bib14" id="ref54">14</reflink>]; Master et al., [<reflink idref="bib44" id="ref55">44</reflink>]; Steele, [<reflink idref="bib65" id="ref56">65</reflink>])—across childhood, has been well‐documented, little is known about the underlying processes that give rise to their development.</p> <p>Children could learn social stereotypes via direct transmission in language (e.g., statements that explicitly compare boys' and girls' abilities); however, many parents and teachers try to avoid perpetuating stereotypes and explicitly endorse egalitarian beliefs about science (Wang et al., [<reflink idref="bib68" id="ref57">68</reflink>]). Here we considered whether subtler features of language might contribute to stereotype acquisition in ways that parents and teachers do not explicitly anticipate or intend (Chestnut et al., [<reflink idref="bib11" id="ref58">11</reflink>]; Rhodes et al., [<reflink idref="bib59" id="ref59">59</reflink>]). Category labels ("This is a famous <emph>scientist</emph>") and generic statements ("<emph>Scientists</emph> work hard to solve problems") generally lead children to view the referenced group—in this case, <emph>scientists</emph>—as a stable and fundamentally distinct kind of person (Gelman et al., [<reflink idref="bib20" id="ref60">20</reflink>]; Rhodes et al., [<reflink idref="bib58" id="ref61">58</reflink>]; Waxman, [<reflink idref="bib69" id="ref62">69</reflink>]). Because children actively seek out information about categories (Gelman, [<reflink idref="bib19" id="ref63">19</reflink>]), especially gender (Halim et al., [<reflink idref="bib23" id="ref64">23</reflink>]; Martin &amp; Ruble, [<reflink idref="bib40" id="ref65">40</reflink>]), once they conclude that one must be a particular <emph>kind</emph> of person to succeed in science, they are likely to search for cues as to <emph>which</emph> kind of person this might be. Combined with books, media, and personal interactions that predominantly depict scientists as male (Charlesworth et al., [<reflink idref="bib9" id="ref66">9</reflink>]; Kerkhoven et al., [<reflink idref="bib28" id="ref67">28</reflink>]; Lewis &amp; Lupyan, [<reflink idref="bib37" id="ref68">37</reflink>]), identity‐focused science language—which is the most common way science is introduced to young children in the United States in prekindergarten classrooms and children's media (Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref69">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref70">68</reflink>])—could lay the foundation for children to conclude that the kind of person who succeeds in science is typically male.</p> <p>This experience may be especially impactful for girls given the mismatch between who is represented in science and their own gender. In this way, commonplace features of language could facilitate the acquisition of stereotyped beliefs, especially among girls—even without directly communicating stereotypes—because of how language interacts with the basic processes underlying children's construction of the world (Rhodes et al., [<reflink idref="bib59" id="ref71">59</reflink>]). In comparison, language that is more action‐oriented (e.g., "This famous person <emph>does science</emph>"; "<emph>Doing science</emph> means working hard to solve problems") does not set‐up an expectation that one has to be a specific kind of person to succeed, and thus should not facilitate stereotype acquisition in the same manner.</p> <p>Guided by this framework, these studies aimed to identify <emph>when</emph> (Study 1) and <emph>how</emph> (Studies 2–3) gender stereotypes about science (as compared with another activity, in this case, art) first emerged and how their initial emergence might be prevented. Complete study materials, data, and code for all studies are available at https://osf.io/3hdxt/?view%5fonly=b0bc115f94364373822429344494435b. All methods and procedures were approved by New York University's Institutional Review Board (IRB‐FY2020‐4169).</p> <hd id="AN0183920700-2">STUDY 1</hd> <p>The earliest that gender‐science stereotypes have been previously documented is around age 6 (Miller et al., [<reflink idref="bib45" id="ref72">45</reflink>]), but gender gaps in science engagement emerge even earlier (Leibham et al., [<reflink idref="bib35" id="ref73">35</reflink>]; Rhodes et al., [<reflink idref="bib59" id="ref74">59</reflink>]). Thus, we first tested whether children have begun to acquire gender‐science stereotypes earlier than previously documented (and if such stereotypes have immediate consequences for science engagement), to guide our efforts in Studies 2 and 3 to reveal <emph>how</emph> stereotypes develop and might be prevented.</p> <hd id="AN0183920700-3">Method</hd> <p></p> <hd id="AN0183920700-4">Participants</hd> <p>Participants included 199 children (102 girls, 97 boys; <emph>M</emph><subs>age</subs> = 4.93 years, SD = 0.62, range = 4.00–5.98 years; data collected from August 2020 to November 2022). Of those children whose parents provided racial and ethnic demographic information (<emph>n</emph> = 193), 61.7% of their parents identified them as White, 16.6% as Biracial or Multiracial, 14.5% as Asian or Asian American, and 4.7% as Black or African American; 8.8% of the sample, across race, were identified as Hispanic. A sensitivity analysis determined that the obtained sample was sufficient to detect a small‐to‐medium effect (<emph>f</emph> = .15) of stimuli gender at 95% power (<emph>α</emph> ≤ .05; Faul et al., [<reflink idref="bib17" id="ref75">17</reflink>]). Most children were enrolled in preschool or prekindergarten (69.3%) or kindergarten (19.1%), with 2.0% in the first grade and 9.5% not enrolled in school. Most children (<reflink idref="bib173" id="ref76">173</reflink>) were from the United States and came from 158 unique ZIP Codes across 31 states (see Figure S1). For Studies 1 to 3, children were all recruited via school partnerships, community outreach efforts, social media, and paid advertisements on Facebook and Instagram to our online recruitment and testing platform PANDA (discoveriesinaction.org; Rhodes, Rizzo, et al., 2020), where they were able to complete studies remotely on their own time and in their own homes.</p> <hd id="AN0183920700-5">Materials and procedure</hd> <p></p> <hd id="AN0183920700-6">Set‐up</hd> <p>Children completed the studies on their own computers with a webcam; data collection sessions were recorded to ensure that the child participant was a real child of the intended age range and that parents and children provided consent. The studies were completely self‐administered and consisted of animations and interactive narrations, so children were able to complete studies remotely, without ever directly interacting with a researcher or experimenter. Families were compensated with a $10 Amazon gift card at the end of the study session.</p> <hd id="AN0183920700-7">Measures</hd> <p></p> <hd id="AN0183920700-8">Gender stereotypes about science versus art</hd> <p>We developed a binary‐choice task appropriate for young children based on adult gender‐science stereotype measures that assess the strength of male‐science/female‐art associations and predict science achievement in adult populations (Charlesworth &amp; Banaji, [<reflink idref="bib8" id="ref77">8</reflink>]; Nosek et al., [<reflink idref="bib50" id="ref78">50</reflink>], [<reflink idref="bib52" id="ref79">52</reflink>]). We presented children with eight pictures of children (see Figure S4) one at a time and asked: "Here is a picture of a kid. Do you think this kid is really good at science (scored 1) or really good at art (scored 0)?". The pictured children consisted of four girls and four boys (one boy and one girl who was Asian, Black, Hispanic, and White, respectively; the main findings in this study and all others in this article were largely consistent across stimuli race and ethnicity, see Supporting Information for further analyses). The order of the eight pictured children, as well as whether the choice option of art or science was presented first, was randomized across participants, and the pictures were matched for perceived attractiveness, happiness, age, and quality of clothes. With this method, children made binary judgments about boys' and girls' science competence, without being required to express a gender stereotype about science at all (e.g., children could choose that all boys and girls are good at science). Also, unlike other methods commonly used to assess gender stereotypes about science (e.g., the "Draw‐A‐Scientist" task; Chambers, [<reflink idref="bib7" id="ref80">7</reflink>]), the task framing did not require children to think of science as associated with a profession or identity.</p> <hd id="AN0183920700-9">Science engagement</hd> <p>Children were presented with two pictures of books (one with a magnifying glass on its cover and one with an art palette on its cover; see Figure S5A) and asked: "Would you rather choose a book about science or a book about art?", and then with two pictures of boxes (one with a magnifying glass on it and one with an art palette on it; see Figure S5B) and asked: (<reflink idref="bib2" id="ref81">2</reflink>) "Would you rather do a science activity or an art activity?". Responses were scored as 1 for the "Book/activity about science" or 0 for the "Book/activity about art".</p> <hd id="AN0183920700-10">Science interest</hd> <p>Children were presented with a four‐point scale containing one dot and three stars of increasing size and asked: (<reflink idref="bib1" id="ref82">1</reflink>) "How much do you like science?" Responses were scored from 1 to 4 (1 = "Not at all", 2 = "A little", 3 = "Sort of", 4 = "A lot").</p> <hd id="AN0183920700-11">Science efficacy</hd> <p>Children were presented with a four‐point scale containing one dot and three stars of increasing size and asked: (<reflink idref="bib1" id="ref83">1</reflink>) "How good do you think you are at science?" Responses were scored from 1 to 4 (1 = "Not at all", 2 = "A little", 3 = "Sort of", 4 = "A lot").</p> <hd id="AN0183920700-12">Analytic strategy</hd> <p>We tested how children's likelihood of associating stimuli depicting boys and girls with being "really good at science" varied by target gender, participant gender, and participant age in a generalized linear mixed‐effects model with a binomial distribution, using the "glmer" function in the <emph>lme4</emph> package (Bates et al., [<reflink idref="bib4" id="ref84">4</reflink>]), with target gender, participant gender, participant age, and their interactions as predictors, and participant ID as a random intercept. We report likelihood ratio tests to evaluate each parameter. These analyses were intended to discover the extent to which children in the age range tested express beliefs consistent with gender stereotypes about science, rather than to test a particular hypothesis about their exact age of emergence.</p> <hd id="AN0183920700-13">Results</hd> <p>Overall, children predicted that boys would be good at science more often than girls (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref85">1</reflink>) = 6.65, <emph>p</emph> = .010, OR = 1.31, 95% CI [1.07, 1.61]). Children predicted that the boy characters would be good at science and art equally often (<emph>M</emph> = 0.47, SD = 0.50, 95% CI [0.43, 0.51]; <emph>b</emph> = −.12, SE = .07, <emph>z</emph> = −1.67, <emph>p</emph> = .094), but that girls would be good at science less often than they would be good at art (and less often than expected by chance; <emph>M</emph> = 0.40, SD = 0.49, 95% CI [0.37, 0.44]; <emph>b</emph> = −.44, SE = .09, <emph>z</emph> = −4.66, <emph>p</emph> &lt; .001). In addition, children became more likely to predict that boys were good at science (rather than art) across age (<emph>b</emph> = .28, SE = .12, <emph>z</emph> = 2.30, <emph>p</emph> = .022), whereas their responses for girls did not vary across age (<emph>b</emph> = −.18, SE = .12, <emph>z</emph> = −1.44, <emph>p</emph> = .150; two‐way interaction between target gender and participant age: <emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref86">1</reflink>) = 7.38, <emph>p</emph> = .007).</p> <p>Although this effect did not interact with participant gender (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref87">1</reflink>) = 1.33, <emph>p</emph> = .249), we explored the slopes between age and gender stereotypes for boys and girls separately, given that we expected the language manipulation in Studies 2 and 3 to be particularly influential for girls and wanted to test if age‐related patterns emerged differently across gender. Indeed, simple slope analyses revealed that girls' male‐science stereotypes increased across age (<emph>b</emph> = .47, SE = .17, <emph>z</emph> = 2.80, <emph>p</emph> = .005; see Figure 1), whereas boys' male‐science stereotypes and children's female‐science stereotypes did not (<emph>p</emph>s &gt; .20).</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/CDV/01mar25/cdev14213-fig-0001.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="cdev14213-fig-0001.jpg" title="1 Gender stereotype data by participant gender and age from Study 1. Boys' and girls' likelihood of associating boy (blue solid) and girl (orange dashed) targets with being &quot;really good at science&quot; in a cross‐sectional sample (ages 4–5) of children from primarily across the United States. Small circles represent average responses of individual participants. Error bands reflect 95% CIs." /> </p> <p></p> <p>For science engagement, overall, girls were less likely than boys to choose to engage in science (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref88">1</reflink>) = 12.61, <emph>p</emph> &lt; .001, OR = 0.30, 95% CI [0.16, 0.55]). To test if children's gender stereotypes related to variation in children's engagement, we created a composite gender stereotype score by subtracting the number of times that children said girls were good at science from the number of times they did so for boys to yield an indicator of the gender stereotype that boys are more likely to excel in science. We then implemented a generalized linear mixed‐effects model with a binomial distribution, using the "glmer" function in the <emph>lme4</emph> package (Bates et al., [<reflink idref="bib4" id="ref89">4</reflink>]), with gender stereotypes about science, participant gender, participant age, and their interactions as predictors, and participant ID and item (book or activity) as random intercepts.</p> <p>Indeed, girls who expressed more gender stereotypes about science were less likely to choose to engage in science (<emph>b</emph> = −1.28, SE = .52, <emph>z</emph> = −2.48, <emph>p</emph> = .013), whereas boys' stereotype expression did not relate to their engagement choices (<emph>b</emph> = .72, SE = .57, <emph>z</emph> = 1.27, <emph>p</emph> = .204; gender by stereotype interaction: <emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref90">1</reflink>) = 7.17, <emph>p</emph> = .007; see Figure 2). Overall, science engagement increased across age (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref91">1</reflink>) = 5.59, <emph>p</emph> = .018, OR = 1.65, 95% CI [1.64, 1.65]), but the effect of age did not interact with participant gender or the extent to which children expressed gender stereotypes.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/CDV/01mar25/cdev14213-fig-0002.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="cdev14213-fig-0002.jpg" title="2 Science engagement data from Study 1. Boys' (blue solid) and girls' (orange dashed) likelihood of choosing a science book or activity as predicted by their own gender stereotypes about science (higher scores indicate a stronger expression of the gender stereotype that boys are more likely to excel in science). Small circles represent average responses of individual participants. Error bands reflect 95% CIs." /> </p> <p></p> <p>To ensure that children's choices on the engagement task were not driven only by art preferences, we tested whether children's independent science interest and efficacy ratings predicted their engagement choices between science versus art in separate generalized linear mixed‐effects models with a binomial distribution, using the "glmer" function in the <emph>lme4</emph> package (Bates et al., [<reflink idref="bib4" id="ref92">4</reflink>]), with science interest/efficacy, participant gender, and their interactions as predictors, and participant ID and item (book or activity) as random intercepts.</p> <p>Indeed, both science interest (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref93">1</reflink>) = 10.08, <emph>p</emph> = .002, OR = 1.72, 95% CI [1.22, 2.43]) and science efficacy (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref94">1</reflink>) = 6.27, <emph>p</emph> = .012, OR = 1.50, 95% CI [1.10, 2.05]) independently predicted children's science engagement (i.e., choice of a science book/activity over an art book/activity), confirming that children's choices on this measure did not reflect their feelings about art alone (for further details on task rationale and development, see the supplemental materials).</p> <hd id="AN0183920700-16">Discussion</hd> <p>Study 1 found that the belief that boys are particularly good at science (more so than another activity, in this case, art) strengthens between ages 4 and 6 (particularly among girls). In addition, these early‐emerging stereotypes related to girls' own choices of whether to do a science activity.</p> <p>Our measure of children's gender stereotypes about science was based on adult measures that assess the strength of male‐science/female‐art associations (Charlesworth &amp; Banaji, [<reflink idref="bib8" id="ref95">8</reflink>]; Nosek et al., [<reflink idref="bib50" id="ref96">50</reflink>], [<reflink idref="bib52" id="ref97">52</reflink>]). In adults, such measures predict science achievement, interest, engagement, efficacy, and aspirations (Cundiff et al., [<reflink idref="bib13" id="ref98">13</reflink>]; Lane et al., [<reflink idref="bib30" id="ref99">30</reflink>]; Nosek et al., [<reflink idref="bib50" id="ref100">50</reflink>], [<reflink idref="bib52" id="ref101">52</reflink>]; Nosek &amp; Smyth, [<reflink idref="bib51" id="ref102">51</reflink>]; Zitelny et al., [<reflink idref="bib71" id="ref103">71</reflink>]). In addition to this connection to the adult literature, the stereotype measure that we used here had several methodological strengths that are relevant to our context. First, it did not include any identity‐based science language (in contrast to the "Draw‐A‐Scientist" task, which adopts an identity‐based framing within the task itself, Chambers, [<reflink idref="bib7" id="ref104">7</reflink>]). Second, it did not make gender especially salient by introducing a gender contrast or require expression of any gender stereotypes. That is, children were not asked to choose between a boy and a girl as good at science; instead, they were asked to select which activity each character was good at individually. Thus, children could have, in principle, responded that everyone is good at science. In addition, having children select activities for each person individually, rather than choose between people, avoided methodological concerns about in‐group bias. In particular, young children have a strong tendency to display in‐group favoritism related to gender in early childhood (Martinot &amp; Désert, [<reflink idref="bib41" id="ref105">41</reflink>]; Master &amp; Meltzoff, [<reflink idref="bib43" id="ref106">43</reflink>]; Miller et al., [<reflink idref="bib46" id="ref107">46</reflink>]), which could have led children to select their own gender on every item if the task had been set‐up to offer a choice between a boy and a girl. The present measure instead allowed us to examine children's domain‐specific stereotypes (whether they endorse the belief that boys and girls are good at <emph>different</emph> subjects; i.e., that science is not for everyone), without using a valence‐based approach (e.g., examining the belief that boys are good at something and girls are not).</p> <p>One question raised by the present measure is the extent to which the responses reflect beliefs about science versus art. Study 1 found, however, that change in gender stereotypes across this age range was particularly driven by increasing choices that boys are especially good at science (we found no change across age in children's beliefs about girls being good at art). In addition, the extent to which girls responded that boys are better at science more often than they did for girls correlated with their own choices of whether to engage in a science activity (which was itself independently predicted by their interest and efficacy in science). Taken together, these patterns suggest that the stereotype that boys are particularly good at science (that they are better at science than other activities, whereas girls are not) develops across early childhood, with immediate implications for engagement. Therefore, we next sought to test how such a tendency develops and might be prevented.</p> <hd id="AN0183920700-17">STUDY 2</hd> <p>To shed light on how gender stereotypes about science develop and might be prevented, in Study 2, we developed a virtual, unmoderated language‐based science intervention. Identity‐focused descriptions (i.e., language that includes noun labels and generic descriptions, such as "<emph>Scientists</emph> work hard to solve problems") are the most common way science is introduced to children in everyday contexts; they are three times as common as action‐oriented descriptions of science (i.e., language describing science in terms of actions, such as "<emph>Doing science</emph> means working hard to solve problems") in PBS Kids television shows and five times as common in prekindergarten classrooms (Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref108">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref109">68</reflink>]). Identity‐focused linguistic cues reduce science engagement and efficacy among children from social groups that are underrepresented in science, in both lab settings and the noisier, more variable contexts in which children are exposed to them in daily life, such as schools (Lei et al., [<reflink idref="bib32" id="ref110">32</reflink>]; Rhodes et al., [<reflink idref="bib59" id="ref111">59</reflink>]; Wang et al., [<reflink idref="bib68" id="ref112">68</reflink>]). No research to‐date has tested <emph>how</emph> language has these effects, however; here we consider that they do so by modifying children's beliefs and directly contributing to their acquisition of social stereotypes. Thus, we tested if commonplace language contributes to stereotype acquisition, and consequently, if replacing it with more action‐oriented language might prevent the development of gender‐science stereotypes in early childhood.</p> <p>The curriculum lasted ~ 1 month (which included four science lessons spaced at least 1 week apart; see Table S1). Because the lessons took place virtually at children's homes and encouraged practicing the concepts at home with their parents (see Figure 3, Panel 6), we considered that parents might generalize the language strategy they heard in the curriculum to how they talk about science more generally (which could then contribute to sustained effects of the intervention over time). To test this possibility, we recorded a brief sample of parents and children discussing science several weeks after the curriculum ended and tracked the development of children's gender stereotypes for a year post‐intervention.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/CDV/01mar25/cdev14213-fig-0003.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="cdev14213-fig-0003.jpg" title="3 Procedure of the science lessons in Studies 2 and 3. Sample screenshots and text from the science intervention, using the lesson on buoyancy as an example. Examples of condition‐varying language are colored (action in green and identity in purple). The actual lessons were presented as video animations and interactive narrations (Panels 1–5; see OSF repository for all videos) and downloadable activity guides (Panel 6; see OSF repository for all guides)." /> </p> <p></p> <hd id="AN0183920700-19">Method</hd> <p></p> <hd id="AN0183920700-20">Participants</hd> <p>Participants included 467 children (242 girls, 225 boys; <emph>M</emph><subs>age</subs> = 4.97 years, SD = 0.56, range = 4.00–5.99 years; an additional 18 children were recruited and began participation but were excluded from analyses because they had participated in Study 1; data collected from December 2020 to December 2021); 460 children's parents provided racial and ethnic demographic information, with 65.2% of those identified as White, 15.2% as Asian or Asian American, 10.9% as Biracial or Multiracial, 4.3% as Black or African American, and 0.4% as Native American; 12.4% of the sample, across race, were identified as Hispanic. Although 467 children began the study, the numbers who participated varied across time; for more information and analyses of attrition, see the supplemental materials. A sensitivity analysis determined that the obtained sample (<emph>n</emph> = 467, with at least one data point per participant) was sufficient to detect an interaction between target gender, participant gender, and condition at 99% power (<emph>α</emph> ≤ .05; <emph>f</emph> = .15). Of the children whose parents provided schooling information for them (<emph>n</emph> = 390) at the beginning of the study, most were enrolled in prekindergarten (60.5%) or kindergarten (35.4%), and 16.4% in other. Most children (<reflink idref="bib365" id="ref113">365</reflink>) were from the United States and came from 321 unique ZIP Codes across 34 states (see Figure S2). Of the parents who provided information on their level of education (<emph>n</emph> = 375), 1.6% completed high school, 4.5% an associate degree, 4.8% some college, 35.7% a bachelor's degree, 34.1% a master's degree, and 19.2% a doctoral or professional degree; 365 parents also provided educational information on the child's other parent, with 0.3% having completed elementary school, 0.5% some high school, 5.5% high school, 5.5% an associate degree, 7.4% some college, 35.3% a bachelor's degree, 27.1% a master's degree, and 18.4% a doctoral or professional degree.</p> <hd id="AN0183920700-21">Materials and procedure</hd> <p></p> <hd id="AN0183920700-22">Set‐up</hd> <p>Procedures for unmoderated remote research were similar to Study 1, with the exception that a small number (<emph>n</emph> = 39) of children participated on a tablet rather than a computer.</p> <hd id="AN0183920700-23">Procedure</hd> <p>Children participated in a month‐long science program with a year‐long follow‐up named, "Pre‐K Science Adventures with Curious Cat," that we fully scripted, animated, and narrated. Children completed up to four virtual, self‐paced science lessons (average lessons completed: 2.95) spaced at least 1 week apart (average length per lesson: 19.3 min), each teaching a new science concept (friction, gravity, buoyancy, transparency; see supplemental materials for full study descriptions and timelines). The lessons were taught by an animated, gender‐neutral teacher, "Curious Cat," who provided interactive narrations, asked questions, and provided automated feedback (see Figure 3). Children were required to click and respond to proceed through the lessons (i.e., the lessons were not passive videos, but rather interactive interfaces; see OSF repository for sample recordings of the full lessons). Children were randomly assigned to one of two versions of the science program: an <emph>identity</emph> condition (<emph>n</emph> = 235) where science was taught with the identity‐focused language that is common in children's media and classrooms (e.g., "Scientists use their hands and eyes to observe"; "Let's be scientists and make a prediction or a thoughtful guess"; Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref114">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref115">68</reflink>]), and an <emph>action</emph> condition (<emph>n</emph> = 232) where science was taught with action‐oriented language that is less common in children's daily contexts, but which we hypothesized would be more beneficial (e.g., "Doing science means using your hands and eyes to observe"; "Let's do science and make a prediction or a thoughtful guess"). Children's assigned language condition remained consistent across all four science lessons (see Tables S1 and S2).</p> <hd id="AN0183920700-24">Science lessons</hd> <p>For each lesson, Curious Cat (i.e., the animated, gender‐neutral teacher) introduced a scientific concept, showed interactive examples, asked related questions, and provided automated feedback. Children were able to navigate the lessons independently. For example, in the lesson on buoyancy (other lessons followed a similar procedure), Curious Cat explained the concept, including why different objects sink or float, and children saw a variety of different examples (e.g., a beach ball and a rock). Children responded to questions about the properties of the object (e.g., how heavy or light it was), and Curious Cat guided them to understand the target concept (e.g., explained that objects sink when they do not have buoyancy and that most heavy objects do not have buoyancy). For each example, children were guided to observe the relevant properties (e.g., whether an object looked heavy or light), and then to make a prediction (defined as a "thoughtful guess") about whether the object would sink or float. As a final step in the scientific process, they checked their guesses to see if they were correct (e.g., saw a video of the object sinking or floating). As feedback, Curious Cat stated what happened (e.g., "[That's right!] The rock sinks"), and then provided a brief explanation (e.g., "This is because the rock is heavy, so it does not have a lot of buoyancy, and the rock sinks"). All four science lessons utilized the three steps of the scientific process, teaching children the importance of (i) observing, (ii) making predictions or "thoughtful guesses", and (iii) checking their predictions. Children were asked how much they liked each lesson (1 = "Not at all", 2 = "A little", 3 = "Sort of", 4 = "A lot") after they completed all four lessons; of those who provided at least one response (<emph>n</emph> = 162), children indicated an overall tendency to like the curriculum (<emph>M</emph> = 3.33, SD = 0.69).</p> <hd id="AN0183920700-25">Gender stereotypes about science versus art</hd> <p>Children's gender stereotypes were measured four times: 1 week after the first science lesson, 1 week after all four science lessons, approximately 6 months after the first science lesson, and ~ 1 year after the first science lesson. Whereas the type of linguistic input children received during the four science lessons was condition‐dependent, the language in the gender stereotype measure was condition‐neutral (this measure was identical to that used in Study 1).</p> <hd id="AN0183920700-26">Storybook task</hd> <p>One week after all the science lessons were completed, families were invited to complete a brief task to assess parent–child conversation about science. The session in which this task was given did not include any condition‐dependent language (see Table S1). Children and their parents were presented with four images, one after another, of people engaging in scientific activities (see Figure 5a; note that the images that families discussed were not from any of the target lessons—and looked quite different—so parents were likely not mimicking Curious Cat directly). Parents were asked to discuss these images with their children however they would normally talk about pictures in a storybook, and families were able to progress through the images at their own pace (<emph>M</emph><subs>length</subs> = 3.35 min; see Table S3 for example conversations). Parents' and children's conversations were video recorded, transcribed, and coded for the use of action‐focused science language and identity‐focused science language (consistent with Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref116">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref117">68</reflink>]). Action‐focused science language cues included descriptions of the target both as "doing science" (e.g., "She's <emph>doing science</emph>!"; "When was the last time you <emph>did science</emph>?") and as doing a science activity in a specific field of science (e.g., "She is <emph>doing astronomy</emph>"; "Do you remember <emph>doing chemistry</emph>?"). Similarly, identity‐focused science language cues included descriptions of the target as "a scientist" or "being a scientist" (e.g., "He's a <emph>scientist</emph>!"; "Can you name any real‐life <emph>scientists</emph>?"), as well as other science identities in a specific field of science (e.g., "She's an <emph>astronomer</emph>"; "That's an <emph>archaeologist</emph>"). Inter‐rater reliability was high (Cohen's <emph>κ</emph> = .89), and all discrepancies were resolved by a third senior researcher.</p> <hd id="AN0183920700-27">Analytic strategy</hd> <p>We tested how children's gender stereotypes about science varied by target gender, participant gender, condition, and across time in a generalized linear mixed‐effects model with a binomial distribution, using the "glmer" function in the <emph>lme4</emph> package (Bates et al., [<reflink idref="bib4" id="ref118">4</reflink>]), with target gender, participant gender, condition, session, and their interactions as predictors, and participant ID as a random intercept. These analyses were designed as confirmatory tests of our unpreregistered hypotheses.</p> <p>We tested how parents' use of action‐ and identity language cues in the Storybook Task varied by participant gender and condition in separate generalized linear models with a binomial distribution, using the "glm" function in the <emph>stats</emph> package (R Core Team, [<reflink idref="bib56" id="ref119">56</reflink>]), with participant gender, condition, and their interactions as predictors. These analyses were also designed as confirmatory tests of our unpreregistered hypotheses.</p> <hd id="AN0183920700-28">Results</hd> <p>Overall, children predicted that boys would be good at science (rather than art) more often than girls (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref120">1</reflink>) = 213.36, <emph>p</emph> &lt; .001, OR = 1.94, 95% CI [1.78, 2.12]). Children predicted that boy characters were good at science more often than expected by chance (<emph>M</emph> = 0.56, SD = 0.50, 95% CI [0.55, 0.58]; <emph>b</emph> = .24, SE = .04, <emph>z</emph> = 5.67, <emph>p</emph> &lt; .001), but that girls were good at science less often than expected by chance (<emph>M</emph> = 0.40, SD = 0.49, 95% CI [0.39, 0.42]; <emph>b</emph> = −.43, SE = .05, <emph>z</emph> = −9.23, <emph>p</emph> &lt; .001).</p> <p>Consistent with our hypothesis, the language children heard modulated their beliefs, particularly among girls. Girls in the identity condition predicted that boys were good at science more often than girls in the action condition (<emph>b</emph> = .26, SE = .10, <emph>z</emph> = 2.61, <emph>p</emph> = .009; all other condition‐based contrasts: <emph>p</emph>s &gt; .15). In addition, in the identity condition, girls were less likely than boys to predict that girls are good at science (<emph>b</emph> = −.41, SE = .10, <emph>z</emph> = −3.91, <emph>p</emph> &lt; .001), whereas in the action condition, boys and girls responded similarly on the stereotype measure (all other gender‐based contrasts: <emph>p</emph>s &gt; .09; three‐way interaction of target gender, participant gender, and language condition: <emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref121">1</reflink>) = 7.11, <emph>p</emph> = .008; see Figure 4a). These effects did not interact with time, meaning that there was no evidence that the effect of condition on children's gender stereotypes diminished over time, even up to 1 year after children completed the curriculum.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/CDV/01mar25/cdev14213-fig-0004.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="cdev14213-fig-0004.jpg" title="4 Longitudinal gender stereotype data by participant gender and type of linguistic input from Studies 2 and 3. Boys' and girls' likelihood of associating boy (blue) and girl (orange) targets with being &quot;really good at science&quot; by assigned language condition (identity: high alpha, solid lines and squares; action: low alpha, dashed lines and triangles) in a longitudinal sample (ages 4–5) of primarily White children (a) and children of color (b) from primarily across the United States. Sample sizes for each session are included (the number of sessions participants chose to complete did not vary by participant gender, condition, or their interaction; see the supplemental materials for more information on attrition). Large shapes represent group means and small shapes represent individual scores. Error bars reflect ±1 SE." /> </p> <p></p> <hd id="AN0183920700-30">Parent language data</hd> <p>Overall, parents used identity‐focused language (<emph>M</emph> = 3.96 statements per conversation, SD = 4.34) 11 times as often as action‐focused language (<emph>M</emph> = 0.36, <emph>SD</emph> = 0.82) in conversations that took place weeks after the end of the curriculum; thus, as in children's media and prekindergarten classrooms, identity‐focused language is also highly prevalent in parent–child conversation. Parents of children in the action condition, however, were significantly more likely to use action‐focused science language (which is rare in children's daily contexts) than parents of children in the identity condition (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref122">1</reflink>) = 6.36, <emph>p</emph> = .012, OR = 2.41, 95% CI [1.19, 5.03]; see Figure 5b; see supplemental materials for additional analyses), confirming that the curriculum indeed modified how parents talked about science with their children on an ongoing basis. This finding bolsters the possibility that the curriculum had sustained effects on children's beliefs because it modified parent–child conversations about science in daily life.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/CDV/01mar25/cdev14213-fig-0005.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="cdev14213-fig-0005.jpg" title="5 Materials and data for natural language task in Study 2. (a) Images used to examine parents' and children's natural language when discussing science (pictures are of astronomy, chemistry, archaeology, and botany). Pictures were presented one at a time, with the order randomized across participants. Families were encouraged to discuss the pictures however they would typically discuss a picture book. (b) Proportion of parents who used action‐ (left) and identity‐focused (right) science language at least once by child's assigned language condition. Horizontal lines represent group means and small circles represent individual responses. Error bars reflect ±1 SE." /> </p> <p></p> <hd id="AN0183920700-32">Discussion</hd> <p>In Study 2, girls in the identity condition responded that boys are particularly good at science—consistent with male‐science gender stereotypes—more often than girls in the action condition. Thus, the language that children heard modulated their stereotype development, even though neither language condition mentioned gender in any manner, and the content of the lessons was identical across conditions and was consistently inclusive and encouraging. We suggest that identity‐focused language leads children to expect that one has to be a certain kind of person to succeed in science, making girls more susceptible to cultural cues that this type of person is male. From this perspective, action‐focused language does not set‐up such an expectation and thus acts as a buffer to such cultural cues.</p> <p>Identity‐focused language is the most common way that parents, prekindergarten teachers, and children's media introduce science to children in the United States (as shown for parents, here, and for teachers and media in Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref123">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref124">68</reflink>]). Thus, the trajectory of stereotyped responses that we found among girls in the identity condition likely reflects the trajectory that is common in the United States (as the language children heard in that condition is consistent with what they hear in daily life). Indeed, the pattern of responses in this condition is similar to that across ages in Study 1, in which children did not receive a language manipulation. In the context of prior work and Study 1, the pattern across conditions in the present study is consistent with the possibility that action‐focused language (which is rarer in children's daily lives) disrupts this typical trajectory.</p> <p>As in Study 1, the effect in the present study was specific to how often children responded that boys are good at science. Although they made this choice relative to art, the curriculum did not mention art at all, so it is difficult to identify a process wherein the identity condition would uniquely affect girls' beliefs about boys' abilities in art. The key effect in the present study is that girls in the identity‐focused condition were more likely to predict that boys are good at science—these predictions were significantly stronger than those among all other children (i.e., both girls in the action condition and boys in both conditions).</p> <p>The data on parent language provides insight into how the effects of the intervention might have been sustained over time (for up to at least 1 year after the curriculum). Although the curriculum focused on children, because it took place in children's homes with parents present, and encouraged practicing the activities with parents, we considered that parents might begin to use the language from the curriculum in further conversations about science with their children. Indeed, in the brief sample of parent–child conversation that we collected after the curriculum was complete, action‐language was more common among parents in the action condition (although identity was still the most common way of talking about science across both conditions). We did not collect enough language data from each family (only several minutes) to test if variation in parent language contributes to the sustained effects of the intervention over time, but the present findings (as well as past research that reveals a dose–response relation between language exposure and relevant outcome variables; Gelman et al., [<reflink idref="bib21" id="ref125">21</reflink>]; Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref126">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref127">68</reflink>]) are consistent with such a possibility. Future research could test such a relation directly; in the meantime, these findings illustrate the potential of unmoderated remote research to create change over time by modifying ongoing processes in parent–child interaction.</p> <hd id="AN0183920700-33">STUDY 3</hd> <p>Because the participants in Study 2 were primarily White (65%), we next sought to replicate the intervention with a second cohort of children from more diverse racial and ethnic backgrounds. This is important because children from more diverse racial and ethnic backgrounds often face multiple, intersecting stereotypes that could interfere with their achievements (Grossman &amp; Porche, [<reflink idref="bib22" id="ref128">22</reflink>]; Lei et al., [<reflink idref="bib34" id="ref129">34</reflink>]; Muradoglu et al., [<reflink idref="bib47" id="ref130">47</reflink>]); thus, we aimed to identify whether this language‐based intervention is powerful enough to affect beliefs across a more diverse population of children.</p> <hd id="AN0183920700-34">Method</hd> <p></p> <hd id="AN0183920700-35">Participants</hd> <p>Participants included 206 children of color (104 girls, 102 boys; <emph>M</emph><subs>age</subs> = 4.87 years, SD = 0.63, range = 4.00–6.10 years; two children who just turned six were accidentally recruited and included in the final sample; the main results did not change with or without these two children; data collected from August 2021 to July 2022). Again, parents provided racial and ethnic background information, with 33.8% of those identified as Biracial or Multiracial, 26.0% as Asian or Asian American, 18.6% as Black or African American, 11.8% as White, and 0.50% as Native American; 28.4% of the sample, of any race (including all White participants), were identified as Hispanic. A sensitivity analysis determined that the obtained sample (<emph>n</emph> = 206, with at least one data point per participant) was sufficient to detect an interaction between target gender, participant gender, and condition at 96% power (<emph>α</emph> ≤ .05; <emph>f</emph> = .15). Of the children whose parents provided schooling information for them (<emph>n</emph> = 158) at the beginning of the study, the majority (70.9%) were enrolled in prekindergarten, 13.3% in kindergarten, and 15.8% in other. Most children (<reflink idref="bib163" id="ref131">163</reflink>) were from the United States and came from 153 unique ZIP Codes across 27 states (see Figure S3). Of the parents who provided information on their level of education (<emph>n</emph> = 154), 5.8% completed high school, 3.9% an associate degree, 8.4% some college, 39% a bachelor's degree, 28.6% a master's degree, and 14.3% a doctoral or professional degree; 140 parents also provided educational information on the child's other parent, with 1.4% having completed elementary school, 0.7% some high school, 12.1% high school, 6.4% an associate degree, 10% some college, 29.3% a bachelor's degree, 23.6% a master's degree, and 16.4% a doctoral or professional degree.</p> <hd id="AN0183920700-36">Materials and procedure</hd> <p>Procedures were identical to Study 2, except that (<reflink idref="bib1" id="ref132">1</reflink>) children's gender stereotypes were measured three times instead of four: 1 week after the first science lesson, 1 week after all four science lessons, and ~ 10 months after the first science lesson (see supplemental materials for full study descriptions and timelines), and (<reflink idref="bib2" id="ref133">2</reflink>) we do not report data from the natural language task because we did not obtain sufficient data for analysis (<emph>n</emph> = 82; see Table S2).</p> <hd id="AN0183920700-37">Results</hd> <p>Consistent with Studies 1 and 2, children predicted that boys would be good at science (rather than art) more often than girls (<emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref134">1</reflink>) = 39.00, <emph>p</emph> &lt; .001, OR = 1.68, 95% CI [1.43, 1.97]). Again, children were above chance at endorsing the stereotype that boys are good at science (<emph>M</emph> = 0.55, SD = 0.50, 95% CI [0.52, 0.57]; <emph>b</emph> = .18, SE = .08, <emph>z</emph> = 2.40, <emph>p</emph> = .016), while below chance at predicting girls would be good at science (<emph>M</emph> = 0.42, SD = 0.49, 95% CI [0.39, 0.45]; <emph>b</emph> = −.32, SE = .08, <emph>z</emph> = −4.16, <emph>p</emph> &lt; .001).</p> <p>Critically, the longitudinal data again revealed stereotype development was modulated by the language that children heard. In the identity condition, girls expressed more male‐science stereotypes over time (<emph>b</emph> = .42, SE = .14, <emph>z</emph> = 2.94, <emph>p</emph> = .003), whereas the stereotypes of girls in the action condition and boys in both conditions remained constant (all other simple slopes across time: <emph>p</emph>s &gt; .06; four‐way interaction of target gender, participant gender, language condition, and time: <emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref135">1</reflink>) = 8.41, <emph>p</emph> = .004; see Figure 4b). Consistent with Study 2, girls in the identity condition were more likely to express male‐science stereotypes than girls in the action condition (<emph>b</emph> = .43, SE = .18, <emph>z</emph> = 2.44, <emph>p</emph> = .015; all other condition‐based contrasts: <emph>p</emph>s &gt; .40; three‐way interaction of target gender, participant gender, and language condition: <emph>χ</emph><sups>2</sups>(<reflink idref="bib1" id="ref136">1</reflink>) = 4.56, <emph>p</emph> = .033). Further illustrating the detrimental consequences of identity‐based science language for girls, in the identity condition, girls predicted that boys were good at science more often than boys did (<emph>b</emph> = .57, SE = .18, <emph>z</emph> = 3.09, <emph>p</emph> = .002) and that girls are good at science less often than boys did (<emph>b</emph> = −.41, SE = .18, <emph>z</emph> = −2.26, <emph>p</emph> = .024), whereas boys and girls in the action condition responded similarly on the stereotype measures (gender‐based contrasts in the action condition: <emph>p</emph>s &gt; .10).</p> <hd id="AN0183920700-38">Discussion</hd> <p>Study 3 replicated the effect of language on the development of girls' beliefs that boys are particularly good at science in a more racially and ethnically diverse sample of children. In the present study, in the identity conditions, girls' beliefs that boys are particularly good at science (compared with art) increased over time. There were no changes across time among boys or among girls in the action condition. These patterns are highly consistent with those from Study 1 (where only girls' male‐science associations strengthened across age) and Study 2 (where only the male‐science associations of girls in the identity condition were stronger than those of other children). Together, these data support the interpretation that commonplace, identity‐emphasizing linguistic cues in children's environments contribute to the development of girls' gender stereotypes about science, and that replacing such language with more action‐oriented language can help buffer girls against these processes in early childhood.</p> <hd id="AN0183920700-39">GENERAL DISCUSSION</hd> <p>Across one cross‐sectional and two longitudinal studies with more than 800 racially, ethnically, and geographically diverse children, we found that (i) gender stereotypes about science emerge before elementary school as children are first learning about science (the earliest they have been detected), and contribute to disparities in science engagement even before children begin formal schooling, (ii) common language used to introduce science to children contributes to the development of these stereotypes, especially among girls, and (iii) subtle changes to children's linguistic contexts durably reduce the acquisition of stereotypes in early childhood. Further, we found that gender stereotypes about science were expressed similarly, and were similarly responsive to language, by children from diverse racial and ethnic backgrounds.</p> <p>The present findings are consistent with a social‐constructive model of stereotype acquisition wherein identity‐focused language that entails frequent noun labels and generic descriptions (e.g., "Scientists use their hands and eyes to observe") leads children to think that one has to be a particular <emph>kind</emph> of person to succeed in science. Given that children actively construct their understanding of the social world (and are especially motivated to do so for gender; Gelman, [<reflink idref="bib19" id="ref137">19</reflink>]; Halim et al., [<reflink idref="bib23" id="ref138">23</reflink>]; Liben &amp; Bigler, [<reflink idref="bib38" id="ref139">38</reflink>]; Martin &amp; Ruble, [<reflink idref="bib40" id="ref140">40</reflink>]), this identity‐based representation of science could place them "on the lookout" for cues as to which kind of person this might be. Combined with frequent experience depicting members of that category as overwhelmingly or exclusively male (Kerkhoven et al., [<reflink idref="bib28" id="ref141">28</reflink>]; Long et al., [<reflink idref="bib39" id="ref142">39</reflink>]), such language facilitates the acquisition of social stereotypes—even though it does not communicate any explicit stereotypic content (and in fact, often appears stereotype‐neutral and inclusive)—<emph>because</emph> it interacts with the basic processes underlying how children actively construct their understanding of the social world (Liben &amp; Bigler, [<reflink idref="bib38" id="ref143">38</reflink>]; Rhodes et al., [<reflink idref="bib59" id="ref144">59</reflink>]).</p> <p>Language that describes science as an identity that people <emph>are</emph>, rather than an activity that people <emph>do</emph>, is highly prevalent in children's daily contexts and is the most common way that prekindergarten teachers and children's media introduce science to young children in the United States (Rhodes, Cardarelli, et al., [<reflink idref="bib57" id="ref145">57</reflink>]; Wang et al., [<reflink idref="bib68" id="ref146">68</reflink>]). The present research (Study 2) further found that such language is also prevalent from parents. The prevalence of this language in daily life bolsters the possibility that such language could be a plausible mechanism in children's daily environments that gives rise to the emergence of gender stereotypes. Indeed, the present findings highlight the pernicious consequences of <emph>identity language</emph>, especially for young girls, as (a) girls who did not participate in the intervention (and likely were exposed to science language that was overwhelmingly or exclusively identity‐based) developed stronger male‐science associations with age (Study 1), and (b) girls who were assigned to the identity condition (and likely received limited or no action‐oriented language input) expressed stronger male‐science associations than those who were not (Studies 2 and 3), with increasing endorsement across time (Study 3). Together, these data point to the possibility that commonplace, identity‐based language in children's daily contexts elicits the stereotype that boys are particularly likely to succeed in science among girls, but that action‐oriented linguistic input can effectively and durably disrupt this process.</p> <p>The present study complements past findings that identity‐focused linguistic cues reduce science engagement and efficacy among girls and children from other groups that are underrepresented in science (Lei et al., 2019; Rhodes et al., [<reflink idref="bib59" id="ref147">59</reflink>]) by revealing a potential mechanism through which it does so—modifying children's beliefs about who is good at science. Given that children's science engagement and efficacy are predictive of their learning outcomes (Jansen et al., [<reflink idref="bib25" id="ref148">25</reflink>]; Leibham et al., [<reflink idref="bib35" id="ref149">35</reflink>]), future work could test whether this type of commonplace language also takes a toll on girls' learning outcomes and achievements in science by eliciting the early emergence of gender stereotypes. And although the present study tested this model of stereotype acquisition specifically within the domain of science, future studies on other related domains (e.g., math, computer science, engineering) could reveal whether such a developmental process unfolds similarly for other kinds of social stereotypes.</p> <p>In the present studies, identity language facilitated stereotype acquisition only among girls. This might be the case because children in this age range tend to view their own gender as the "default" person (Laosa et al., [<reflink idref="bib31" id="ref150">31</reflink>]; Lei et al., [<reflink idref="bib33" id="ref151">33</reflink>]). Thus, once children infer (from language) that one must be a particular kind of person to succeed in science and are "on the lookout" for cues to what kind of person this might be, girls might be more likely to notice the mismatch between who they see represented in science (mostly men) and their default concept of people, and thus encode that information more deeply. In addition, as children acquire vast amounts of information from books, media, and personal interactions that exclusively or overwhelmingly depicts scientists as male (Charlesworth et al., [<reflink idref="bib9" id="ref152">9</reflink>]; Lewis &amp; Lupyan, [<reflink idref="bib37" id="ref153">37</reflink>]), girls may feel especially demotivated and begin to question whether they themselves belong to this natural and distinct social category. This negative socioemotional experience among girls who were exposed to identity language can be especially powerful in reinforcing the notion that girls cannot succeed in science (Vaish et al., [<reflink idref="bib67" id="ref154">67</reflink>]) and contribute to increasingly high levels of stereotype expression across time.</p> <p>Although it is possible that some portion of children's responses on our measure is driven by their gender‐art stereotypes in addition to their gender‐science stereotypes, the science lessons in Studies 2 and 3 exclusively focused on science (i.e., did not teach or discuss art or gender at all). Indeed, only girls' male‐science associations (not female‐art associations) increased across age (Study 1) and were strengthened by identity language input (Studies 2 and 3). The present measure had several important strengths; it did not require children to directly compare boys and girls (which might have artificially increased attention to gender), avoided patterns of response fully explainable by in‐group bias (Martinot &amp; Désert, [<reflink idref="bib41" id="ref155">41</reflink>]; Master &amp; Meltzoff, [<reflink idref="bib43" id="ref156">43</reflink>]; Miller et al., [<reflink idref="bib46" id="ref157">46</reflink>]), and was modeled on adult measures that have been found to predict variation in achievement (Charlesworth &amp; Banaji, [<reflink idref="bib8" id="ref158">8</reflink>]; Nosek et al., [<reflink idref="bib50" id="ref159">50</reflink>], [<reflink idref="bib52" id="ref160">52</reflink>]). Nevertheless, it would be helpful to confirm these patterns with additional measures of children's beliefs in future work.</p> <p>The present research found that gender‐science stereotypes emerge early across children from diverse backgrounds and are also endorsed consistently for stimuli from diverse racial and ethnic backgrounds (see supplemental materials). These patterns suggest that gender stereotypes about science may develop earlier and more robustly than other similar gender stereotypes, such as those about brilliance (Bian et al., [<reflink idref="bib6" id="ref161">6</reflink>]; Leslie et al., [<reflink idref="bib36" id="ref162">36</reflink>]), which appear to develop a little later in childhood and to be endorsed only for White targets (Jaxon et al., [<reflink idref="bib26" id="ref163">26</reflink>]; Shu et al., [<reflink idref="bib63" id="ref164">63</reflink>]). Gender stereotypes about science might be endorsed earlier and more consistently than other academic gender stereotypes due to the ubiquity of both identity‐focused language and other cues to gender‐science stereotypes across diverse contexts (Charlesworth et al., [<reflink idref="bib9" id="ref165">9</reflink>]; Lewis &amp; Lupyan, [<reflink idref="bib37" id="ref166">37</reflink>]). Differences in measurement across studies makes these comparisons somewhat difficult to interpret, however, suggesting the need for future work on the developmental trajectory of different academic stereotypes, and how they might relate to one another, across early childhood. For example, it would be interesting for future research to examine if identity‐based language about art also elicits gender‐art stereotypes in a similar way. Past research showing that children indicate less interest and efficacy after experiencing a setback in <emph>drawing</emph> when they are asked to "be a drawer" than "to draw" (Cimpian et al., [<reflink idref="bib12" id="ref167">12</reflink>]) supports the general perspective that identity language can be problematic, especially when children have reason to question if they are part of the successful group; however, children did not respond in gendered ways to identity‐based language about art, suggesting that perhaps for art, such "reason to question" comes from children's own personal experiences (e.g., feedback about their skills in art) rather than gender stereotypes.</p> <p>The present studies also demonstrate the effectiveness of unmoderated remote research (Rhodes, Rizzo, et al., [<reflink idref="bib60" id="ref168">60</reflink>]) for advancing developmental, social, and intervention science. This approach allowed us to recruit three large samples of racially and geographically diverse children to complete a multi‐session intervention, ensure that the intervention was administered with complete fidelity, collect measures from children as well as natural language data from parents via a single platform, and track the developmental trajectory of children's science stereotypes up to a year. In addition, the curriculum developed here is virtual and self‐guided; therefore, although the present studies were primarily limited to the United States and middle‐class, English‐speaking populations, this approach could be easily implemented and disseminated across broader populations (including, with technological support, children from economically disadvantaged communities, children in non‐Western contexts). Further, although the present research was focused on children, the curriculum had enduring effects on how parents referred to science in conversations with their children, thus illustrating how unmoderated remote research can lead to change that is sustained over time by modifying ongoing processes. More generally, this research highlights how modifying the language used to introduce science in children's books, television shows, classrooms, and beyond, could be an effective step toward creating more inclusive science learning environments for all children.</p> <hd id="AN0183920700-40">ACKNOWLEDGMENTS</hd> <p>We are grateful to the participating families. We thank Naffeza Ali, Ian Cropley, Catherine Hu, Christine Lott, Alexandria Pena, Maria Togonidze, and Ashley Winegarden for research assistance; Bob Rehder and members of the Conceptual Development and Social Cognition Lab for manuscript feedback; and Tessa West for statistical guidance. Research reported in this publication was supported by the National Science Foundation grant NSF EHR‐2000617.</p> <hd id="AN0183920700-41">FUNDING INFORMATION</hd> <p>National Science Foundation grant NSF EHR‐2000617 (MR).</p> <hd id="AN0183920700-42">DATA AVAILABILITY STATEMENT</hd> <p>All data, code, and materials are openly available on the project's Open Science Framework repository: https://osf.io/3hdxt/?view%5fonly=b0bc115f94364373822429344494435b. Additional data and analyses are reported in the supplemental materials. The analyses presented here were not preregistered.</p> <p>GRAPH: Data S1.</p> <ref id="AN0183920700-43"> <title> REFERENCES </title> <blist> <bibl id="bib1" idref="ref50" type="bt">1</bibl> <bibtext> Agars, M. D. (2004). Reconsidering the impact of gender stereotypes on the advancement of women in organizations. Psychology of Women Quarterly, 28, 103 – 111. https://doi.org/10.1111/j.1471‐6402.2004.00127.x</bibtext> </blist> <blist> <bibl id="bib2" idref="ref46" type="bt">2</bibl> <bibtext> Ainley, M., &amp; Ainley, J. (2011). Student engagement with science in early adolescence: The contribution of enjoyment to students' continuing interest in learning about science. Contemporary Educational Psychology, 36, 4 – 12. https://doi.org/10.1016/j.cedpsych.2010.08.001</bibtext> </blist> <blist> <bibl id="bib3" idref="ref36" type="bt">3</bibl> <bibtext> Baram‐Tsabari, A., Sethi, R. J., Bry, L., &amp; Yarden, A. (2006). Using questions sent to an Ask‐A‐Scientist site to identify children's interests in science. Science Education, 90, 1050 – 1072. https://doi.org/10.1002/sce.20163</bibtext> </blist> <blist> <bibl id="bib4" idref="ref84" type="bt">4</bibl> <bibtext> Bates, D., Mächler, M., Bolker, B., &amp; Walker, S. (2015). Fitting linear mixed‐effects models using lme4. Journal of Statistical Software, 67, 1 – 48. https://doi.org/10.48550/arXiv.1406.5823</bibtext> </blist> <blist> <bibl id="bib5" idref="ref3" type="bt">5</bibl> <bibtext> Bell, A., Chetty, R., Jaravel, X., Petkova, N., &amp; Van Reenen, J. (2019). Joseph Schumpeter Lecture, EEA Annual Congress 2017: Do tax cuts produce more Einsteins? The impacts of financial incentives versus exposure to innovation on the supply of inventors. Journal of the European Economic Association, 17, 651 – 677. https://doi.org/10.3386/w25493</bibtext> </blist> <blist> <bibl id="bib6" idref="ref14" type="bt">6</bibl> <bibtext> Bian, L., Leslie, S. J., &amp; Cimpian, A. (2017). Gender stereotypes about intellectual ability emerge early and influence children's interests. Science, 355, 389 – 391. https://doi.org/10.1126/science.aah6524</bibtext> </blist> <blist> <bibl id="bib7" idref="ref80" type="bt">7</bibl> <bibtext> Chambers, D. W. (1983). Stereotypic images of the scientist: The draw‐a‐scientist test. Science Education, 67, 255 – 265.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref77" type="bt">8</bibl> <bibtext> Charlesworth, T. E., &amp; Banaji, M. R. (2022). Patterns of implicit and explicit stereotypes III: Long‐term change in gender stereotypes. Social Psychological and Personality Science, 13, 14 – 26. https://doi.org/10.1177/1948550620988425</bibtext> </blist> <blist> <bibl id="bib9" idref="ref66" type="bt">9</bibl> <bibtext> Charlesworth, T. E., Yang, V., Mann, T. C., Kurdi, B., &amp; Banaji, M. R. (2021). Gender stereotypes in natural language: Word embeddings show robust consistency across child and adult language corpora of more than 65 million words. Psychological Science, 32, 218 – 240. https://doi.org/10.1177/0956797620963619</bibtext> </blist> <blist> <bibtext> Cheryan, S., Master, A., &amp; Meltzoff, A. N. (2015). Cultural stereotypes as gatekeepers: Increasing girls' interest in computer science and engineering by diversifying stereotypes. Frontiers in Psychology, 6, 1 – 8. https://doi.org/10.3389/fpsyg.2015.00049</bibtext> </blist> <blist> <bibtext> Chestnut, E. K., Zhang, M. Y., &amp; Markman, E. M. (2021). "Just as good": Learning gender stereotypes from attempts to counteract them. Developmental Psychology, 57, 114 – 125. https://doi.org/10.1037/dev0001143</bibtext> </blist> <blist> <bibtext> Cimpian, A., Arce, H. M. C., Markman, E. M., &amp; Dweck, C. S. (2007). Subtle linguistic cues affect children's motivation. Psychological Science, 18, 314 – 316. https://doi.org/10.1111/j.1467‐9280.2007.01896.x</bibtext> </blist> <blist> <bibtext> Cundiff, J. L., Vescio, T. K., Loken, E., &amp; Lo, L. (2013). Do gender–science stereotypes predict science identification and science career aspirations among undergraduate science majors? Social Psychology of Education, 16, 541 – 554. https://doi.org/10.1007/s11218‐013‐9232‐8</bibtext> </blist> <blist> <bibtext> Cvencek, D., Meltzoff, A. N., &amp; Greenwald, A. G. (2011). Math–gender stereotypes in elementary school children. Child Development, 82, 766 – 779. https://doi.org/10.1111/j.1467‐8624.2010.01529.x</bibtext> </blist> <blist> <bibtext> Cvencek, D., Nasir, N. I. S., O'Connor, K., Wischnia, S., &amp; Meltzoff, A. N. (2015). The development of math–race stereotypes: "They say Chinese people are the best at math". Journal of Research on Adolescence, 25, 630 – 637. https://doi.org/10.1111/jora.12151</bibtext> </blist> <blist> <bibtext> Desy, E. A., Peterson, S. A., &amp; Brockman, V. (2011). Gender differences in science‐related attitudes and interests among middle school and high school students. Science Educator, 20, 23 – 30.</bibtext> </blist> <blist> <bibtext> Faul, F., Erdfelder, E., Lang, A. G., &amp; Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175 – 191. https://doi.org/10.3758/BF03193146</bibtext> </blist> <blist> <bibtext> Ferrant, G., &amp; Nowacka, K. (2015). Measuring the drivers of gender inequality and their impact on development: The role of discriminatory social institutions. Gender and Development, 23, 319 – 332. https://doi.org/10.1080/13552074.2015.1053221</bibtext> </blist> <blist> <bibtext> Gelman, S. A. (2009). Learning from others: Children's construction of concepts. Annual Review of Psychology, 60, 115 – 140. https://doi.org/10.1146/annurev.psych.59.103006.093659</bibtext> </blist> <blist> <bibtext> Gelman, S. A., Ware, E. A., &amp; Kleinberg, F. (2010). Effects of generic language on category content and structure. Cognitive Psychology, 61, 273 – 301. https://doi.org/10.1016/j.cogpsych.2010.06.001</bibtext> </blist> <blist> <bibtext> Gelman, S. A., Ware, E. A., Kleinberg, F., Manczak, E. M., &amp; Stilwell, S. M. (2014). Individual differences in children's and parents' generic language. Child Development, 85, 924 – 940. https://doi.org/10.1111/cdev.12187</bibtext> </blist> <blist> <bibtext> Grossman, J. M., &amp; Porche, M. V. (2014). Perceived gender and racial/ethnic barriers to STEM success. Urban Education, 49, 698 – 727. https://doi.org/10.1177/0042085913481364</bibtext> </blist> <blist> <bibtext> Halim, M. L., Ruble, D. N., Tamis‐LeMonda, C. S., Zosuls, K. M., Lurye, L. E., &amp; Greulich, F. K. (2014). Pink frilly dresses and the avoidance of all things "girly": Children's appearance rigidity and cognitive theories of gender development. Developmental Psychology, 50, 1091 – 1101. https://doi.org/10.1037/a0034906</bibtext> </blist> <blist> <bibtext> Hall, C., Dickerson, J., Batts, D., Kauffmann, P., &amp; Bosse, M. (2011). Are we missing opportunities to encourage interest in STEM fields? Journal of Technology Education, 23, 32 – 46.</bibtext> </blist> <blist> <bibtext> Jansen, M., Scherer, R., &amp; Schroeders, U. (2015). Students' self‐concept and self‐efficacy in the sciences: Differential relations to antecedents and educational outcomes. Contemporary Educational Psychology, 41, 13 – 24. https://doi.org/10.1016/j.cedpsych.2014.11.002</bibtext> </blist> <blist> <bibtext> Jaxon, J., Lei, R. F., Shachnai, R., Chestnut, E. K., &amp; Cimpian, A. (2019). The acquisition of gender stereotypes about intellectual ability: Intersections with race. Journal of Social Issues, 75, 1192 – 1215. https://doi.org/10.1111/josi.12352</bibtext> </blist> <blist> <bibtext> Kang, J., Hense, J., Scheersoi, A., &amp; Keinonen, T. (2019). Gender study on the relationships between science interest and future career perspectives. International Journal of Science Education, 41, 80 – 101. https://doi.org/10.1080/09500693.2018.1534021</bibtext> </blist> <blist> <bibtext> Kerkhoven, A. H., Russo, P., Land‐Zandstra, A. M., Saxena, A., &amp; Rodenburg, F. J. (2016). Gender stereotypes in science education resources: A visual content analysis. PLoS One, 11, e0165037. https://doi.org/10.1371/journal.pone.0165037</bibtext> </blist> <blist> <bibtext> Kurtz‐Costes, B., Copping, K. E., Rowley, S. J., &amp; Kinlaw, C. R. (2014). Gender and age differences in awareness and endorsement of gender stereotypes about academic abilities. European Journal of Psychology of Education, 29, 603 – 618. https://doi.org/10.1007/s10212‐014‐0216‐7</bibtext> </blist> <blist> <bibtext> Lane, K. A., Goh, J. X., &amp; Driver‐Linn, E. (2012). Implicit science stereotypes mediate the relationship between gender and academic participation. Sex Roles, 66, 220 – 234. https://doi.org/10.1007/s11199‐011‐0036‐z</bibtext> </blist> <blist> <bibtext> Laosa, L. M., Swartz, J. D., &amp; Holtzman, W. H. (1973). Human figure drawings by normal children: A longitudinal study of perceptual‐cognitive and personality development. Developmental Psychology, 8, 350 – 356. https://doi.org/10.1037/h0034334</bibtext> </blist> <blist> <bibtext> Lei, R. F., Green, E. R., Leslie, S. J., &amp; Rhodes, M. (2019). Children lose confidence in their potential to "be scientists", but not in their capacity to "do science". Developmental Science, 22 (6), e12837. https://doi.org/10.1111/desc.12837</bibtext> </blist> <blist> <bibtext> Lei, R. F., Leshin, R. A., Moty, K., Foster‐Hanson, E., &amp; Rhodes, M. (2022). How race and gender shape the development of social prototypes in the United States. Journal of Experimental Psychology: General, 151, 1956 – 1971. https://doi.org/10.1037/xge0001164</bibtext> </blist> <blist> <bibtext> Lei, R. F., Leshin, R. A., &amp; Rhodes, M. (2020). The development of intersectional social prototypes. Psychological Science, 31 (8), 911 – 926. https://doi.org/10.1177/0956797620920360</bibtext> </blist> <blist> <bibtext> Leibham, M. B., Alexander, J. M., &amp; Johnson, K. E. (2013). Science interests in preschool boys and girls: Relations to later self‐concept and science achievement. Science Education, 97, 574 – 593. https://doi.org/10.1002/sce.21066</bibtext> </blist> <blist> <bibtext> Leslie, S. J., Cimpian, A., Meyer, M., &amp; Freeland, E. (2015). Expectations of brilliance underlie gender distributions across academic disciplines. Science, 347, 262 – 265. https://doi.org/10.1126/science.1261375</bibtext> </blist> <blist> <bibtext> Lewis, M., &amp; Lupyan, G. (2020). Gender stereotypes are reflected in the distributional structure of 25 languages. Nature Human Behaviour, 4, 1021 – 1028. https://doi.org/10.1038/s41562‐020‐0918‐6</bibtext> </blist> <blist> <bibtext> Liben, L. S., &amp; Bigler, R. S. (2002). The developmental course of gender differentiation: Conceptualizing, measuring, and evaluating constructs and pathways. Monographs of the Society for Research in Child Development, 67, 1 – 183. https://doi.org/10.1111/1540‐5834.00189</bibtext> </blist> <blist> <bibtext> Long, M., Steinke, J., Applegate, B., Knight Lapinski, M., Johnson, M. J., &amp; Ghosh, S. (2010). Portrayals of male and female scientists in television programs popular among middle school‐age children. Science Communication, 32, 356 – 382. https://doi.org/10.1177/1075547009357779</bibtext> </blist> <blist> <bibtext> Martin, C. L., &amp; Ruble, D. (2004). Children's search for gender cues: Cognitive perspectives on gender development. Current Directions in Psychological Science, 13, 67 – 70. https://doi.org/10.1111/j.0963‐7214.2004.00276.x</bibtext> </blist> <blist> <bibtext> Martinot, D., &amp; Désert, M. (2007). Awareness of a gender stereotype, personal beliefs and self‐perceptions regarding math ability: When boys do not surpass girls. Social Psychology of Education, 10, 455 – 471. https://doi.org/10.1007/s11218‐007‐9028‐9</bibtext> </blist> <blist> <bibtext> Master, A. (2021). Gender stereotypes influence children's STEM motivation. Child Development Perspectives, 15, 203 – 210. https://doi.org/10.1111/cdep.12424</bibtext> </blist> <blist> <bibtext> Master, A., &amp; Meltzoff, A. N. (2020). Cultural stereotypes and sense of belonging contribute to gender gaps in STEM. International Journal of Gender, Science and Technology, 12, 152 – 198.</bibtext> </blist> <blist> <bibtext> Master, A., Meltzoff, A. N., &amp; Cheryan, S. (2021). Gender stereotypes about interests start early and cause gender disparities in computer science and engineering. Proceedings of the National Academy of Sciences of the United States of America, 118, e2100030118. https://doi.org/10.1073/pnas.2100030118</bibtext> </blist> <blist> <bibtext> Miller, D. I., Nolla, K. M., Eagly, A. H., &amp; Uttal, D. H. (2018). The development of children's gender‐science stereotypes: A meta‐analysis of 5 decades of US Draw‐a‐Scientist studies. Child Development, 89, 1943 – 1955. https://doi.org/10.1111/cdev.13039</bibtext> </blist> <blist> <bibtext> Miller, P. H., Slawinski Blessing, J., &amp; Schwartz, S. (2006). Gender differences in high‐school students' views about science. International Journal of Science Education, 28, 363 – 381. https://doi.org/10.1080/09500690500277664</bibtext> </blist> <blist> <bibtext> Muradoglu, M., Horne, Z., Hammond, M. D., Leslie, S.‐J., &amp; Cimpian, A. (2022). Women—Particularly underrepresented minority women—And early‐career academics feel like impostors in fields that value brilliance. Journal of Educational Psychology, 114, 1086 – 1100. https://doi.org/10.1037/edu0000669</bibtext> </blist> <blist> <bibtext> National Science Foundation. (2002). Women, minorities, and persons with disabilities in science and engineering: 2002. U.S. Government Printing Office.</bibtext> </blist> <blist> <bibtext> Newton, D. P., &amp; Newton, L. D. (1992). Young children's perceptions of science and the scientist. International Journal of Science Education, 14, 331 – 348. https://doi.org/10.1080/0950069920140309</bibtext> </blist> <blist> <bibtext> Nosek, B. A., Banaji, M. R., &amp; Greenwald, A. G. (2002). Math = male, me = female, therefore math ≠ me. Journal of Personality and Social Psychology, 83, 44 – 59. https://doi.org/10.1037/0022‐3514.83.1.44</bibtext> </blist> <blist> <bibtext> Nosek, B. A., &amp; Smyth, F. L. (2011). Implicit social cognitions predict sex differences in math engagement and achievement. American Educational Research Journal, 48, 1125 – 1156. https://doi.org/10.3102/0002831211410683</bibtext> </blist> <blist> <bibtext> Nosek, B. A., Smyth, F. L., Sriram, N., Lindner, N. M., Devos, T., Ayala, A., Bar‐Anan, Y., Bergh, R., Cai, H., Gonsalkorale, K., Kesebir, S., Maliszewski, N., Neto, F., Olli, E., Park, J., Schnabel, K., Shiomura, K., Tulbure, B. T., Wiers, R. W., ... Greenwald, A. G. (2009). National differences in gender–science stereotypes predict national sex differences in science and math achievement. Proceedings of the National Academy of Sciences of the United States of America, 106, 10593 – 10597. https://doi.org/10.1073/pnas.0809921106</bibtext> </blist> <blist> <bibtext> O'Dea, R. E., Lagisz, M., Jennions, M. D., &amp; Nakagawa, S. (2018). Gender differences in individual variation in academic grades fail to fit expected patterns for STEM. Nature Communications, 9, 3777. https://doi.org/10.1038/s41467‐018‐06292‐0</bibtext> </blist> <blist> <bibtext> Plante, I., De la Sablonnière, R., Aronson, J. M., &amp; Théorêt, M. (2013). Gender stereotype endorsement and achievement‐related outcomes: The role of competence beliefs and task values. Contemporary Educational Psychology, 38, 225 – 235. https://doi.org/10.1016/j.cedpsych.2013.03.004</bibtext> </blist> <blist> <bibtext> Plante, I., O'Keefe, P. A., Aronson, J., Fréchette‐Simard, C., &amp; Goulet, M. (2019). The interest gap: How gender stereotype endorsement about abilities predicts differences in academic interests. Social Psychology of Education, 22, 227 – 245. https://doi.org/10.1007/s11218‐018‐9472‐8</bibtext> </blist> <blist> <bibtext> R Core Team. (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R‐project.org/</bibtext> </blist> <blist> <bibtext> Rhodes, M., Cardarelli, A., &amp; Leslie, S. J. (2020). Asking young children to "do science" instead of "be scientists" increases science engagement in a randomized field experiment. Proceedings of the National Academy of Sciences of the United States of America, 117, 9808 – 9814. https://doi.org/10.1073/pnas.1919646117</bibtext> </blist> <blist> <bibtext> Rhodes, M., Leslie, S. J., &amp; Tworek, C. M. (2012). Cultural transmission of social essentialism. Proceedings of the National Academy of Sciences of the United States of America, 109, 13526 – 13531. https://doi.org/10.1073/pnas.1208951109</bibtext> </blist> <blist> <bibtext> Rhodes, M., Leslie, S. J., Yee, K. M., &amp; Saunders, K. (2019). Subtle linguistic cues increase girls' engagement in science. Psychological Science, 30, 455 – 466. https://doi.org/10.1177/0956797618823670</bibtext> </blist> <blist> <bibtext> Rhodes, M., Rizzo, M. T., Foster‐Hanson, E., Moty, K., Leshin, R. A., Wang, M., Benitez, J., &amp; Ocampo, J. D. (2020). Advancing developmental science via unmoderated remote research with children. Journal of Cognition and Development, 21, 477 – 493. https://doi.org/10.1080/15248372.2020.1797751</bibtext> </blist> <blist> <bibtext> Sadler, P. M., Sonnert, G., Hazari, Z., &amp; Tai, R. (2012). Stability and volatility of STEM career interest in high school: A gender study. Science Education, 96, 411 – 427. https://doi.org/10.1002/sce.21007</bibtext> </blist> <blist> <bibtext> Sax, L. J. (2001). Undergraduate science majors: Gender differences in who goes to graduate school. The Review of Higher Education, 24, 153 – 172. https://doi.org/10.1353/rhe.2000.0030</bibtext> </blist> <blist> <bibtext> Shu, Y., Hu, Q., Xu, F., &amp; Bian, L. (2022). Gender stereotypes are racialized: A cross‐cultural investigation of gender stereotypes about intellectual talents. Developmental Psychology, 58, 1345 – 1359. https://doi.org/10.1037/dev0001356</bibtext> </blist> <blist> <bibtext> Starr, C. R., &amp; Simpkins, S. D. (2021). High school students' math and science gender stereotypes: Relations with their STEM outcomes and socializers' stereotypes. Social Psychology of Education, 24, 273 – 298. https://doi.org/10.1007/s11218‐021‐09611‐4</bibtext> </blist> <blist> <bibtext> Steele, J. (2003). Children's gender stereotypes about math: The role of stereotype stratification 1. Journal of Applied Social Psychology, 33, 2587 – 2606. https://doi.org/10.1111/j.1559‐1816.2003.tb02782.x</bibtext> </blist> <blist> <bibtext> Steffens, M. C., Jelenec, P., &amp; Noack, P. (2010). On the leaky math pipeline: Comparing implicit math‐gender stereotypes and math withdrawal in female and male children and adolescents. Journal of Educational Psychology, 102, 947 – 963. https://doi.org/10.1037/a0019920</bibtext> </blist> <blist> <bibtext> Vaish, A., Grossmann, T., &amp; Woodward, A. (2008). Not all emotions are created equal: The negativity bias in social‐emotional development. Psychological Bulletin, 134, 383 – 403. https://doi.org/10.1037/0033‐2909.134.3.383</bibtext> </blist> <blist> <bibtext> Wang, M. M., Cardarelli, A., Leslie, S.‐J., &amp; Rhodes, M. (2022). How children's media and teachers communicate exclusive and essentialist views of science and scientists. Developmental Psychology, 58, 1455 – 1471. https://doi.org/10.1037/dev0001364</bibtext> </blist> <blist> <bibtext> Waxman, S. R. (2010). Names will never hurt me? Naming and the development of racial and gender categories in preschool‐aged children. European Journal of Social Psychology, 40, 593 – 610. https://doi.org/10.1002/ejsp.732</bibtext> </blist> <blist> <bibtext> Zhao, S., Setoh, P., Storage, D., &amp; Cimpian, A. (2022). The acquisition of the gender‐brilliance stereotype: Age trajectory, relation to parents' stereotypes, and intersections with race/ethnicity. Child Development, 93, e581 – e597. https://doi.org/10.1111/cdev.13809</bibtext> </blist> <blist> <bibtext> Zitelny, H., Shalom, M., &amp; Bar‐Anan, Y. (2017). What is the implicit gender‐science stereotype? Exploring correlations between the gender‐science IAT and self‐report measures. Social Psychological and Personality Science, 8, 719 – 735. https://doi.org/10.1177/1948550616683017</bibtext> </blist> </ref> <aug> <p>By Michelle M. Wang; Amanda Cardarelli; Jonah Brenner; Sarah‐Jane Leslie and Marjorie Rhodes</p> <p>Reported by Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib10" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib44" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib18" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib68" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib42" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib45" firstref="ref8"></nolink> <nolink nlid="nl7" bibid="bib64" firstref="ref9"></nolink> <nolink nlid="nl8" bibid="bib70" firstref="ref10"></nolink> <nolink nlid="nl9" bibid="bib14" firstref="ref13"></nolink> <nolink nlid="nl10" bibid="bib15" firstref="ref15"></nolink> <nolink nlid="nl11" bibid="bib29" firstref="ref16"></nolink> <nolink nlid="nl12" bibid="bib35" firstref="ref22"></nolink> <nolink nlid="nl13" bibid="bib49" firstref="ref23"></nolink> <nolink nlid="nl14" bibid="bib62" firstref="ref24"></nolink> <nolink nlid="nl15" bibid="bib54" firstref="ref25"></nolink> <nolink nlid="nl16" bibid="bib55" firstref="ref28"></nolink> <nolink nlid="nl17" bibid="bib66" firstref="ref29"></nolink> <nolink nlid="nl18" bibid="bib13" firstref="ref30"></nolink> <nolink nlid="nl19" bibid="bib30" firstref="ref31"></nolink> <nolink nlid="nl20" bibid="bib51" firstref="ref32"></nolink> <nolink nlid="nl21" bibid="bib59" firstref="ref38"></nolink> <nolink nlid="nl22" bibid="bib16" firstref="ref40"></nolink> <nolink nlid="nl23" bibid="bib46" firstref="ref41"></nolink> <nolink nlid="nl24" bibid="bib48" firstref="ref42"></nolink> <nolink nlid="nl25" bibid="bib61" firstref="ref43"></nolink> <nolink nlid="nl26" bibid="bib53" firstref="ref45"></nolink> <nolink nlid="nl27" bibid="bib24" firstref="ref47"></nolink> <nolink nlid="nl28" bibid="bib27" firstref="ref48"></nolink> <nolink nlid="nl29" bibid="bib65" firstref="ref56"></nolink> <nolink nlid="nl30" bibid="bib11" firstref="ref58"></nolink> <nolink nlid="nl31" bibid="bib20" firstref="ref60"></nolink> <nolink nlid="nl32" bibid="bib58" firstref="ref61"></nolink> <nolink nlid="nl33" bibid="bib69" firstref="ref62"></nolink> <nolink nlid="nl34" bibid="bib19" firstref="ref63"></nolink> <nolink nlid="nl35" bibid="bib23" firstref="ref64"></nolink> <nolink nlid="nl36" bibid="bib40" firstref="ref65"></nolink> <nolink nlid="nl37" bibid="bib28" firstref="ref67"></nolink> <nolink nlid="nl38" bibid="bib37" firstref="ref68"></nolink> <nolink nlid="nl39" bibid="bib57" firstref="ref69"></nolink> <nolink nlid="nl40" bibid="bib17" firstref="ref75"></nolink> <nolink nlid="nl41" bibid="bib173" firstref="ref76"></nolink> <nolink nlid="nl42" bibid="bib50" firstref="ref78"></nolink> <nolink nlid="nl43" bibid="bib52" firstref="ref79"></nolink> <nolink nlid="nl44" bibid="bib71" firstref="ref103"></nolink> <nolink nlid="nl45" bibid="bib41" firstref="ref105"></nolink> <nolink nlid="nl46" bibid="bib43" firstref="ref106"></nolink> <nolink nlid="nl47" bibid="bib32" firstref="ref110"></nolink> <nolink nlid="nl48" bibid="bib365" firstref="ref113"></nolink> <nolink nlid="nl49" bibid="bib56" firstref="ref119"></nolink> <nolink nlid="nl50" bibid="bib21" firstref="ref125"></nolink> <nolink nlid="nl51" bibid="bib22" firstref="ref128"></nolink> <nolink nlid="nl52" bibid="bib34" firstref="ref129"></nolink> <nolink nlid="nl53" bibid="bib47" firstref="ref130"></nolink> <nolink nlid="nl54" bibid="bib163" firstref="ref131"></nolink> <nolink nlid="nl55" bibid="bib38" firstref="ref139"></nolink> <nolink nlid="nl56" bibid="bib39" firstref="ref142"></nolink> <nolink nlid="nl57" bibid="bib25" firstref="ref148"></nolink> <nolink nlid="nl58" bibid="bib31" firstref="ref150"></nolink> <nolink nlid="nl59" bibid="bib33" firstref="ref151"></nolink> <nolink nlid="nl60" bibid="bib67" firstref="ref154"></nolink> <nolink nlid="nl61" bibid="bib36" firstref="ref162"></nolink> <nolink nlid="nl62" bibid="bib26" firstref="ref163"></nolink> <nolink nlid="nl63" bibid="bib63" firstref="ref164"></nolink> <nolink nlid="nl64" bibid="bib12" firstref="ref167"></nolink> <nolink nlid="nl65" bibid="bib60" firstref="ref168"></nolink> |
|---|---|
| Header | DbId: eric DbLabel: ERIC An: EJ1461468 AccessLevel: 3 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Maladaptive but Malleable: Gender-Science Stereotypes Emerge Early but are Modifiable by Language – Name: Language Label: Language Group: Lang Data: English – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Michelle+M%2E+Wang%22">Michelle M. Wang</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0001-7256-4268">0000-0001-7256-4268</externalLink>)<br /><searchLink fieldCode="AR" term="%22Amanda+Cardarelli%22">Amanda Cardarelli</searchLink><br /><searchLink fieldCode="AR" term="%22Jonah+Brenner%22">Jonah Brenner</searchLink><br /><searchLink fieldCode="AR" term="%22Sarah-Jane+Leslie%22">Sarah-Jane Leslie</searchLink><br /><searchLink fieldCode="AR" term="%22Marjorie+Rhodes%22">Marjorie Rhodes</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-4664-6056">0000-0002-4664-6056</externalLink>) – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="SO" term="%22Child+Development%22"><i>Child Development</i></searchLink>. 2025 96(2):865-880. – Name: Avail Label: Availability Group: Avail Data: Wiley. Available from: John Wiley & Sons, Inc. 111 River Street, Hoboken, NJ 07030. Tel: 800-835-6770; e-mail: cs-journals@wiley.com; Web site: https://www.wiley.com/en-us – Name: PeerReviewed Label: Peer Reviewed Group: SrcInfo Data: Y – Name: Pages Label: Page Count Group: Src Data: 16 – Name: DatePubCY Label: Publication Date Group: Date Data: 2025 – Name: SourceSuprt Label: Sponsoring Agency Group: SrcSuprt Data: National Science Foundation (NSF), Directorate for Education and Human Resources (EHR) – Name: NumberContract Label: Contract Number Group: NumCntrct Data: 2000617 – Name: TypeDocument Label: Document Type Group: TypDoc Data: Journal Articles<br />Reports - Research – Name: Subject Label: Descriptors Group: Su Data: <searchLink fieldCode="DE" term="%22Sex+Stereotypes%22">Sex Stereotypes</searchLink><br /><searchLink fieldCode="DE" term="%22Scientists%22">Scientists</searchLink><br /><searchLink fieldCode="DE" term="%22Preschool+Children%22">Preschool Children</searchLink><br /><searchLink fieldCode="DE" term="%22Sciences%22">Sciences</searchLink><br /><searchLink fieldCode="DE" term="%22Language+Usage%22">Language Usage</searchLink><br /><searchLink fieldCode="DE" term="%22Intervention%22">Intervention</searchLink> – Name: DOI Label: DOI Group: ID Data: 10.1111/cdev.14213 – Name: ISSN Label: ISSN Group: ISSN Data: 0009-3920<br />1467-8624 – Name: Abstract Label: Abstract Group: Ab Data: Gender-science stereotypes emerge early in childhood, but little is known about the developmental processes by which they arise. The present study tested the hypothesis that language implying scientists are a special and distinct kind of person contributes to the development of gender-science stereotypes, even when it does not communicate stereotypic content. One cross-sectional and two longitudinal studies with racially and geographically diverse children (primarily from the United States; ages 4-5; N = 872, tested 2020-2022) revealed that gender stereotypes about science versus art (a) emerge before elementary school, (b) arise from commonplace identity-emphasizing language, especially among girls, and (c) can be durably disrupted by subtle changes to language. This study identifies a promising way to counteract stereotypes at their roots. – Name: AbstractInfo Label: Abstractor Group: Ab Data: As Provided – Name: Note Label: Notes Group: Note Data: https://osf.io/3hdxt/?view_only=b0bc115f94364373822429344494435b – Name: DateEntry Label: Entry Date Group: Date Data: 2025 – Name: AN Label: Accession Number Group: ID Data: EJ1461468 |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1461468 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1111/cdev.14213 Languages: – Text: English PhysicalDescription: Pagination: PageCount: 16 StartPage: 865 Subjects: – SubjectFull: Sex Stereotypes Type: general – SubjectFull: Scientists Type: general – SubjectFull: Preschool Children Type: general – SubjectFull: Sciences Type: general – SubjectFull: Language Usage Type: general – SubjectFull: Intervention Type: general Titles: – TitleFull: Maladaptive but Malleable: Gender-Science Stereotypes Emerge Early but are Modifiable by Language Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Michelle M. Wang – PersonEntity: Name: NameFull: Amanda Cardarelli – PersonEntity: Name: NameFull: Jonah Brenner – PersonEntity: Name: NameFull: Sarah-Jane Leslie – PersonEntity: Name: NameFull: Marjorie Rhodes IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 03 Type: published Y: 2025 Identifiers: – Type: issn-print Value: 0009-3920 – Type: issn-electronic Value: 1467-8624 Numbering: – Type: volume Value: 96 – Type: issue Value: 2 Titles: – TitleFull: Child Development Type: main |
| ResultId | 1 |
