Testing the impact of two afterschool museum outreach interventions on elementary children’s STEM outcomes: hands-on STEM alone or with STEM stories

Introduction: This randomized control trial (RCT) evaluated an afterschool program with 24 weeks of hands-on science, technology, engineering and math (STEM) activities that were developed and delivered by museum educators at sites where the majority of students experienced poverty. The program also encouraged parent involvement in STEM-related activities.

Methods: We contrasted two treatment approaches to understand conditions that best support informal STEM learning during the elementary school period of kindergarten to Grade 5 (K-5). Specifically, this included a 60-min weekly “Basic” program, as well as an enhanced “Stories” version with an additional 15-min read-aloud featuring women and girls doing STEM activities. We randomized 36 afterschool sites with 541 children to determine the benefits of the basic afterschool program and the added benefits of the stories on STEM attitudes, aspects of science achievement, and other outcomes.

Results: There were improvements or interaction effects in three of the five measured child outcomes: STEM value, career aspirations, and perceived math abilities. The most notable finding was that children’s STEM value increased significantly in the Basic treatment (effect size, g = 0.25, p = 0.027) and trended positively in the Stories condition (g = 0.18, p = 0.66), relative to the control group. Parent involvement in STEM also showed some changes.

Discussion: These findings underscore the potential for informal, hands-on experiences to positively influence children’s STEM-related attitudes. However, findings also suggest that interventions may need to be more intensive or sustained across the late elementary grades to achieve more substantial impact.

Introduction

In U.S. public schools, 85% offer afterschool programs (U.S. Department of Education et al., 2024) and one in three elementary children attend these programs (National Center for Education Statistics [NCES], 2023). About 42% of U.S. schools offer academically-focused afterschool programs, with a slightly higher percentage (52%) of schools in higher poverty neighborhoods currently reporting special funds for academically-focused afterschool enrichment programs (e.g., COVID pandemic recovery funds; National Center for Education Statistics [NCES], 2024). For these reasons, afterschool programs provide an ideal outside-of-school setting to explore how to best broaden access to early science, technology, engineering, and math (STEM) opportunities for girls and boys. We used a rigorous design to evaluate the causal impacts of two afterschool STEM approaches compared to a business-as-usual condition, aligning with priorities of policymakers, program developers, and other stakeholders who emphasize rigorous research to understand what works and conditions that best support informal STEM learning (National Research Council [NRC], 2015).

First, we evaluated how girls and boys benefited from a hands-on afterschool STEM program designed by museum-based informal learning experts. Relative to other STEM-related experiences, afterschool STEM programs offer important, accessible experiences for children experiencing poverty who might not have other opportunities for STEM enrichment experiences otherwise (Allen et al., 2019; National Research Council [NRC], 2009). Second, we evaluated an approach that added stories of women and girls doing STEM activities. Given that narratives are theorized to increase attitudes such as agency and identity as someone who can do STEM (e.g., Adler, 2012; Kim et al., 2018), we examined whether layering in stories of historical women in STEM careers, as well as fictional narratives of girls doing STEM activities, would improve outcomes. Recognizing the importance of families in promoting STEM interest pathways for girls and boys experiencing economic disadvantage (e.g., Pattison and Ramos Montañez, 2022), we also offered simple parent supports to extend learning at home. The next sections explain the theoretical underpinnings and past research that guided our evaluation.

Theoretical frameworks for understanding children’s STEM attitudes

This study is grounded in theoretical perspectives that include contextual factors that shape children’s STEM experiences. Ecological frameworks recognize that children benefit from both formal and informal learning experiences at school, home, and in-between places (Bronfenbrenner and Morris, 2006). Within informal spaces, adults are key socializers (e.g., Vygotsky, 1978) who provide activities and conversations to build STEM knowledge, guide inquiry, explain observations, and plan future investigations (e.g., Allen et al., 2019; Eberbach and Crowley, 2005, 2017; Haden, 2010). Inquiry-based STEM activities focus on asking questions, observing, and advanced reasoning, with formats that range from structured inquiry to open inquiry (Bell et al., 2005). The afterschool STEM program in this study used structured, hands-on inquiry approaches alongside STEM extensions families could explore at home or at a local museum. The program’s primary goals were to promote children’s positive STEM attitudes (e.g., I like STEM activities; Science is fun; I am good at math).

Our evaluation utilized the Situated Expectancy-Value Theory (SEVT), developed by Eccles and Wigfield (2023) and extensively applied in STEM research and programming, which posits that an individual’s behaviors and achievements are influenced by their expectancy of success and the perceived value of the task. These expectancy-value attitudes are situated in the child’s unique social, cultural, and contextual experiences (Eccles and Wigfield, 2020). Specifically, the first key construct within our use of SEVT is expectancy of success in STEM, such as attitudes that “I will do well in science/math activities or careers.” The second central construct within SEVT, value, includes enjoyment and interest in doing STEM activities, such as “I like doing science/math,” or “Science is important” (Ball et al., 2017; Rosenzweig and Chen, 2023). These expectancy and value factors are linked to STEM achievement and pursuit of STEM courses and careers (e.g., Fernandez et al., 2024; Rosenzweig and Chen, 2023; Wang and Degol, 2013). Expectancy and value aspects of STEM motivation have been used in diverse study designs and age groups from elementary-age students through college and career stages (e.g., Fielding-Wells et al., 2017; Master et al., 2025; Morgan et al., 2013; Sullivan and Bers, 2019). Thus, SEVT provided a useful lens for this study to evaluate the extent to which a museum outreach, afterschool STEM program improved students’ attitudes of STEM expectancy and value.

The importance of early experiences for children’s STEM outcomes

Reasons for broadening access to informal STEM experiences are that fostering early, positive STEM attitudes and competencies can, in turn, promote both immediate and longer-term STEM learning or pursuit of STEM-related courses or careers (National Research Council [NRC], 2015). Substantial evidence shows that providing early experiences that improve children’s STEM interest, knowledge, or career pathways is a promising approach for enhancing longer-term STEM outcomes (e.g., Maltese and Cooper, 2017; Maltese and Tai, 2011; Wang and Degol, 2013). It is well-established that early STEM knowledge relates to long-term STEM interest (e.g., Bustamante et al., 2023; Duncan et al., 2007; Kaderavek et al., 2020; Pellegrini et al., 2021). However, formal learning during the typical U.S. school day is often not of sufficient intensity or rigor to promote long-term science and math achievement as well as STEM career pathways for children experiencing poverty (e.g., Curran and Kitchin, 2019; Morgan et al., 2016). Thus, considerable investments have been made in understanding how informal learning enhances STEM knowledge and the factors that contribute to child outcomes, including students from underrepresented backgrounds envisioning pathways in which their future self succeeds in STEM fields (e.g., Dorsen et al., 2006; National Research Council [NRC], 2015; Offenstein et al., 2009).

Research on afterschool STEM programs

There is variation across U.S. states in the systems and quality of afterschool STEM programs. About 78% of afterschool programs deliver STEM programs (Afterschool Alliance, 2024). A large multistate study showed that most students retrospectively report these experiences improved their STEM engagement, career interest and other skills, particularly when students participated in quality programs (Allen et al., 2019). Meta-analyses with less than 20 studies showed that informal STEM programs (e.g., afterschool, summer camps, clubs) promote children’s STEM interest with small mean effect sizes (0.21–0.39), but few studies have taken place with elementary students and in the afterschool space (Xia et al., 2025; Young et al., 2017). Aggregated effect sizes for elementary students were smaller than for middle and high schoolers (Xia et al., 2025). Researchers evaluating afterschool STEM programs with older middle and high school students—with a variety of engineering, robotics, science experiences, or museum field trips—consistently report increases in student STEM motivation and career interests, but call for additional research with larger samples, robust measures, and causal impact approaches (e.g., Baran et al., 2019; Chittum et al., 2017; Meschede et al., 2022). This study adds to prior research on afterschool STEM by using a randomized control trial with validated observation and child measures for an age group of elementary students where there has been a paucity of research.

STEM career pathways

About 24% of the U.S. workforce is employed in STEM occupations, and there are improving diversification trends in recent years, with women representing 68% of workers in these occupations (National Center for Education Statistics [NCES], 2023). However, these averages may mask continued STEM pathway issues where girls and women who show promising STEM abilities choose not to pursue careers in STEM fields (Holian and Kelly, 2020; National Center for Science and Engineering Statistics [NCSES], 2023). For example, STEM fields like physics, engineering, and computer science are dominated by men in ways that are unexpected given women’s comparable achievement potential in these areas (Cimpian et al., 2020). This problem might be addressed by improving early motivational attitudes about STEM that can support pathways to longer-term interest in STEM fields (Le and Robbins, 2016). Yet a review of over 500 interventions designed to promote girls’ and women’s pursuit of STEM careers found that less than 10 studies used experimental designs that produce causal effects (van den Hurk et al., 2019). This study adds to the literature by using an experimental design to understand the short-term impacts of an afterschool STEM program on elementary-aged girls’ and boys’ STEM-related attitudes, career aspirations, and aspects of STEM achievement.

Although children’s early occupational interests are likely to change, exploring their early interests in careers may be meaningful. Adults ask children what they want to be when they grow up in order to provide aligned enrichment and academic opportunities and to convey their own aspirations or expectations for the child (Irwin and Elley, 2013). Children’s awareness of diverse STEM careers is key to successful future STEM pathways (Jiang et al., 2024). Informal educators and parents may shape children’s career awareness and aspirations. For example, when families encourage their child to take part in early STEM hobbies, programs, clubs, and museum visits, this can support later STEM interests and career decisions (e.g., Callanan et al., 2017; Dabney et al., 2016). Outside-of-school exposure to STEM programs can generate STEM interest and increase awareness of potential STEM careers (e.g., Goff et al., 2020; Habig et al., 2020; VanMeter-Adams et al., 2014). The current study emphasizes the importance of children’s STEM career expectancies during the elementary grades by investigating their interests in both core STEM fields (e.g., scientists, engineers) as well as fields adjacent to STEM that require applications of math or science knowledge and skills (e.g., health care workers, electricians, architects).

The elementary grades may be a key developmental period for bolstering STEM interest, as this is an initial period of identifying vocational interests (Low et al., 2005). Adults who hold STEM careers often retrospectively recount how their interests in STEM were shaped by early formal and informal learning experiences (e.g., Calabrese Barton et al., 2013; Dabney et al., 2016). A large, longitudinal study in Croatia found that STEM career interest was generally low and stable from middle childhood onward, despite slight decreases in science interest from elementary to middle school for all students and slight increases in engineering interest for boys during this transition (Babarović, 2022). Although STEM career tracks in the U.S. are less specified at the high school level, women with STEM career readiness in high school do not maintain pathways toward STEM careers at the same rates as men through college and mid-career (Speer, 2023). Taken together, STEM attitudes and career goals may be stable from the late elementary grades onward, thus making the elementary grades an important developmental period to evaluate the efficacy of STEM programs.

Promoting positive perceptions of girls’ and boys’ STEM abilities

Creating inclusive spaces for both girls and boys is considered an essential quality standard for STEM learning experiences (e.g., National Center for Education Statistics [NCES], 2023; United Nations Children International Children’s Fund [UNICEF], 2020). SEVT recognizes that children’s experiences are situated within a larger cultural milieu that includes unique perceptions about girls’ and boys’ respective abilities in STEM (Eccles and Wigfield, 2020). For example, children become increasingly aware over time that other people hold gender-specific stereotypes about ability in math, science, and engineering (also referred to as stereotype awareness or knowledge), whether or not they personally believe in or endorse them (often referred to as stereotype endorsement or personal belief; Kurtz-Costes et al., 2014). Although children are aware of gender-specific ability perceptions in society, we focus on students’ perceived STEM ability for girls and boys (personal endorsement of those beliefs), as this is malleable. Multiple studies report longitudinal relations between perceived STEM ability endorsement and children’s decisions to enroll in STEM courses or to persist in STEM careers (e.g., Riegle-Crumb and Peng, 2021; Sansone, 2019).

Perceptions of gender-specific STEM abilities emerge in the elementary school grades and can have long-term implications for STEM motivation, achievement, and career pathway decisions (Miller et al., 2024; Starr and Simpkins, 2021; Steffens et al., 2010). Specifically, beliefs that boys are more capable or more interested in STEM are linked to girls’ lower expectations of success and interest in STEM in late elementary and middle school (Cvencek et al., 2014; Martinot and Désert, 2007; Master et al., 2021, 2025). A meta-analysis of 98 studies found that children and adolescents’ perception of STEM-abilities slightly favored boys for math and science (Miller et al., 2024); however, there were gender differences and interactions with age, such that younger girls, in particular, were more likely to report perception of STEM-ability favoring their own group (i.e., in-group bias). The current study recruited students across the elementary grades because this is a period when perceptions of STEM abilities may begin to influence STEM academic and occupational outcomes. For example, students who perceive their gender as having lower STEM ability view themselves as having lower STEM ability as well (Master, 2021). Elementary students were also chosen because their perceptions of others’ and their own STEM abilities can be positively impacted by intervention (e.g., growth mindset training to improve perceived math ability, see Lee et al., 2021). We also recruited afterschool sites where most children were experiencing poverty because negative perceptions of STEM ability may be particularly disruptive to children from low-income backgrounds (e.g., Heberle and Carter, 2015; Miller et al., 2024).

Theoretically, both implicit and explicit strategies could reduce endorsement of gender-specific STEM ability. Implicit strategies include broadly welcoming approaches for all learners (e.g., creating spaces with roughly equal numbers of girls and boys; providing activities both girls and boys find engaging or relevant; avoiding gendered activities; Burgstahler and Cauce, 2020; Master et al., 2025), whereas explicit strategies directly address how girls and women take part in STEM (e.g., learning about stories of women who hold STEM careers; examining women’s historical impact in STEM fields; Sullivan, 2019). Explicit strategies to disrupt gender-specific STEM ability perceptions have been explored with older, middle- and high-school girls, such as offering informal experiences that specifically target girls (e.g., Bleacher et al., 2014; Santiago et al., 2019; Weisgram and Bigler, 2006). Other explicit approaches provide girls with opportunities to interact with women in STEM careers as role models (Olsson and Martiny, 2018; Riegle-Crumb et al., 2017; Steinke and Duncan, 2023). The current experiment contrasted a business-as-usual condition against two variations of an afterschool STEM program, with the two treatments differing only in that one used explicit strategies of adding stories to promote positive early STEM perceptions for girls and boys. While both treatment conditions included implicit strategies (e.g., approximately equal numbers of girls and boys, providing non-gendered activities that both girls and boys likely find engaging), we tested the effect of explicitly exposing children to stories highlighting women and girls succeeding in STEM by randomly assigned half of the treatment sites to a story-enhanced version of the STEM program.

Story-enhanced discussions within informal STEM learning

Our goal was to improve perceptions of gender-specific STEM abilities for girls alongside boys by sharing STEM-related stories with female role models (e.g., Bartholomew and Santana, 2021; Marcus et al., 2023). Specifically, after the hands-on STEM experience, boys and girls discussed a text read-aloud or a video featuring either a non-fiction story of women in STEM careers (e.g., Amelia Earhart, Mae Jemison) or a fictional narrative that included girls doing STEM activities (e.g., picture book titled Rosie Revere, Engineer by Andrea Beaty). Emerging evidence suggests that using narratives in STEM instruction may enhance STEM outcomes such as interest and recall (Golke and Wittwer, 2024; Tobler et al., 2024). Stories are also thought to motivate learners by helping them make sense of hands-on learning activities or content and providing a common language for communication with non-expert audiences about STEM topics (e.g., Dahlstrom, 2014; Leech et al., 2020; Shaby et al., 2025). Despite the potential for enhancing hands-on STEM activities with stories, further research is needed to document the feasibility of implementation alongside its causal impacts, as this involves practical challenges such as additional educator preparation and instructional time (e.g., Walan, 2019). Our work differed from other informal storytelling approaches that focus on children telling their own stories as they engage STEM with their families (Acosta and Haden, 2023; Marcus et al., 2023) because our family extensions were a light-touch addition to the primary, afterschool program.

Learners can find stories in science particularly motivating when they are able to see aspects of themselves in the narrative (Neeley et al., 2020). In this study, we added stories of women and girls doing STEM activities to promote children’s ability to identify with these narratives. For example, each unit included read-alouds of an illustrated children’s book, where about half featured historical women and half were fictional characters that featured girls or girls alongside boys. Read-alouds are an extensively studied context for promoting various child outcomes such as vocabulary and content knowledge (for reviews, see Dowdall et al., 2019; Mol et al., 2009), particularly when texts feature content-specific vocabulary (e.g., Cabell and Hwang, 2020; Neuman et al., 2021; Wright et al., 2022). A set of videos of women in STEM careers (e.g., chemist, astronomer, volcanologist) were also used in this study to expose children to explicit messages that girls and women can do STEM, which could, in turn, reduce the likelihood of STEM ability perceptions favoring boys (e.g., Tang et al., 2024).

Museum-based, informal STEM educators curated the content for this study’s read-alouds, and they facilitated interactive discussions about these stories, as well as other aspects of the afterschool STEM program. Informal STEM educators are experts in creating engaging STEM learning within informal spaces like afterschool programs, museums, libraries, or science centers. Informal STEM educators’ background and training require competencies to deliver accurate STEM information using engaging strategies (Morrissey, 2020; Mulvey et al., 2020). Although most STEM-related stories research to date has focused on families (e.g., Acosta and Haden, 2023; Marcus et al., 2023) or classroom teachers (e.g., French, 2004; Neuman et al., 2021), there is other evidence that informal STEM educators can guide learning in ways that improve learning (e.g., Carol-Ann Burke, 2020; Mulvey et al., 2020).

Family roles in STEM perceptions and pathways

Finally, we were interested in using simple approaches to engage families in extending STEM learning at home because guiding children’s learning is an important parent socialization practice (Grusec and Davidov, 2010). Families are often the purveyors of informal STEM experiences in ways that may relate to students’ achievement or career pathways (Eccles, 2015). It is well-documented that children’s STEM learning is enhanced when families receive resources to support their child’s science and engineering interests and explorations (e.g., Haden et al., 2014; Kaderavek et al., 2020; Marcus et al., 2017, 2018). This project invited families of participating children to visit their local children’s museum because museums provide a unique context for families to bolster their child’s interest in STEM (e.g., Anderson et al., 2002; Callanan et al., 2017; Haden et al., 2014). We recognized that museum visits alone were not of sufficient intensity to influence STEM career pathways (Suter, 2014), but that broadening access to museums was important for families experiencing poverty (Crispin and Beck, 2025).

We also used a text messaging approach to communicate with families during this afterschool program. Although more intensive family approaches are often needed to impact student achievement (Grindal et al., 2016), there are other benefits to light-touch approaches that provide information about STEM learning and prompt at-home activities (e.g., Haden et al., 2014; Marcus et al., 2018). Light-touch approaches may enhance STEM outcomes, particularly for families who are experiencing poverty or have less experience in STEM (Edwards and Danridge, 2020; Ennes et al., 2023; Zucker et al., 2022). For example, text messaging approaches that encourage parents to participate in hands-on STEM activities at home are promising (e.g., Hurwitz et al., 2015; Snell et al., 2020; Zucker et al., 2024). The current study included family text messages to explain what their child was learning in the afterschool STEM program and offered tips and linked activities to explore similar STEM concepts at home.

Current study goals and expectations

This project evaluated a program called Afterschool Science, Technology, Engineering, Arts and Math (A’STEAM), developed by the Children’s Museum [blinded] and delivered via outreach events for more than 15 years at a sizeable local scale (e.g., 50–100 afterschool sites per year). Using a cluster randomized control trial, researchers and museum educators collaborated to investigate the impact of the A’STEAM curriculum, an outside-of-school time program designed to ignite children’s interest in STEM via hands-on, inquiry activities during the elementary years of kindergarten to Grade 5 (K-5). We were interested in how girls’ and boys’ STEM attitudes, perceived abilities, and achievement were impacted. Children experienced one of three randomly assigned conditions: business-as-usual/control afterschool programming; A’STEAM Basic that used the standard hands-on activities; or A’STEAM Stories that added stories of women and girls doing STEM. Our primary outcomes were children’s STEM attitudes within SEVT (i.e., science/math liking and expectancies focused on STEM career interest). Secondary child outcomes were aspects of science achievement and perceived STEM abilities for girls and boys. Parents whose children took part in A’STEAM were also supported to extend this learning at home; therefore, parent involvement in STEM was another secondary outcome. Our evaluation described implementation outcomes and addressed four primary research questions (RQ). Implementation outcomes are of interest because the Stories condition was a novel treatment that we had not previously studied, so we did not want to assume it would be delivered as intended. We also explored the extent to which effects varied by gender and grade level to understand impacts for boys and girls across the K-2 early elementary grades and 3–5 later elementary grades.

1. To what extent does condition assignment (Control, Basic, or Stories) predict change in children’s attitudes, including STEM value, career expectancies, or aspirations?

2. To what extent does condition assignment predict changes in children’s science achievement, focused on science vocabulary?

3. To what extent does condition assignment predict change in children’s perceptions about how good girls and boys are at STEM?

4. To what extent does condition assignment predict change in parental involvement in STEM?

Hypotheses

We expected the A’STEAM program would provide high-quality STEM experiences to all children. Informed by SEVT (Eccles and Wigfield, 2023), for RQ1, we hypothesized that all children would benefit from both treatment conditions on the primary outcomes of STEM career expectancies and values because of the interesting, positive, and exciting experiences from participating in the hands-on A’STEAM activities (e.g., Babarović, 2022; Jiang et al., 2024). It was possible we would see a slightly larger benefit from the A’STEAM Stories condition for girls’ STEM expectancies and career awareness because of the more explicit approaches of discussing stories of women in STEM fields and careers (Santiago et al., 2019; Sullivan, 2019). For RQ2 addressing science achievement, we hypothesized very small effects for both treatment groups relative to the control group because impacting achievement outcomes requires long-term, system level changes with SEVT. We further expected slightly larger science achievement impacts favoring the A’STEAM Stories condition that exposed children to more scientific vocabulary within STEM read-alouds and stories (e.g., Mol et al., 2009; Neuman et al., 2021). For RQ3, we expected that both STEM treatments would reduce endorsement of gender-specific STEM abilities because of the implicit approaches to create an inclusive environment for all students; however, we expected a slightly larger impact of the more explicit A’STEAM Stories condition on perceived STEM abilities for men and women (Master et al., 2025; Miller et al., 2024). Finally, we expected that for RQ4 both treatment groups could have small benefits on parent STEM involvement via the use of text messaging that encouraged families to extend STEM learning at home and visit their local children’s museum (e.g., Haden et al., 2014; Snell et al., 2020; Zucker et al., 2024). The change model for the intervention that leverages SEVT is outlined in the Supplementary Figure SM1.

Materials and methods

Study design

In an urban locale within the South-Central U.S., we used an experimental design that randomized a total of 36 afterschool sites to one of three conditions: Control (J = 12), A’STEAM Basic (J = 12), or A’STEAM Stories (J = 12). We recruited participants across two cohorts – cohort 1 participated in the 2022–23 school year and cohort 2 took part in the 2023–2024 school year. The study was monitored by our local Institutional Review Board (IRB; study #Blinded-HMC-MS-21-0449).

Participants and procedures

Amongst the 36 sites and two cohorts, we received parent/guardian consent for 820 children. Of these, 158 were excluded due to low site enrollment, ineligibility, or prior participation in cohort 1. This yielded a randomized sample of 665 children across three conditions: Basic (n = 225), Stories (n = 220), and Control (n = 220). After accounting for attrition between pre- and post-test, the final analytic sample consisted of 69 site educators and 541 children (n = 250 in K–2 and n = 291 in grades 3–5). Participant demographics are summarized in Table 1. We attempted to enroll an equal number of boys and girls across Kindergarten (K) to Grade 5, but there was some variability in particular grades. Parents/guardians completed a demographic form that included checkboxes with three choices: male, female, or other. We assume parents/guardians reported sex, although the form was not explicit as to whether to report sex assigned at birth or socially constructed gender roles. Overall, 56% of participants were girls. No parents selected “other”; however, sex was not reported for eight children. English was the predominant home language across all groups (>90%). There was variability in the race/ethnicity of participating children however, race and ethnicity were not reported for 66 children.

TABLE 1

Table 1. Demographic characteristics.

Recruitment and eligibility

To improve recruitment, we used a waitlist control approach so all schools received the A’STEAM curriculum sooner or later. That is, sites randomly assigned to the Control condition continued their afterschool programming as usual until posttest was complete; then, Control sites were offered “waitlist” or delayed opportunities to access the A’STEAM training and materials. To be an eligible site, the school had to: (a) already offer an afterschool program for children in grades K-5, as we needed an existing infrastructure to offer the STEM enrichment one day per week; (b) serve >50% economically disadvantaged children in the school population (e.g., qualify for free/reduced lunch); and (c) have one employee agree to serve as the site educator to provide supervision and instructional support during the activities. These site staff were important to ensure that an adult was present who met the local requirements for child supervision and to provide support to the museum educators, including co-facilitating STEM activities. Children were eligible if they were: (a) enrolled in kindergarten to Grade 5 during the focal school year; (b) had no severe disabilities or an individualized education plan for whom the curriculum would be inappropriate (children with mild disabilities such as ADHD or dyslexia were eligible); and (c) had parents who spoke English or Spanish, as the family supports were only available in those languages.

Retention strategies and attrition

Site educators received $25 for completing the pretest and $25 for posttest surveys. Site educators and site coordinators each received $35 for pretest interviews and $35 for posttest interviews. Most site educators were employed by the after-school provider, but for a small number of sites (J = 4) that did not have sufficient staff, we paid $35 per hour to classroom teachers who stayed after school to deliver the program. Parents received $20 at pretest and again at post-test for completing surveys. In both the cohorts, the 18 participating sites were randomized to one of the three conditions (A’STEAM Basic, A’STEAM Stories, Control). Randomization was conducted by one of the authors of this paper using a random number generator on Excel, with randomization blocked by recruitment order. Across both cohorts, we observed 0% attrition for sites and 17.1% attrition for children. Across all sites, 665 children were eligible, consented, and completed pre-test assessments. At post-test, 114 children had attrited (e.g., absent on testing days, moved to a new school, resulting in the child attrition rate of 17.1%). Differential attrition was assessed by calculating attrition rates for each condition from pre- to post-assessment. A’STEAM Basic had a 16.4% attrition rate (225–188 participants), A’STEAM Stories had a 15.0% attrition rate (220–187), and Control had a 20.0% attrition rate (220–176). This differential attrition rate meets the What Works Clearinghouse Group Design Standards Without Reservations using the conservative attrition standard (What Works Clearinghouse, 2015).

Description of conditions and materials

Control condition

Participants in the business-as-usual condition participated in a waitlist control design, where they received no treatment during the study period but were later invited to a 2-h training at the Children’s Museum (blinded) the following fall, which covered one unit and included corresponding supplies. We observed the after-school activities at Control sites with the Dimensions of Success measure, as detailed below, to describe the nature of the Control condition. Families in the Control sites received family museum passes at mid-year (December).

Treatment conditions

The A’STEAM curriculum included up to 24 sessions of hands-on, after-school, child activities led by an informal STEM educator from the museum with support from the site-based educators. Due to recruitment challenges at some sites, weekly sessions varied from 15 to 24 total. We offered the treatments to boys and girls in two age-based groups at each site – Kindergarten (K) to Grade 2 (K-2) and Grade 3 to 5 (3-5) – to ensure activities were tailored to appropriate learning expectations. For example, this allowed up to 20 Grade K-2 children to take part at 3 p.m. and then another 20 Grade 3–5 children to take part at 4 p.m. The A’STEAM Basic condition required 45–60 min and was delivered 1 day per week at each site. The A’STEAM Stories condition included the same activities and structure but added a 15-min component about girls and women doing STEM (60–75 min total). The informal STEM educators who worked for the Children’s Museum (Blinded) were responsible for leading the delivery of the treatments at the afterschool program sites. Site educators who worked for the site’s afterschool program were present during each session to ensure compliance with local standards and to co-facilitate aspects of the STEM activities with the museum staff. Figure 1 summarizes the differences in the treatment components across conditions and includes a photo of a site in each condition. Sites randomly assigned to A’STEAM Basic or A’STEAM Stories delivered these programs using procedures detailed in the next sections.

FIGURE 1

Figure 1. Photos illustrate educators introducing hands-on science, technology, engineering and math (STEM) activities in both treatment groups and discussing STEM during interactive read-alouds for the Stories condition. Key components of Basic and Stories treatments are summarize beneath photos.

A’STEAM basic

First, site educators assigned to the treatment conditions were invited to participate in a 2-h professional learning experience (PLE) at the museum to learn how to deliver the A’STEAM program. Most PLEs were scheduled on weekends and educators received a $70 stipend to attend. This PLE was a hands-on experience in which the lead museum informal STEM educator modeled the STEM activities with the goal of increasing site educators’ procedural knowledge of the activities, how to manage child behaviors and safety, and to discuss disciplinary knowledge of the STEM concepts. To reinforce and extend these concepts, site educators were also offered an online module that included high-level information about sparking children’s interest in STEM and practical tips for managing high-quality afterschool STEM activities. Examples of these professional learning components are shown in online Supplementary Figures SM2, 3.

Next, the A’STEAM delivery began at sites with varying program start dates due to local calendars or requirements. Up to 24 sessions were offered covering up to six units that had approximately four sessions each: Chemistry, Aerodynamics, Astronomy, Engineering, Magnetism, Earth and Water. The units in the A’STEAM informal learning program followed informal STEM approaches (National Research Council [NRC], 2015) but had a secondary goal of general alignment with formal learning standards (e.g., National Research Council [NRC], 2012) to improve adoption by school-based afterschool providers. The hands-on child activities used these steps: (1) introduce the activity; (2) explicitly teach key vocabulary using picture cards; (3) preview the focus question to guide inquiry; (4) model key activity steps and set behavioral expectations; (5) guide children to complete the activity while providing support and checking for understanding; and (6) close the activity by answering the inquiry question and summarizing key ideas. For example, in a Engineering unit activity called Scribble Bots, the inquiry questions were “What is making? Why is it important?” The activity challenged students to design and construct a robot using a simple circuit, motor, and markers that could make scribbles on paper. The vocabulary words taught included: making, simple circuit, load, power source, and conductor. Students were challenged to make the robot without the educator modeling possible designs so they could experience the engineering design process of testing and redesign. Samples of the educator resources and materials used with students are provided in Supplementary Figures SM3–6.

Concurrent with A’STEAM delivery, the research staff delivered simple family engagement resources consisting of family museum passes and text messages. Around the middle of the program (December), all treatment families received passes to visit the local museum at no cost (value of $85). Families received two text messages per week (56 total) with information about the after-school activities their child completed, key vocabulary they learned, and tips for extending exploration of these concepts with home activities that required no specialized materials. For example, a new unit text message said, “Your child started a new unit called Aerodynamics this week! Your child is going to learn about energy, the law of motion, and what causes weight versus gravity.” Other text messages provided facts and tips designed to increase parent involvement in informal STEM, such as: “Fact: Children spend 20% of their time in school. Parents support learning during the other 80% of time. Encourage your child to visit the library, museum, or parks!” These family text messaging approaches were modeled in our past work and show promise for improving parent and child outcomes (Cabell et al., 2019). More detailed samples of text messages are shown in the top panel of Supplementary Table SM7.

A’STEAM stories

The A’STEAM Stories condition was the same as the A’STEAM Basic condition, but with an added 15-min read-aloud or discussion component focused on girls/women in STEM. We conceptualized this as a light-touch intervention (Lewis, 2019) to improve perceptions of gender-specific STEM abilities for girls as well as boys (e.g., Bartholomew and Santana, 2021; Marcus et al., 2023). Adding stories to STEM activities may also improve learning because research shows that discussing books and narratives builds interest, language, and knowledge (e.g., Golke and Wittwer, 2024; Mol et al., 2009; Wright et al., 2022). These additive treatment activities were usually added after children completed the hands-on STEM activity and before closing the lesson. During a pilot phase, we explored child engagement and interest in a variety of stories of famous women in STEM, as well as fictional girls as characters in commercially available books. This resulted in a set of materials children found interesting and relevant: (a) interactive read-alouds with women doing science or engineering activities related to the unit of study; (b) discussing a famous woman scientist whose key contributions were summarized on a framed handout placed in a double sided acrylic frame and referred to as a “table topper” placed at each small-group table; or (c) discussing a video of a woman working in a STEM career. The list of books read aloud and scientists featured on table toppers are in Supplementary Table SM8; a sample handout is in Supplementary Table SM9. Links to the videos of women in STEM careers are in Supplementary Table SM10.

To learn to deliver the A’STEAM Stories treatment, site educators received a stipend ($70) to attend the same 2-h PLE as other educators but stayed for an additional 20-min breakout session in which the lead museum informal STEM educator explained the rationale and procedures for the additive stories component. To reinforce and extend these concepts of using stories to improve children’s attitudes about STEM, site educators were also offered an online module that is outlined in Supplementary Figure SM11 that explains: (a) how to create inclusive afterschool STEM learning environments; (b) how children form ideas about gender participation in STEM and when they become aware of others’ gender-specific STEM attitudes; and (c) how these ideas might influence girls’ and boys’ interest and participation in STEM activities and careers.

The family components were the same across the two conditions, with one exception – that is, 25% (14) of parent text messages were replaced with facts about women in STEM and videos of women in STEM. For example, one text message said: “Fact: Did you know that when compared to girls, boys are over three times more likely to be interested in STEM majors and careers? Let’s work together to change that (link to Women’s Museum of CA).” Other parent texts encouraged them to discuss videos with the child, such as “Check it out: Watch this video at home with your child about Dr. Bridgette Shannon, a famous African-American chemist! (link to Dr. Bridgette Shannon video).” Additional examples of parent text messages in the Stories condition are shown in the lower panel of Supplementary Table SM7.

Background of A’STEAM educators

Museum educators were staff with expertise in informal education and a background in science, engineering, or related careers, with a minimum education requirement of a bachelor’s degree. Museum educators were two White men, one White woman, one South Asian man, and one South Asian woman.

Implementation outcome measures

Adherence

Educator adherence to the treatment protocols was measured using a structured checklist with 14 items for A’STEAM Basic and 15 items for A’STEAM Stories. Each adherence item was scored as yes (1) or no (0) to indicate completion of the component. Adherence items corresponded to the protocol for A’STEAM activity delivery including: (a) Activity Setup: educator introduced the activity, pre-taught two vocabulary words, set expectations, explained materials and safety procedures, and posed a guiding question (six items); (b) Activity Implementation: educators modeled the activity, allowed independent completion, monitored group engagement (four items); (c) Closing: educator provided a summary or debrief (one item). A’STEAM Stories included one additional item: (d) Story: educator read and discussed that unit’s story-based component There were also items assessing general educator guidelines (three items). All items are listed in Supplementary Table SM12.

Dosage of sessions

Dosage data for up to 24 sessions were collected at the site level, with museum educators documenting the number of sessions delivered each day; however, consistent reporting in RedCap was hindered because museum educators generally left site immediately after sessions and therefore lacked stable internet access or sufficient time to complete the survey at afterschool sites. Therefore, we supplemented these dosage data with a record review of museum educators’ time reporting trackers to fill gaps.

Quality of informal learning

Dimensions of Success (DoS; Shah et al., 2018; Papazian et al., 2013) was used to measure the quality of STEM instruction during afterschool site observations. Across conditions, each site was observed twice and a range of programming was documented for a total of 84 observations (23 Control, 31 Basic, 30 Stories). The DoS measures 12 dimensions related to the features of the learning environment, activity engagement, STEM knowledge and practices, and youth development in STEM; each dimension is given a score ranging from 1 (low) to 4 (high) quality using practices deemed evidence-based within informal learning research. All items are outlined in Supplementary Table SM13. A team of four observers was certified to use the measure after completing the DoS Observer Certification training provided by the developers of the measure. During an in-person observation of A’STEAM, for each cohort, all raters demonstrated interrater reliability of 92.86%–100% (M = 98.21%), using the developer’s criteria of within one-score for each dimension as satisfactory agreement. An intraclass correlation coefficient (ICC) was calculated to assess the consistency of ratings provided by four raters across 14 items. The analysis using a two-way mixed-effects model ICC was ≥0.937, 95% CI [0.854, 0.976], indicating excellent inter-rater reliability.

For sites with no STEM activities occurring, the DoS was not appropriate even with a “Low/1” score because of the absence of a true zero on this measure. Therefore, researchers created a descriptive approach to document the types of adult-led, non-STEM activities. We categorized these activities as: (a) academic (e.g., reading, homework help); (b) physically active (e.g., sports, games), and/or (c) creative activity (e.g., arts, music, dance). Activity categories were not mutually exclusive, such that a single observation could be assigned more than one classification code; descriptive statistics are in Supplementary Table SM13.

Sustainment

We surveyed afterschool site educators who supported the treatment groups about 6 months after the intervention period ended (Nov./Dec. of each cohort) to ask if they were still delivering any A’STEAM components and how feasible it was to integrate A’STEAM into their typical afterschool offerings. Due to the end of the grant period, the survey was sent at the same time to waitlist control sites, which was only about 2 months after the BAU training (i.e., delivered in the Sept of each cohort and after post-test). The response rate was 30 site educators completing the survey; however, site educators’ sustainability survey data were excluded if the educator changed sites between enrollment and sustainability phases. This resulted in 24 site educator sustainability surveys collected across all 19 sites (BAU n = 5; Basic n = 8; Stories n = 6). For ease of interpretation, sites with multiple sustainment survey reports are reduced to a single report by selecting the data that was most complete, as each site had some missing data. More details are provided in Supplementary Table SM14.

Measures

All child and parent outcome measures were administered at pre-test and post-test. The primary child outcomes aimed to increase positive attitudes within SEVT (Eccles and Wigfield, 2023) because these are proximal or closely tied to the A’STEAM intervention targets. Secondary outcomes varied, such as a distal measure of science achievement where we expect longer-term, downstream effects in the SEVT that could take years to emerge. Other secondary outcomes like the parent involvement in STEM measure were closely related to the intervention, but considered a secondary target because of the light-touch nature of the family components.

Child measures

STEM value

A set of 12 items measured children’s STEM subjective task value and focused on the intrinsic value or enjoyment children perceived in doing science and related tasks. These items were adapted from several existing measures that measured science liking, interest, and similar constructs (Patrick et al., 2009; Weible and Zimmerman, 2016). Sample items include: “I like science,” “I want to know more about science,” “I like to use different tools,” “I would like to invent something new,” “I take things apart to see how they work.” Children responded to questions using a five-point scale, which included words and smiley faces to ensure all age groups could respond regardless of literacy skills (1 = Strongly Disagree, 3 = Neutral, 5 = Strongly Agree). An exploratory factor analysis demonstrated that these items loaded onto a single factor at both pre- and post-test, with factor loadings ranging from 0.503 to 0.704 at pre-test and from 0.433 to 0.750 at post-test. We used a factor score derived from children’s post-test responses, instead of an average score, as the dependent variable to better reflect how items differently related to the construct of STEM value (see Supplementary Figure 15). Internal consistency ranged from α = 0.84–0.86.

STEM career expectancies

We used two approaches to measure children’s STEM career expectancies or beliefs about whether they would want a science or engineering career. Two items for STEM career expectancies asked children to rate how interested they would be in becoming a scientist and engineer (“How interested would you be in being a scientist?”/ “How interested would you be in being an engineer?”), again using the five-point scale with smiley faces (1 = Strongly Disagree, 3 = Neutral, 5 = Strongly Agree). Correlations among items for science and engineering expectancies were moderate and ranged from r = 0.33 to 0.34 (see Table 2).

TABLE 2

Table 2. Correlations among measures at pre-test (below diagonal) and post-test (above diagonal).

STEM career aspirations

To further assess career expectancies, we also used a researcher-developed approach. We first used an unscored, priming question asking them to think of all the jobs they know. Next, we asked them, “What do you want to be when you grow up?” and coded responses as one of three career fields adapted from the 2019 National Survey of College Graduates (National Center for Science and Engineering Statistics [NCSES], 2021): (a) STEM careers (e.g., astronaut, engineer, marine biologist, zookeeper); (b) STEM-related careers (e.g., doctor, nurse, vet, electrician, architect); or (c) non-STEM career (e.g., fireman, gym teacher, sports player, movie star, cook, judge, teacher general). A senior research staff member worked with the lead author to develop methods for categorizing children’s sometimes unclear responses (e.g., “video game maker”), following the principle of assigning the most advanced STEM careers that could be inferred from that response (e.g., coded as STEM software developer) rather than STEM-related occupations (e.g., computer programmer). This was an iterative process across six coding team trainings/meetings, with additional cycles of review and consensus-based revisions. This process resulted in an average agreement of 98% across a team of five coders. Supplementary Table SM16 provides more details on coding.

Science achievement

We administered the Woodcock-Johnson IV Tests of Achievement Science subtest (Schrank et al., 2014) to measure children’s science knowledge. This measure requires naming a set of pictures and answering several questions to test their vocabulary knowledge. This subtest has strong psychometric properties with internal reliability of 0.84 and concurrent validity of 0.89 with the Wechsler Individual Achievement Test-II.

STEM ability perceptions

We asked students how good they think most girls/boys are at science and math (Master et al., 2017). Again, a five-point scale with smiley faces was used (1 = not very good, 2 = not good, 3 = good, 4 = very good, 5 = super good). Correlations among these items ranged from r = 0.44 to 0.46 for perceptions of girls’ abilities and r = 0.42 to 0.50 for perceptions of boys’ abilities (see Table 2).

Parent measures

A 10-min parent survey captured parent involvement in STEM activities. The parent involvement scale, which was adapted from its original scale (West et al., 2007) and used in our prior museum studies (Zucker et al., 2024), used a five-point scale (“rarely or never” to “most days per week”) to measure how frequently parents engaged in math- and science-related talk and activities with their child. Sample items include “I encourage my child to do math in their head” and “In my interactions with my child, I try to connect science to everyday life.” Internal consistency for parent involvement in STEM talk ranged from 0.89 to 0.92. Three other items used a four-point scale (“not at all” to “more than once a week”) to rate the frequency of visiting informal STEM experiences in the past 6 months, such as the museum, zoo, or science events for children. Due to limited resources for reminding and incentivizing parents to complete surveys, we surveyed a random selection of about one-third of the focal sample (276 total, n = 137 in Cohort 1; n = 139 in Cohort 2). Note that the survey response rate was acceptable (258 total, Cohort 1, pretest n = 79, 58%; posttest n = 87, 64%; Cohort 2 pretest n = 87, 63%; posttest n = 83, 60%); however, there was substantial missing data for these measures with only 124 parents in the analysis sample for both pre- and post-surveys. As expected given the brevity of the scale, reliability was moderate for visits to informal STEM experiences (α = 0.55–0.57).

Analysis approach

Given the nested structure of children within school sites, we employed multilevel modeling to examine treatment effects while accounting for the hierarchical data structure and potential clustering effects. The models included child-level covariates at Level 1 (gender, parent education level, race, ethnicity, grade group [K-2, 3–5], study cohort, and pretest scores to calculate residualized gains) and the site-level treatment condition predictor at Level 2. We tested condition × child gender interactions to examine potential differential treatment effects for girls versus boys. Although we accounted for site-level variation in the multilevel structure, our models focused on child-level predictors and treatment condition. Main analyses used intent-to-treat (ITT), estimating effects of treatment assignment. To control for false discovery rate, we applied the Benjamini-Hochberg (BH) procedure when evaluating post hoc pairwise comparisons among the conditions (Basic vs. Control, Stories vs. Control, and Basic vs. Stories) for each outcome (Benjamini and Hochberg, 1995). Additionally, for outcomes with significant main effects of categorical predictors (e.g., gender, grade group), we conducted exploratory pairwise comparisons across levels of those predictors. For interpreting interaction effects, we present Δd (the difference between subgroup standardized mean differences), as Cohen’s d has been recommended for effect size assessment in regression-based longitudinal and multilevel models (Feingold, 2013).

Results

We first present descriptive findings for implementation and child outcomes and then results for each research question.

Implementation outcomes

Starting with fidelity of implementing A’STEAM as designed, overall adherence was good in both conditions, with similar mean scores for Basic (78%, M = 7.84 out of 14, SD = 2.80) and Stories (85%, M = 8.48 out of 15, SD = 2.57). However, item-level data (see Supplementary Table SM12) related only to the Stories components showed these read-alouds or video discussions were only implemented in 73% of observations. Museum educators indicated they often ran out of time because some or most children were picked up during this final portion of the afterschool program.

Turning to the dosage or amount of 24 A’STEAM sessions delivered during the intervention period, due to cancelations by the site (e.g., site educator shortages/sick, conflicting events/field trips at afterschool site) or delays in recruitment, the site average was approximately 20 sessions (SD = 2.83; range = 15–24), which was similar across the Basic (M = 19.93, SD = 1.98) and Stories (M = 19.47, SD = 3.78) conditions. Across conditions, 58% of sites received a high dosage (i.e., defined as 20+ sessions or >80%; Basic = 67%, Stories = 50%); whereas 42% were in the moderate range (13–19 sessions; Basic = 33%, Stories = 50%) and none were low [i.e., <12 sessions, using fidelity definitions from Hill and Erickson (2019)]. There was no significant difference in dosage between Basic and Stories conditions, F(1,22) = 0.18, p = 0.675.

Regarding the site-level measure on quality of the informal learning provided, DoS scores were similar between treatment conditions when averaged across DoS measurement criteria (A’STEAM Basic: M = 3.40, SD = 0.30; A’STEAM Stories: M = 3.44, SD = 0.29). These scores were in the moderate-high (four-point rating), suggesting A’STEAM resulted in approaches aligned with evidence-based informal learning practices. Item-level DoS details are in Supplementary Table SM13. Noteworthy items are participation ratings for girls and boys with a higher quality score for girls’ engagement in the Basic condition (f = 4.677; p = 0.0389); participation was not significantly different for girls versus boys in the Stories condition. Importantly, these quality observations revealed a paucity of STEM activities in 96% of business-as-usual sites. That is, across 23 Control observations, only one instance involved a STEM-related activity (i.e., use of a technology, which received a DoS score of 2.93). The remaining 22 observations were categorized as follows. Most control sites (55%) offered activities in multiple categories, with 14% of sites offering both academic and physical activities, 18% offering academic and creative activities, and 14% offering creative and physical activities. For the other control sites (45%) with only one non-STEM activity observed, these included: 9% academic, 18% physical activity, and 18% creative activity.

In terms of site educators’ sustainment of the intervention delivery, at 83% of Stories sites and 75% of Basic sites, educators reported continued use of the program. Perhaps due to recency, at 100% of waitlist control sites, they reported initiating the A’STEAM program. When asked about integrating this program into their typical workflow, site educators indicated that this fit moderately well into their routines. All conditions had some sites that reported limited resources or materials for the STEM activities were a barrier to continuing to deliver the program (40% BAU, 50% Basic, 33% Stories). But the other reported barriers to sustainment differed by condition with lack of educator training as a more prominent issue at 40% of BAU sites that did not have up to 24 sessions co-faciliated with the museum educator (cf. 13% Basic, 0% Stories). At 25% of Basic sites and 33% of Stories sites a barrier was shortage of personnel, whereas this was reported as a barrier at 0% of BAU sites.

Descriptives: child-level outcomes

To describe the child-level outcome measures, Table 2 presents correlations among child and parent outcome measures at pretest and posttest, pooled across conditions. A stable cluster emerged among the STEM motivation measures [i.e., attitudes about STEM value and career expectancies (scientist, engineer)], with consistently moderate correlations across time points (r = 0.28–0.52, p < 0.001). STEM career aspirations were modestly associated with these motivation variables (r = 0.11–0.21), indicating related but distinct constructs. Notably, STEM achievement (WJ Science Achievement) was positively associated with career aspirations at both pre-test (r = 0.23, p < 0.01) and post-test (r = 0.26, p < 0.001), suggesting a consistent link between children’s science knowledge and interest in STEM careers, particularly in being an engineer. Correlations among perceived STEM ability indicators (perceived science and math ability of girls and boys) were moderate to strong (r = 0.42–0.50 for ratings about boys; r = 0.44–0.46 for ratings about girls). These variables were only weakly associated with STEM motivation or career aspirations. Notably, perceived girls’ math ability was negatively correlated with STEM achievement at post-test (r = −0.19, p < 0.01), a pattern not observed at pre-test. Lastly, the informal STEM engagement (parent STEM talk and activities and visits to informal STEM experiences) were moderately correlated, particularly at post-test (r = 0.52, p < 0.001), but showed limited correlations to other domains. Descriptive statistics and results of the experimental conditions’ impacts on the STEM outcomes are presented in Table 3.

TABLE 3

Table 3. Pre-test and post-test performance on science, technology, engineering and math (STEM) outcomes with multi-level modeling statistics.

RQ1: Which A’STEAM conditions impacted children’s STEM attitudes?

STEM value

Aligned with hypotheses, A’STEAM showed positive effects on STEM value compared to the control condition (see Table 3). The Basic A’STEAM condition produced a significant medium effect (d = 0.25, p = 0.028), with participants moving from a pre-test mean of 0.12 to post-test mean of 0.14, while control participants declined from −0.07 to −0.17. However, this contrast did not remain statistically significant after applying the Benjamini-Hochberg correction. The Stories condition demonstrated a small effect that was not statistically significant (d = 0.18, p = 0.066), with participants maintaining positive STEM value (pre-M = 0.00, post-M = 0.03) compared to the control decline. Results from the ITT model indicate that no covariates or interactions with gender were significant in the model other than children’s STEM value at pretest (B = 0.41; p < 0.0001; see Table 4).

TABLE 4

Table 4. Estimates (standard errors) for science, technology, engineering and math (STEM) outcomes from multi-level models.

STEM career expectancies

To measure career STEM expectancies, children were asked to rate their interest in becoming a scientist or an engineer when they grow up. Contrary to hypotheses, results from the multilevel regression analyses in Table 4 indicate that treatment condition was not significantly related to these outcomes, and the only significant covariates were children’s career expectancies at pretest for both scientist (B = 0.40, p < 0.0001) and engineer (B = 0.33, p < 0.0001), respectively.

STEM career aspirations

The multilevel regression for career aspirations found no significant effect for condition or gender. Although we did not observe the hypothesized main effect for condition, there was a significant condition by gender interaction, such that girls in the Stories condition experienced significantly greater gains in the number of STEM and STEM-related careers they reported compared to boys in the Stories condition (B_boys = −0.31, p < 0.05; see Table 4). Follow-up contrasts indicated that boys in the Stories condition named fewer careers than boys in Control (d = −0.35, p = 0.041) and Basic (d = −0.28, p = 0.040), whereas the corresponding contrasts for girls were small and non-significant (Stories vs. Control: d = 0.14; Stories vs. Basic: d = 0.08). The difference between boys and girls in the Stories condition (Δd = –0.49 vs. Control; Δd = −0.36 vs. Basic) represents a moderate interaction effect. These contrasts did not remain significant after Benjamini–Hochberg correction, though the pattern of effects and associated effect sizes suggest a consistent moderation by gender. As expected, the number of STEM and STEM-related careers children named at pretest (B = 0.45; p < 0.0001) was a significant predictor of the number of these careers children named at post-test, as well as grade group, with children in grades 3–5 naming a higher number of STEM and STEM-related careers compared to children in grades K-2 (B = 0.17, p < 0.05).

RQ2: Which A’STEAM conditions impacted children’s science achievement?

While we only hypothesized potentially small effects on this distal measure, there were no significant impacts. The model for children’s STEM achievement showed that pretest (B = 0.80, p < 0.0001) was the only significant predictor of post-test achievement besides race (B = 3.35; p < 0.05), with children identified as White demonstrating significantly higher gains compared to children from other racial groups, after controlling for pretest scores, gender, grade level, cohort, parental education, and Hispanic ethnicity. Importantly, because the model included Hispanic ethnicity as a separate covariate, the effect of White racial identity reflects the contrast between White and non-White students regardless of Hispanic status. This distinction allows us to conclude that White children on average demonstrated stronger posttest STEM performance than children who were non-White independent of baseline achievement and other covariates, a finding that is consistent with prior research using WJ assessments (e.g., Edwards and Oakland, 2006; Morgan et al., 2016).

RQ3: Which A’STEAM conditions impacted children’s perceived STEM abilities?

Four multilevel regression models were estimated for perceived girls’ and boys’ ability in science and math, respectively. The science and math items were modeled separately because of their moderate correlations (see Table 1). Children’s ratings for both of these items exhibited ingroup bias, with boys rating girls lower than boys in both science (B = −0.45, p < 0.01) and math (B = −0.63, p < 0.0001), and rating themselves higher in both domains (B_Science = 0.48, B_Math = 0.72, ps < 0.007). However, we observed a significant gender by treatment interaction for boys’ ratings of girls’ math ability, with boys in the Basic condition rating girls’ ability higher compared to boys in the Control condition (B = 0.48, p < 0.05; d = 0.21 vs. −0.25, Δd = 0.46), indicating a moderate positive shift in perceptions among boys exposed to the Basic program. This interaction did not remain significant after Benjamini–Hochberg correction. In addition, post hoc pairwise comparisons indicated a significant positive effect of the Stories condition (d = 0.20, p = 0.041) on perceived girls’ science ability, though this contrast also did not remain significant after correction. Covariates beyond pretest in Table 4 show that children in grades 3–5 had higher perceptions of boys’ science abilities relative to the lower grades. Parental higher education (B = 0.19, p = 0.045) was also a significant predictor of perceived girls’ math ability, indicating that children with more highly educated parents perceived girls as more capable in math.

RQ4: Which A’STEAM conditions impacted parent STEM involvement?

Parent-reported STEM involvement was measured using items capturing participation in STEM-related community activities (e.g., zoos, museums, or science events) and frequency of STEM talk at home. Assignment to the Stories condition was associated with greater gains in parent-reported STEM talk for parents of both girls and boys (B = 0.48, p = 0.036), whereas parents in the Basic condition did not differ significantly from Control (B = 0.29, p = 0.21). Several additional variables significantly predicted STEM talk at home. Boys’ parents reported higher gains in STEM talk at home than girls’ parents (B = 0.56, p = 0.018), and parents of White children reported lower gains in STEM talk compared to parents of non-White children (B = −0.27, p = 0.044), a finding that stands in contrast to the STEM achievement results. Parents of non-Hispanic children did not differ significantly from parents of Hispanic children (of any race; B = −0.13, p > 0.05). Importantly, a significant interaction between condition and gender emerged, indicating that the gender difference (greater gains in STEM talk for boys) was significantly reduced in the Basic (B = −0.68, p = 0.033) and Stories conditions (B = −0.60, p = 0.052), albeit marginally significantly, compared to Control. This suggests that parents of girls in the Basic or Stories condition showed greater increases in math talk than those in the Control condition. Follow-up contrasts indicated a gender-specific effect: parents of girls showed increases in the Basic condition relative to Control (d = 0.39), whereas parents of boys showed decreases (d = −0.48), yielding a large subgroup difference (Δd = 0.87). Similarly, parents of girls in the Stories condition reported significantly greater increases (d = 0.63, p = 0.036) than parents of boys (d = −0.14, p = 0.579), yielding a moderate-to-large subgroup difference (Δd = 0.77). However, the interaction did not remain significant after the correction. Pre-test scores were significant predictors of STEM talk (B = 0.78, p < 0.0001) and the only significant predictor for the measure of parent involvement in STEM-related community activities (B = 0.42, p < 0.0001).

Discussion

Given the importance of early STEM experiences to long-term STEM pathways (e.g., Kaderavek et al., 2020; Pellegrini et al., 2021), this study evaluated the impact of a STEM-focused afterschool program on girls’ and boys’ STEM outcomes while experimentally manipulating the treatments to include hands-on science alone or in combination with stories that feature women or girls doing STEM. The most important findings were that the afterschool STEM programs produced small effects on all children’s value of STEM and moderate effects on girls’ STEM career aspirations, albeit non-significant after corrections for multiple contrasts. Some parents of children in the treatment conditions also showed significant increases in their STEM talk. In the next sections, we discuss how these findings relate to prior research and what implications these results have for informal STEM programs.

Impacts on child outcomes

Motivation: STEM value and career expectancies

Using Situated Expectancy-Value Theory (SEVT; Eccles and Wigfield, 2020), our primary outcomes examined two dimensions of elementary children’s motivation to do science and engineering: value and career expectancies. Both the A’STEAM Basic and Stories approaches showed promising trends for improving children’s value of STEM, with potentially meaningful effect sizes (d = 0.25 and d = 0.18, respectively); however these effects were not significant. These findings underscore the potential for informal, hands-on STEM experiences to positively shape children’s motivational beliefs in under-resourced communities. This aligns with other STEM research that shows, when children and adults work together to solve problems or understand inquiry processes, it can be enjoyable in ways that increase interest, curiosity, or desire to participate in STEM learning (Fielding-Wells et al., 2017).

We used two measures to assess treatment impact on STEM career expectancies: ratings of desire to become a scientist or engineer and an open-ended career aspirations question. Although STEM career expectancy ratings were not significantly influenced by either intervention, the findings indicate that children’s pretest attitudes were strong predictors of posttest outcomes, highlighting the stability of early motivational beliefs. The null treatment effects suggest that to achieve deeper and more lasting change, interventions may need to be extended to expose children to more snapshots of STEM careers or be more focused on helping children identify STEM career pathways they find more appealing (Le and Robbins, 2016; Rosenzweig and Chen, 2023). It is worth noting that other studies that use SEVT to theorize about ways to improve the STEM pipeline show that value is a stronger predictor of children’s STEM motivation than success expectancies, which is likely due to the reality that career expectancies can shift slowly or unevenly over time (Ball et al., 2017; Jiang et al., 2024).

Interestingly and in contrast to the null effects on STEM career expectancy ratings, the Stories condition produced a gender-specific change in career aspirations. That is, we found a trend that girls in this condition reported greater gains in the number of STEM careers they could envision pursuing compared to boys in the same condition; however, these effects were non-significant after corrections. This suggests that when educators expose children to narratives about women in STEM careers, it may be useful in expanding girls’ awareness of STEM career possibilities, which aligns with relevant research (e.g., Kim et al., 2018; Riegle-Crumb and Peng, 2021; Sansone, 2019).

Additionally, the significant effect of grade level on the career aspirations, which indicates that K-2 children had smaller gains on this outcome, suggests that young children’s career aspirations may be less developed, require more focused exploration to support development, and/or primarily reflect their immediate social contexts, such as parents and teachers. Brief exposures to inspiring stories of women who are scientists and engineers may not be sufficient for younger children to see themselves in future STEM careers (Cimpian et al., 2020; Wang and Degol, 2013). Thus, to effectively shift early-elementary grade children’s perceptions about their roles in STEM fields, it may be more impactful to explicitly provide more exploration of STEM careers and/or involve caregivers and teachers in STEM-related activities (e.g., Acosta and Haden, 2023; Marcus et al., 2023). Experiences like A’STEAM that promote positive STEM experiences are likely important but not sufficient for selecting STEM careers in the absence of other key factors like positive social persuasion or understanding the personal utility of these career pathways (Zhou and Shirazi, 2025).

Science achievement

Given that science achievement gaps emerge early and persist over time (e.g., Morgan et al., 2016), we measured potential changes in science achievement for these children, most of whom were experiencing poverty. We found no significant treatment effects on the standardized measure of science achievement. This is not surprising given that measures of achievement are difficult to shift in the short-term because achievement builds cumulatively over time and there is limited potential for change after only 24 afterschool STEM sessions (Duncan et al., 2007; Kaderavek et al., 2020). Another mechanism to consider is difficulties in learning transfer. Informal STEM learning often does not generalize to more formal assessments unless connections to formal academic material are made explicit through avenues such as narrative reflection (Marcus et al., 2021). We also found racial disparities in impacts such that White children benefited more than non-White children, which is consistent with other reports of systemic disparities (e.g., Edwards and Oakland, 2006; Heberle and Carter, 2015). These disparities might result from factors such as differential exposure to school-aligned STEM practices or familiarity with the structure of standardized science assessments (see White et al., 2016).

STEM ability perceptions and gender

This study also sheds light on how informal STEM programming may affect children’s gendered perceptions of ability. Overall, children in grades 3–5 showed greater increases in their rating of boys as capable in science compared to children in grades K-2, whereas changes in the ratings of girls’ abilities were not significantly different between grades. This may indicate that gendered ability beliefs may intensify with age, as shown in the extant literature (e.g., Martinot and Désert, 2007; Miller et al., 2024). However, the finding that boys in the Basic condition made significantly greater increases in their ratings of girls’ math abilities than boys in the Control group also shows that identity-neutral, hands-on STEM experiences may help counteract early gender stereotypes and modestly disrupt these patterns, especially among boys (c.f., Kim et al., 2018; Steffens et al., 2010; Tang et al., 2024). It is particularly interesting that the Basic condition, and not the Stories condition, had this significant effect, even when it did not include explicit gender messaging.

In contrast, no improvements were observed in girls’ perceptions of girls’ abilities or in how they viewed boys’ abilities. This asymmetry suggests that implicit approaches may affect peer perceptions without meaningfully altering ingroup-perceptions, particularly among girls. Accumulated research shows that these perceptions are resistant to change without intentional efforts (Miller et al., 2024). Future approaches may need to more directly address these topics, perhaps by inviting conversation about what children believe girls and women can achieve in STEM fields (e.g., Master et al., 2021; Sullivan, 2019). While the Stories condition sought to showcase women’s STEM achievements, it may not have been developmentally aligned or explicit enough to fully disrupt early gender stereotypes. Reading the stories at the end of the session, rather than the beginning, may have also reduced children’s engagement with the narratives. It is also possible that educators in the Stories condition may have used language about girls and women than inadvertently reinforced gendered beliefs (Chestnut and Markman, 2018; Wang et al., 2025).

Parents’ education levels were also a significant predictor of children’s perceptions: children with more highly educated parents were more likely to perceive girls as capable in math. This highlights how family background may shape children’s early gender beliefs about STEM. This finding aligns with prior research that parent education is one socialization factor that influences children’s self-perceptions (Guo et al., 2024; Martinot and Désert, 2007). Although we did not identify whether the reported education level referred to a mother or a father, it is possible that children with more highly educated parents are more frequently exposed to women in STEM fields, which may influence their beliefs about girls’ abilities.

Impact on parent involvement

We provided treatment families with text messages and family museum passes to extend the A’STEAM concepts and further bolster STEM interest pathways (e.g., Dabney et al., 2016; Pattison and Ramos Montañez, 2022). In contrast to parents in the Basic condition, parents in the Stories condition, who received text messages that included facts and videos about women in STEM in 25% of the messages, reported significantly greater gains in STEM-related conversations at home. This finding may suggest that facts about women in STEM resonated more with some parents, potentially prompting more active STEM dialogue with children. There were also significant interaction effects between treatment and gender, suggesting that gender-based differences in STEM talk at home were reduced for parents of girls in either treatment condition. Additionally, White parents reported significantly less STEM talk at home than non-White parents, and parents of boys indicated significantly greater increases in their STEM talk at home than parents of girls. These findings may reflect cultural and gender-based differences in how STEM topics are discussed at home or how program messages were interpreted across households (e.g., Callanan et al., 2017).

Taken together, these promising findings align with other emerging approaches that rely on text messaging and provision of resources to involve parents in informal learning (e.g., Ennes et al., 2023; Hurwitz et al., 2015; Snell et al., 2020; Zucker et al., 2022, 2024). The findings highlight the potential power of informal and light-touch family STEM engagement approaches as an opportunity to enhance informal learning via family involvement. Findings also underscore the potential to refine such family engagement approaches to account for cultural and gender-based variation in response.

Implications for quality STEM experiences in U.S. afterschool programs

Studies with nationally representative U.S. datasets show that frequent, high-quality STEM learning experiences are essential for promoting achievement from elementary through high school (Bustamante et al., 2023; Curran and Kitchin, 2019). Our descriptive findings from observing business-as-usual afterschool programs revealed a near-total absence of STEM-related experiences. This aligns with prior findings that many children living in poverty have insufficient access to high-quality STEM experiences (e.g., Morgan et al., 2016; Papazian et al., 2013). These observations confirm that implementing the A’STEAM program fundamentally altered the nature of the afterschool environment, a key factor for understanding how this causal impact study led to promising improvements or interactions for three of the five measured child outcomes: STEM value, career aspirations, and perceived math abilities. Our finding that A’STEAM improved students’ perceived math abilities is especially promising because it shows that hands-on STEM based programs have the potential to increase students’ perceptions of others’ ability to do STEM, which may counteract field-specific ability beliefs (see Jenifer et al., 2024) that one has to be particularly brilliant to succeed in math and other STEM fields.

The scarcity of STEM-related afterschool programming for control sites is especially notable given that this study was conducted at sites serving a majority of children experiencing poverty, during a period of increased funding for afterschool academic enrichment (Levine, 2024; National Center for Education Statistics [NCES], 2024). Systematic efforts are needed to improve the quality of out-of-school STEM experiences (National Research Council [NRC], 2015). Policymakers should consider ways to expand and sustain informal STEM learning opportunities through diverse local and federal channels (e.g., Allen et al., 2019; Pinkard et al., 2025). Afterschool providers could consider implementing hands-on science and engineering programs like A’STEAM (see freely available A’STEAM resources at University of Texas Health Science Center at Houston, 2024) once a week to meaningfully support children’s STEM development.

This study offers valuable guidance for designing inclusive, informal STEM experiences in under-resourced communities. Both the A’STEAM Basic and Stories-enhanced versions showed promise in improving children’s value of STEM. Informal educators should align their program goals with the needs of their students: explicitly presenting identity-affirming stories of women and girls in STEM supported girls’ career aspirations, while the identity-neutral, implicitly inclusive Basic approach was associated with boys rating girls’ math abilities more highly. These findings are consistent with evidence that affirming girls’ identities in STEM during elementary years helps to establish longer-term engagement and pathways (e.g., Habig et al., 2020; Pattison and Ramos Montañez, 2022).

There are also practical considerations. While both conditions were implemented with adequate adherence to the A’STEAM protocols, the Basic approach was slightly more feasible than the Stories-enhanced version that with this component only delivered in 73% of observations, due to limited time or attendance during the added 15 min required for discussing a book, video, or handout featuring women in STEM. Indeed, our follow-up surveys of afterschool educators confirm that adding STEM stories was not always a feasible addition to their workflow due to the extra time needed. However, we do not know if this was specific to read-alouds of commercial books, discussion of table toppers, or discussion of videos. Although read-alouds are a widely used activity with elementary students (Trelease, 2013), the usability of video learning material is promising but less clear (e.g., Bulca et al., 2022; Lu et al., 2022), particularly in afterschool sites that may not have video projection equipment available. Nevertheless, given growing evidence that shared story-based discussions can significantly support children’s knowledge development (e.g., Acosta and Haden, 2023; Cabell and Hwang, 2020; Wright et al., 2022), adding a Stories component is promising when time and resources allow. Further research is needed on how much professional development is required for afterschool educators to successfully lead programs like A’STEAM without co-facilitation with the museum-based informal STEM educators, as integrating story-based approaches can present major challenges for educators (e.g., Barchas-Lichtenstein et al., 2023).

Limitations and future directions

It is important to recall that this research represents a self-selected group that takes part in afterschool STEM programs; thus, findings may not apply to all children in elementary grades or enrolled in afterschool programs. This study was also somewhat impacted by the COVID pandemic that delayed recruitment in cohort 1 due to ongoing challenges in afterschool partner sites. Also, caution should be taken in interpreting findings because several effects that were initially significant did not remain significant after applying the Benjamini-Hochberg correction for multiple comparisons.

There are several limitations to this causal impact study. First, although participants at all sites were given free passes to visit the museum, the museum team was unable to track if participants used these passes to visit the museum; thus, we do not know if families used this treatment component. Instead, we only measured parent self-report of typical frequency of trips to museums and other STEM-related experiences. Second, dosage data relied on site-level reporting by museum educators, which was sometimes hindered by internet connectivity and limited time immediately after sessions. Since dosage data were gathered from retrospective record reviews, rather than captured immediately after sessions, this blocked us from completing treatment-on-treated analyses. Third, child engagement during program sessions was also not captured at a granular level, such that we did not measure student engagement at every session. Instead, researchers measured engagement during two implementation quality observations using the DoS observation framework, providing a general measure of engagement (see Table SM14 of the appendix, “Engagement with STEM”). However, these data do not allow us to test the extent to which each participant’s engagement in the program influenced outcome measures. Future research should systematically capture child-level engagement to more fully test its impact on child outcomes.

Next, the child measures in this study may not have been sufficiently aligned with the treatments to detect all impacts in all age groups. For example, we modified some existing measures for use with students younger or older than in previous research (e.g., Patrick et al., 2009; Weible and Zimmerman, 2016). On the researcher-developed child career aspiration measure, we coded their responses and then considered various ways to score the responses. We decided to calculate a total score of STEM-related careers, rather than the proportion of STEM-related careers they named when they had multiple possible career goals, which may have created bias in favor of students with larger vocabularies. Also, although we labeled the items about pursuing STEM careers (e.g., “I want to be a scientist/engineer”) as STEM career expectancies, they may be more reflective of career interest than expectancies of success, as defined by SEVT, because they do not indicate expectations of success. We assumed that children would express interest in careers they believed they would succeed in, but this assumption may not have been held by participants. Finally, we did not survey site educators about implementation outcomes at posttest, but only after several months of delivering on their own without the museum educators’ support; therefore, we have limited understanding of the immediate barriers to delivery.

Finally, although the parent survey rate was acceptable within the subsamples, it was only sent to a random selection of around 50% to 60% of parents due to limited resources for this type of data collection. Thus, these findings may underestimate or overestimate parent effects given that only a subgroup was selected and that these data may represent parents who were more likely to respond to the parent involvement survey. Also, these measures were gathered only at immediate post-test, rather than after a longer-term, follow up period. Building on previous research indicating that peers can influence students’ motivation and interest (Hazari et al., 2017; Raabe et al., 2019), future work should also investigate whether peer characteristics (e.g., STEM interest, knowledge) influence students’ own development.

Conclusion

These findings underscore the potential for afterschool, hands-on experiences like A’STEAM to positively influence children’s STEM-related attitudes. However, findings also suggest that afterschool informal learning approaches need to be more intensive, sustained or tailored to achieve more substantial impact. Educators and policymakers should seek funding for museum partnerships and other resources that allow afterschool providers in elementary schools to offer hands-on STEM activities and stories of how STEM is relevant to children across the lifespan.

Data availability statement

The raw data supporting the conclusions of this article are available at the Center for Open Science in data repository #YWJNH, available here: https://osf.io/ywjnh/.

Ethics statement

The studies involving humans were approved by the Committee for the Protection of Human Subjects (CPHS) at UTHealth Houston: HSC-MS–21-0449. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin. Written informed consent was obtained from the individual(s), and minor(s)’ legal guardian/next of kin, for the publication of any potentially identifiable images or data included in this article.

Author contributions

TZ: Conceptualization, Funding acquisition, Investigation, Resources, Supervision, Writing – original draft, Writing – review & editing, Methodology. MM: Conceptualization, Formal analysis, Supervision, Visualization, Writing – original draft, Writing – review & editing. VB: Data curation, Investigation, Resources, Validation, Writing – original draft, Writing – review & editing. DD: Data curation, Investigation, Supervision, Writing – review & editing. YA: Data curation, Formal analysis, Investigation, Validation, Writing – original draft, Writing – review & editing. AM: Conceptualization, Writing – original draft, Writing – review & editing. JH: Project administration, Resources, Supervision, Writing – review & editing. CM: Conceptualization, Funding acquisition, Project administration, Resources, Writing – review & editing.

Funding

The author(s) declare financial support was received for the research and/or publication of this article. This work was supported by National Science Foundation (NSF) Advancing Informal STEM Learning (AISL) program under Award Number 2115579 and the Harriet and endowments to UTHealth from the Joe Foster Foundation and the Albert and Margaret Alkek Foundation.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The authors declare that Generative AI was used in the creation of this manuscript. To check references were accurately used and to suggest where the original writing (by human authors) could be clarified.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/feduc.2025.1679669/full#supplementary-material

References

Acosta, D. I., and Haden, C. A. (2023). Supporting Latine children’s informal engineering learning through tinkering and oral storytelling. Dev. Psychol. 59, 2342–2355. doi: 10.1037/dev0001648

PubMed Abstract | Crossref Full Text | Google Scholar

Adler, J. M. (2012). Living into the story: Agency and coherence in a longitudinal study of narrative identity development and mental health over the course of psychotherapy. J. Personal. Soc. Psychol. 102:367. doi: 10.1037/a0025289

PubMed Abstract | Crossref Full Text | Google Scholar

Afterschool Alliance (2024). Evaluating afterschool: The latest research on the impact of afterschool and summer programs. Washington, DC: Afterschool Alliance.

Google Scholar

Allen, P. J., Chang, R., Gorrall, B. K., Waggenspack, L., Fukuda, E., Little, T. D., et al. (2019). From quality to outcomes: A national study of afterschool STEM programming. Int. J. STEM Educ. 6:37. doi: 10.1186/s40594-019-0180-9

Crossref Full Text | Google Scholar

Anderson, D., Piscitelli, B., Weier, K., Everett, M., and Tayler, C. (2002). Children’s museum experiences: Identifying powerful mediators of learning. Curator: Museum J. 45, 213–231. doi: 10.1111/j.2151-6952.2002.tb00057.x

Crossref Full Text | Google Scholar

Babarović, T. (2022). Development of STEM vocational interests during elementary and middle school: A cohort-sequential longitudinal study. J. Career Dev. 49, 1230–1250. doi: 10.1177/08948453211036986

Crossref Full Text | Google Scholar

Ball, C., Huang, K. T., Cotten, S. R., and Rikard, R. V. (2017). Pressurizing the STEM pipeline: An expectancy-value theory analysis of youths’ STEM attitudes. J. Sci. Educ. Technol. 26, 372–382. doi: 10.1007/s10956-017-9685-1

Crossref Full Text | Google Scholar

Baran, E., Canbazoglu Bilici, S., Mesutoglu, C., and Ocak, C. (2019). The impact of an out-of-school STEM education program on students’ attitudes toward STEM and STEM careers. Sch. Sci. Mathemat. 119, 223–235. doi: 10.1111/ssm.12330

Crossref Full Text | Google Scholar

Barchas-Lichtenstein, J., Sherman, M., Voiklis, J., and Clapman, L. (2023). Science through storytelling or storytelling about science? Identifying cognitive task demands and expert strategies in cross-curricular STEM education. Front. Educ. 8:1279861. doi: 10.3389/feduc.2023.1279861

Crossref Full Text | Google Scholar

Bartholomew, S. R., and Santana, V. E. (2021). Investigating the impact of children’s STEM literature on perceptions of abilities and potential for employment in STEM fields and careers. SN Soc. Sci. 1:217. doi: 10.1007/s43545-021-00218-6

Crossref Full Text | Google Scholar

Bell, R. L., Smetana, L., and Binns, I. (2005). Simplifying inquiry instruction. Sci. Teach. 72, 30–33.

Google Scholar

Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Statist. Soc. Ser. B 57, 289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x

Crossref Full Text | Google Scholar

Bleacher, L. V., Meinke, B., Hauck, K., Soeffing, C., and Spitz, A. (2014). “NASA Science4Girls and their families: Connecting local libraries with NASA scientists and education programs to engage girls in STEM (No. GSFC-E-DAA-TN13875),” in Paper presented at the Lunar and Planetary Sciences Conference, (The Woodlands, TX).

Google Scholar

Bronfenbrenner, U., and Morris, P. A. (2006). “The bioecological model of human development,” in Handbook of child psychology: Theoretical models of human development, 6th Edn, eds W. Damon and R. M. Lerner (Hoboken, NJ: Wiley), 793–828.

Google Scholar

Bulca, Y., Bilgin, E., Altay, F., and Demirhan, G. (2022). Effects of a short video physical activity program on physical fitness among physical education students. Percept. Motor Skills 129, 932–945. doi: 10.1177/00315125221088069

PubMed Abstract | Crossref Full Text | Google Scholar

Burgstahler, S., and Cauce, A. M. (2020). Creating inclusive learning opportunities in higher education: A universal design toolkit. Cambridge, MA: Harvard Education Press, 47–48.

Google Scholar

Bustamante, A. S., Bermudez, V. N., Ochoa, K. D., Belgrave, A. B., and Vandell, D. L. (2023). Quality of early childcare and education predicts high school STEM achievement for students from low-income backgrounds. Dev. Psychol. 59, 1440–1451. doi: 10.1037/dev0001546

PubMed Abstract | Crossref Full Text | Google Scholar

Cabell, S. Q., and Hwang, H. (2020). Building content knowledge to boost comprehension in the primary grades. Read. Res. Quar. 55, S99–S107. doi: 10.1002/rrq.304

Crossref Full Text | Google Scholar

Cabell, S. Q., Zucker, T. A., DeCoster, J., Copp, S. B., and Landry, S. (2019). Impact of a parent text messaging program on pre-kindergarteners’ literacy development. AERA Open 5:2332858419833339. doi: 10.1177/2332858419833339

Crossref Full Text | Google Scholar

Calabrese Barton, A., Kang, H., Tan, E., O’Neill, T. B., Bautista-Guerra, J., and Brecklin, C. (2013). Crafting a future in science: Tracing middle school girls’ identity work over time and space. Am. Educ. Res. J. 50, 37–75. doi: 10.3102/0002831212458142

PubMed Abstract | Crossref Full Text | Google Scholar

Callanan, M. A., Castañeda, C. L., Luce, M. R., and Martin, J. L. (2017). Family science talk in museums: Predicting children’s engagement from variations in talk and activity. Child Dev. 88, 1492–1504. doi: 10.1111/cdev.12702

PubMed Abstract | Crossref Full Text | Google Scholar

Carol-Ann Burke, L. E. (2020). Informal science educators and children in a low-income community describe how children relate to out-of-school science education. Int. J. Sci. Educ. 42, 1673–1696. doi: 10.1080/09500693.2020.1781574

Crossref Full Text | Google Scholar

Chestnut, E. K., and Markman, E. M. (2018). “Girls are as good as boys at math” implies that boys are probably better: A study of expressions of gender equality. Cogn. Sci. 42, 2229–2249. doi: 10.1111/cogs.12642

Crossref Full Text | Google Scholar

Chittum, J. R., Jones, B. D., Akalin, S., and Schram, ÁB. (2017). The effects of an afterschool STEM program on students’ motivation and engagement. Int. J. STEM Educ. 4:11. doi: 10.1186/s40594-017-0065-4

PubMed Abstract | Crossref Full Text | Google Scholar

Cimpian, J. R., Kim, T. H., and McDermott, Z. T. (2020). Understanding persistent gender gaps in STEM. Science 368, 1317–1319. doi: 10.1126/science.abb2763

Crossref Full Text | Google Scholar

Crispin, L. M., and Beck, M. I. (2025). Disparities in museum attendance among youth over two decades: An empirical analysis of who attends and how often. Arts Educ. Policy Rev. 126, 25–37. doi: 10.1080/10632913.2023.1234567

Crossref Full Text | Google Scholar

Curran, F. C., and Kitchin, J. (2019). Early elementary science instruction: Does more time on science or science topics/skills predict science achievement in the early grades? AERA Open 5:2332858419861081. doi: 10.1177/2332858419861081

Crossref Full Text | Google Scholar

Cvencek, D., Meltzoff, A. N., and Kapur, M. (2014). Cognitive consistency and math–gender stereotypes in Singaporean children. J. Exp. Child Psychol. 117, 73–91. doi: 10.1016/j.jecp.2013.09.002

PubMed Abstract | Crossref Full Text | Google Scholar

Dabney, K. P., Tai, R. H., and Scott, M. R. (2016). Informal science: Family education, experiences, and initial interest in science. Int. J. Sci. Educ. Part B 6, 263–282. doi: 10.1080/21548455.2015.1093679

Crossref Full Text | Google Scholar

Dahlstrom, M. F. (2014). Using narratives and storytelling to communicate science with nonexpert audiences. Proc. Natl. Acad. Sci. U. S. A. 111, 13614–13620. doi: 10.1073/pnas.1320645111

PubMed Abstract | Crossref Full Text | Google Scholar

Dorsen, J., Carlson, B., and Goodyear, L. (2006). Connecting informal STEM experiences to career choices: Identifying the pathway. Waltham, MA: ITEST Learning Resource Center.

Google Scholar

Dowdall, N., Melendez-Torres, G. J., Murray, L., Gardner, F., Hartford, L., and Cooper, P. J. (2019). Shared picture book reading interventions for child language development: A systematic review and meta-analysis. Child Dev. 91, e383–e399. doi: 10.1111/cdev.13225

PubMed Abstract | Crossref Full Text | Google Scholar

Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C., Klebanov, P., et al. (2007). School readiness and later achievement. Dev. Psychol. 43, 1428–1446. doi: 10.1037/0012-1649.43.6.1428

PubMed Abstract | Crossref Full Text | Google Scholar

Eberbach, C., and Crowley, K. (2005). From living to virtual: Learning from museum objects. Curator: Museum J. 48, 317–338. doi: 10.1111/j.2151-6952.2005.tb00175.x

Crossref Full Text | Google Scholar

Eberbach, C., and Crowley, K. (2017). From seeing to observing: How parents and children learn to see science in a botanical garden. J. Learn. Sci. 26, 608–642. doi: 10.1080/10508406.2017.1308867

Crossref Full Text | Google Scholar

Eccles, J. S. (2015). Gendered socialization of STEM interests in the family. Int. J. Gender Sci. Technol. 7, 116–132.

Google Scholar

Eccles, J. S., and Wigfield, A. (2020). From expectancy-value theory to situated expectancy-value theory: A developmental, social cognitive, and sociocultural perspective on motivation. Contemp. Educ. Psychol. 61:101859. doi: 10.1016/j.cedpsych.2020.101859

Crossref Full Text | Google Scholar

Eccles, J. S., and Wigfield, A. (2023). Expectancy-value theory to situated expectancy-value theory: Reflections on the legacy of 40+ years of working together. Motivat. Sci. 9, 1–12. doi: 10.1037/mot0000275

Crossref Full Text | Google Scholar

Edwards, O. W., and Oakland, T. D. (2006). Factorial invariance of Woodcock-Johnson III scores for African Americans and caucasian Americans. J. Psychoeduc. Assess. 24, 358–366. doi: 10.1177/0734282906289595

Crossref Full Text | Google Scholar

Edwards, P. A., and Danridge, J. C. (2020). “Developing collaboration with culturally diverse parents,” in Collaboration for diverse learners, eds V. J. Risko and K. Bromley (England: Routledge), 251–272.

Google Scholar

Ennes, M. E., Jones, M. G., Childers, G. M., Cayton, E. M., and Chesnutt, K. M. (2023). Children and parents’ perceptions of access to science tools at home and their role in science self-efficacy. Res. Sci. Educ. 53, 671–687. doi: 10.1007/s11165-021-10002-2

Crossref Full Text | Google Scholar

Feingold, A. (2013). A regression framework for effect size assessments in longitudinal modeling of group differences. Rev. General Psychol. 17, 111–121. doi: 10.1037/a0030048

PubMed Abstract | Crossref Full Text | Google Scholar

Fernandez, F., Froschl, M., Lorenzetti, L., and Stimmer, M. (2024). Investigating the importance of girls’ mathematical identity within United States STEM programmes: A systematic review. Int. J. Mathemat. Educ. Sci. Technol. 55, 650–690. doi: 10.1080/0020739X.2021.2022229

Crossref Full Text | Google Scholar

Fielding-Wells, J., O’Brien, M., and Makar, K. (2017). Using expectancy-value theory to explore aspects of motivation and engagement in inquiry-based learning in primary mathematics. Mathemat. Educ. Res. J. 29, 237–254. doi: 10.1007/s13394-016-0182-7

Crossref Full Text | Google Scholar

French, L. (2004). Science as the center of a coherent, integrated early childhood curriculum. Early Childhood Res. Quar. 19, 138–149. doi: 10.1016/j.ecresq.2004.01.005

Crossref Full Text | Google Scholar

Goff, E. E., Mulvey, K. L., Irvin, M. J., and Hartstone-Rose, A. (2020). The effects of prior informal science and math experiences on undergraduate STEM identity. Res. Sci. Technol. Educ. 38, 272–288. doi: 10.1080/02635143.2019.1690712

Crossref Full Text | Google Scholar

Golke, S., and Wittwer, J. (2024). Informative narratives increase students’ situational interest in science topics. Learn. Instruct. 93:101973. doi: 10.1016/j.learninstruc.2024.101973

Crossref Full Text | Google Scholar

Grindal, T., Bowne, J. B., Yoshikawa, H., Schindler, H. S., Duncan, G. J., Magnuson, K., et al. (2016). The added impact of parenting education in early childhood education programs: A meta-analysis. Child. Youth Serv. Rev. 70, 238–249. doi: 10.1016/j.childyouth.2016.09.018

Crossref Full Text | Google Scholar

Grusec, J. E., and Davidov, M. (2010). Integrating different perspectives on socialization theory and research: A domain-specific approach. Child Dev. 81, 687–709. doi: 10.1111/j.1467-8624.2010.01426.x

PubMed Abstract | Crossref Full Text | Google Scholar

Guo, J., Marsh, H. W., Parker, P. D., and Hu, X. (2024). Cross-Cultural patterns of gender differences in STEM: Gender stratification, gender equality and gender-equality paradoxes. Educ. Psychol. Rev. 36:37. doi: 10.1007/s10648-023-09777-1

Crossref Full Text | Google Scholar

Habig, B., Gupta, P., Levine, B., and Adams, J. (2020). An informal science education program’s impact on STEM major and STEM career outcomes. Res. Sci. Educ. 50, 1051–1074. doi: 10.1007/s11165-019-09862-7

Crossref Full Text | Google Scholar

Haden, C. A. (2010). Talking about science in museums. Child Dev. Perspect. 4, 62–67. doi: 10.1111/j.1750-8606.2009.00119.x

Crossref Full Text | Google Scholar

Haden, C. A., Jant, E. A., Hoffman, P. C., Marcus, M., Geddes, J. R., and Gaskins, S. (2014). Supporting family conversations and children’s STEM learning in a children’s museum. Early Childhood Res. Quar. 29, 333–344. doi: 10.1016/j.ecresq.2014.04.006

Crossref Full Text | Google Scholar

Hazari, Z., Potvin, G., Cribbs, J. D., Godwin, A., Scott, T. D., and Klotz, L. (2017). Interest in STEM is contagious for students in biology, chemistry, and physics classes. Sci. Adv. 3:e1700046. doi: 10.1126/sciadv.1700046

PubMed Abstract | Crossref Full Text | Google Scholar

Heberle, A. E., and Carter, A. S. (2015). Cognitive aspects of young children’s experience of economic disadvantage. Psychol. Bull. 141, 723–744. doi: 10.1037/a0039463

PubMed Abstract | Crossref Full Text | Google Scholar

Hill, H. C., and Erickson, A. S. G. (2019). Using implementation fidelity data to inform teacher professional learning. Educ. Res. 48, 677–686. doi: 10.3102/0013189X19891436

PubMed Abstract | Crossref Full Text | Google Scholar

Holian, L., and Kelly, E. (2020). STEM occupational intentions stability and change through high school. Stats in Brief. NCES 2020-167. Washington, D.C: National Center for Education Statistics.

Google Scholar

Hurwitz, L. B., Lauricella, A. R., Hanson, A., Raden, A., and Wartella, E. (2015). Supporting head start parents: Impact of a text message intervention on parent–child activity engagement. Early Child Dev. Care 185, 1373–1389. doi: 10.1080/03004430.2015.1014492

Crossref Full Text | Google Scholar

Irwin, S., and Elley, S. (2013). Parents’ hopes and expectations for their children’s future occupations. Sociol. Rev. 61, 111–130. doi: 10.1111/1467-954X.12000

Crossref Full Text | Google Scholar

Jenifer, J. B., Jaxon, J., Levine, S. C., and Cimpian, A. (2024). You need to be super smart to do well in math!” Young children’s field-specific ability beliefs. Dev. Sci. 27:e13429. doi: 10.1111/desc.13429

PubMed Abstract | Crossref Full Text | Google Scholar

Jiang, H., Zhang, L., and Zhang, W. (2024). Influence of career awareness on STEM career interests: Examining the roles of self-efficacy, outcome expectations, and gender. Int. J. STEM Educ. 11:22. doi: 10.1186/s40594-024-00400-x

Crossref Full Text | Google Scholar

Kaderavek, J. N., Paprzycki, P., Czerniak, C. M., Hapgood, S., Mentzer, G., Molitor, S., et al. (2020). Longitudinal impact of early childhood science instruction on 5th grade science achievement. Int. J. Sci. Educ. 42, 1124–1143. doi: 10.1080/09500693.2020.1737505

Crossref Full Text | Google Scholar

Kim, A. Y., Sinatra, G. M., and Seyranian, V. (2018). Developing a STEM identity among young women: A social identity perspective. Rev. Educ. Res. 88, 589–625. doi: 10.3102/0034654318791587

PubMed Abstract | Crossref Full Text | Google Scholar

Kurtz-Costes, B., Copping, K. E., Rowley, S. J., and Kinlaw, C. R. (2014). Gender and age differences in awareness and endorsement of gender stereotypes about academic abilities. Eur. J. Psychol. Educ. 29, 575–594. doi: 10.1007/s10212-014-0216-7

Crossref Full Text | Google Scholar

Le, H., and Robbins, S. B. (2016). Building the STEM pipeline: Findings of a 9-year longitudinal research project. J. Vocat. Behav. 95, 21–30. doi: 10.1016/j.jvb.2016.06.004

Crossref Full Text | Google Scholar

Lee, J., Lee, H. J., Song, J., and Bong, M. (2021). Enhancing children’s math motivation with a joint intervention on mindset and gender stereotypes. Learn. Instruct. 73:101416. doi: 10.1016/j.learninstruc.2020.101416

Crossref Full Text | Google Scholar

Leech, K. A., Haber, A. S., Jalkh, Y., and Corriveau, K. H. (2020). Embedding scientific explanations into storybooks impacts children’s scientific discourse and learning. Front. Psychol. 11:1016. doi: 10.3389/fpsyg.2020.01016

PubMed Abstract | Crossref Full Text | Google Scholar

Levine, R. S. (2024). Out-of-School time sponsors and partners: A review of programs for low-income adolescents. Afterschool Matters 38, 19–28.

Google Scholar

Lewis, N. (2019). On ‘light-touches’ and ‘heavy-hands’: 2 strategies to tackle educational inequities. Washington, D.C: The Brookings Institute Brown Chalkboard Center.

Google Scholar

Low, K. S., Yoon, M., Roberts, B. W., and Rounds, J. (2005). The stability of vocational interests from early adolescence to middle adulthood: A quantitative review of longitudinal studies. Psychol. Bull. 131, 713–737. doi: 10.1037/0033-2909.131.5.713

PubMed Abstract | Crossref Full Text | Google Scholar

Lu, J., Schmidt, M., Lee, M., and Huang, R. (2022). Usability research in educational technology: A state-of-the-art systematic review. Educ. Technol. Res. Dev. 70, 1951–1992. doi: 10.1007/s11423-022-10152-6

Crossref Full Text | Google Scholar

Maltese, A. V., and Cooper, C. S. (2017). STEM pathways: Do men and women differ in why they enter and exit? AERA Open 3:2332858417727276. doi: 10.1177/2332858417727276

Crossref Full Text | Google Scholar

Maltese, A. V., and Tai, R. H. (2011). Pipeline persistence: Examining the association of educational experiences with earned degrees in STEM among U.S. children. Sci. Educ. 95, 877–907. doi: 10.1002/sce.20402

Crossref Full Text | Google Scholar

Marcus, M., Haden, C. A., and Uttal, D. H. (2017). STEM learning and transfer in a children’s museum and beyond. Merrill-Palmer Quar. 63, 155–180. doi: 10.13110/merrpalmquar1982.63.2.0155

Crossref Full Text | Google Scholar

Marcus, M., Haden, C. A., and Uttal, D. H. (2018). Promoting children’s learning and transfer across informal science, technology, engineering, and mathematics learning experiences. J. Exp. Child Psychol. 175, 80–95. doi: 10.1016/j.jecp.2018.06.003

PubMed Abstract | Crossref Full Text | Google Scholar

Marcus, M., Solis, G., Sellars, S., and Haden, C. A. (2023). Promoting children’s science, technology, engineering, and mathematics learning at home through tinkering and storytelling. Front. Psychol. 14:1146063. doi: 10.3389/fpsyg.2023.1146063

PubMed Abstract | Crossref Full Text | Google Scholar

Marcus, M., Tõugu, P., Haden, C. A., and Uttal, D. H. (2021). Advancing opportunities for children’s informal STEM learning transfer through parent–child narrative reflection. Child Dev. 92, e1075–e1084. doi: 10.1111/cdev.13522

PubMed Abstract | Crossref Full Text | Google Scholar

Martinot, D., and Désert, M. (2007). Awareness of a gender stereotype, personal beliefs and self-perceptions regarding math ability: When boys do not surpass girls. Soc. Psychol. Educ. 10, 455–471. doi: 10.1007/s11218-007-9035-3

Crossref Full Text | Google Scholar

Master, A. (2021). Gender stereotypes influence children’s STEM motivation. Child Dev. Perspect. 15, 203–210. doi: 10.1111/cdep.12414

Crossref Full Text | Google Scholar

Master, A., Alexander, T., Thompson, J., Fan, W., Meltzoff, A. N., and Cheryan, S. (2025). Causes and consequences of stereotypes: Interest stereotypes reduce adolescent girls’ motivation to enroll in computer science classes. J. Res. Technol. Educ. 57, 56–83. doi: 10.1080/15391523.2024.2402355

Crossref Full Text | Google Scholar

Master, A., Cheryan, S., Moscatelli, A., and Meltzoff, A. N. (2017). Programming experience promotes higher STEM motivation among first-grade girls. J. Exp. Child Psychol. 160, 92–106. doi: 10.1016/j.jecp.2017.03.013

PubMed Abstract | Crossref Full Text | Google Scholar

Master, A., Meltzoff, A. N., and Cheryan, S. (2021). Gender stereotypes about interests start early and cause gender disparities in computer science and engineering. Proc. Natl. Acad. Sci. U. S. A. 118:e2100030118. doi: 10.1073/pnas.2100030118

PubMed Abstract | Crossref Full Text | Google Scholar

Meschede, T., Haque, Z., Warfield, M. E., Melchior, A., Burack, C., and Hoover, M. (2022). Transforming STEM outcomes: Results from a seven-year follow-up study of an after-school robotics program’s impacts on freshman students. School Sci. Mathemat. 122, 343–357. doi: 10.1111/ssm.12552

Crossref Full Text | Google Scholar

Miller, D. I., Lauer, J. E., Tanenbaum, C., and Burr, L. (2024). The development of children’s gender stereotypes about STEM and verbal abilities: A preregistered meta-analytic review of 98 studies. Psychol. Bull. 150, 1363–1396. doi: 10.1037/bul0000440

Crossref Full Text | Google Scholar

Mol, S. E., Bus, A. G., and De Jong, M. T. (2009). Interactive book reading in early education: A tool to stimulate print knowledge as well as oral language. Rev. Educ. Res. 79, 979–1007. doi: 10.3102/0034654309332561

PubMed Abstract | Crossref Full Text | Google Scholar

Morgan, P. L., Farkas, G., Hillemeier, M. M., and Maczuga, S. (2016). Science achievement gaps begin very early, persist, and are largely explained by modifiable factors. Educ. Res. 45, 18–35. doi: 10.3102/0013189X16633182

PubMed Abstract | Crossref Full Text | Google Scholar

Morgan, S. L., Gelbgiser, D., and Weeden, K. A. (2013). Feeding the pipeline: Gender, occupational plans, and college major selection. Soc. Sci. Res. 42, 989–1005. doi: 10.1016/j.ssresearch.2013.04.005

PubMed Abstract | Crossref Full Text | Google Scholar

Morrissey, K. (2020). A Guide to the ISL professional competency framework. Available online at: http://www.islframework.org/ (accessed April 1, 2025).

Google Scholar

Mulvey, K. L., McGuire, L., Hoffman, A. J., Goff, E., Rutland, A., Winterbottom, M., et al. (2020). Interest and learning in informal science learning sites: Differences in experiences with different types of educators. PLoS One 15:e0236279. doi: 10.1371/journal.pone.0236279

PubMed Abstract | Crossref Full Text | Google Scholar

National Center for Education Statistics [NCES] (2023). National assessments of educational progress. School pulse panel: Surveying high-priority, education-related topics: Results. Washington, D.C: U.S. Department of Education.

Google Scholar

National Center for Education Statistics [NCES] (2024). Most U.S. Public K–12 schools offer after–school programs but many cannot accommodate all students who want to participate. School Pulse Panel. Available online at: https://nces.ed.gov/whatsnew/press_releases/11_14_2024.asp (accessed April 1, 2025).

Google Scholar

National Center for Science and Engineering Statistics [NCSES] (2021). National survey of college graduates: 2019 (NSF 22-310). Alexandria, VA: National Science Foundation.

Google Scholar

National Center for Science and Engineering Statistics [NCSES] (2023). Diversity and STEM: Women, minorities, and persons with disabilities 2023 (NSF 23-315). Alexandria, VA: National Science Foundation.

Google Scholar

National Research Council [NRC] (2009). Learning science in informal environments: People, places, and pursuits. Washington, DC: The National Academies Press. doi: 10.17226/12190

Crossref Full Text | Google Scholar

National Research Council [NRC] (2012). A framework for K-12 science education: Practices, crosscutting concepts, and core ideas. Washington, DC: The National Academies Press.

Google Scholar

National Research Council [NRC] (2015). Identifying and supporting productive STEM programs in out-of-school settings. Washington, DC: The National Academies Press, doi: 10.17226/21740

Crossref Full Text | Google Scholar

Neeley, L., Barker, E., Bayer, S. R., Maktoufi, R., Wu, K. J., and Zaringhalam, M. (2020). Linking scholarship and practice: Narrative and identity in science. Front. Commun. 5:35. doi: 10.3389/fcomm.2020.00035

Crossref Full Text | Google Scholar

Neuman, S. B., Samudra, P., and Danielson, K. (2021). Effectiveness of scaling up a vocabulary intervention for low-income children, pre-K through first grade. Elemen. School J. 121, 385–409. doi: 10.1086/713099

Crossref Full Text | Google Scholar

Offenstein, J., Moore, C., and Shulock, N. (2009). Pathways to success: Lessons from the literature on career technical education. Sacramento, CA: California State University, Institute for Higher Education Leadership & Policy.

Google Scholar

Olsson, M., and Martiny, S. E. (2018). Does exposure to counterstereotypical role models influence girls’ and women’s gender stereotypes and career choices? A review of social psychological research. Front. Psychol. 9:2264. doi: 10.3389/fpsyg.2018.02264

PubMed Abstract | Crossref Full Text | Google Scholar

Papazian, A. E., Noam, G. G., Shah, A. M., and Rufo-McCormick, C. (2013). The quest for quality in afterschool science: The development and application of a new tool. Afterschool Matters 18, 17–24. Available online at: https://www.niost.org/images/pdf/afterschoolmatters/asm_2013_18_fall/Pages

Google Scholar

Patrick, H., Mantzicopoulos, P., and Samarapungavan, A. (2009). Motivation for learning science in kindergarten: Is there a gender gap and does integrated inquiry and literacy instruction make a difference. J. Res. Sci. Teach. 46, 166–191. doi: 10.1002/tea.20276

Crossref Full Text | Google Scholar

Pattison, S., and Ramos Montañez, S. (2022). “Diverse STEM interest development pathways in early childhood,” in Play and STEM education in the early years: International policies and practices, eds I. S. Kucirkova, E. B. Skarstein, and S. J. Uden (Berlin: Springer International Publishing), 439–457. doi: 10.1007/978-3-030-98153-4_23

Crossref Full Text | Google Scholar

Pellegrini, M., Lake, C., Neitzel, A., and Slavin, R. E. (2021). Effective programs in elementary mathematics: A meta-analysis. AERA Open 7:2332858420986211. doi: 10.1177/2332858420986211

Crossref Full Text | Google Scholar

Pinkard, N., Erete, S., Caitlin, M., Majors, Y., and Walker, N. (2025). Increasing STEM engagement through opportunity landscaping. Acta Psychol. 253:104705. doi: 10.1016/j.actpsy.2025.104705

PubMed Abstract | Crossref Full Text | Google Scholar

Raabe, I. J., Boda, Z., and Stadtfeld, C. (2019). The Social Pipeline: How friend influence and peer exposure widen the STEM gender gap. Sociol. Educ. 92, 105–123. doi: 10.1177/0038040718824095

Crossref Full Text | Google Scholar

Riegle-Crumb, C., and Peng, M. (2021). Examining high school children’ gendered beliefs about math: Predictors and implications for choice of STEM college majors. Sociol. Educ. 94, 227–248. doi: 10.1177/00380407211005462

Crossref Full Text | Google Scholar

Riegle-Crumb, C., Moore, C., and Buontempo, J. (2017). Shifting STEM stereotypes? Considering the role of peer and teacher gender. J. Res. Adolesc. 27, 492–505. doi: 10.1111/jora.12286

Crossref Full Text | Google Scholar

Rosenzweig, E. Q., and Chen, X. Y. (2023). Which STEM careers are most appealing? Examining high school students’ preferences and motivational beliefs for different STEM career choices. Int. J. STEM Educ. 10:40. doi: 10.1186/s40594-023-00424-6

Crossref Full Text | Google Scholar

Sansone, D. (2019). Teacher characteristics, child beliefs, and the gender gap in STEM fields. Educ. Eval. Policy Anal. 41, 127–144. doi: 10.3102/0162373719834063

PubMed Abstract | Crossref Full Text | Google Scholar

Santiago, A., Pederson, K., and Karl, R. (2019). The scigirls strategies: How to engage girls in stem. Connected Sci. Learn. 1:12420566. doi: 10.2505/connectedsciencelearning.2019.12420566

Crossref Full Text | Google Scholar

Schrank, F. A., Mather, N., and McGrew, K. S. (2014). Woodcock-Johnson IV tests of achievement, extended Version, Test 18 science. Rolling Meadows, IL: Riverside.

Google Scholar

Shaby, N., Dillon, J., Peleg, R., Assaraf, O. B. Z., Pattison, S., Pierroux, P., et al. (2025). Telling tales: The use of narratives in informal STEM settings. Res. Sci. Technol. Educ. doi: 10.1080/02635143.2025.2469065

Crossref Full Text | Google Scholar

Shah, A. M., Wylie, C., Gitomer, D., and Noam, G. (2018). Improving STEM program quality in out-of-school-time: Tool development and validation. Sci. Educ. 102, 238–259. doi: 10.1002/sce.21327

Crossref Full Text | Google Scholar

Snell, E. K., Hindman, A. H., and Wasik, B. A. (2020). Exploring the use of texting to support family-school engagement in early childhood settings: Teacher and family perspectives. Early Child Dev. Care 190, 447–460. doi: 10.1080/03004430.2018.1479401

Crossref Full Text | Google Scholar

Speer, J. D. (2023). Bye bye Ms. American Sci: Women and the leaky STEM pipeline. Econ. Educ. Rev. 93:102371. doi: 10.1016/j.econedurev.2023.102371

Crossref Full Text | Google Scholar

Starr, C. R., and Simpkins, S. D. (2021). High school children’ math and science gender stereotypes: Relations with their STEM outcomes and socializers’ stereotypes. Soc. Psychol. Educ. 24, 273–298. doi: 10.1007/s11218-020-09586-5

Crossref Full Text | Google Scholar

Steffens, M. C., Jelenec, P., and Noack, P. (2010). On the leaky math pipeline: Comparing implicit math-gender stereotypes and math withdrawal in female and male children and adolescents. J. Educ. Psychol. 102, 947–963. doi: 10.1037/a0019920

Crossref Full Text | Google Scholar

Steinke, J., and Duncan, T. (2023). Challenging media stereotypes of STEM: Examining an intervention to change adolescent girls’ gender stereotypes of STEM professionals. Int. J. Gender Sci. Technol. 15, 136–165. Available online at: https://genderandset.open.ac.uk/index.php/genderandset/article/view/1428

Google Scholar

Sullivan, A., and Bers, M. U. (2019). Investigating the use of robotics to increase girls’ interest in engineering during early elementary school. Int. J. Technol. Design Educ. 29, 1033–1051. doi: 10.1007/s10798-018-9482-z

Crossref Full Text | Google Scholar

Sullivan, A. A. (2019). Breaking the STEM stereotype: Reaching girls in early childhood. Lanham, MD: Rowman & Littlefield.

Google Scholar

Suter, L. E. (2014). Visiting science museums during middle and high school: A longitudinal analysis of student performance in science. Sci. Educ. 98, 815–839. doi: 10.1002/sce.21114

Crossref Full Text | Google Scholar

Tang, D., Meltzoff, A. N., Cheryan, S., Fan, W., and Master, A. (2024). Longitudinal stability and change across a year in children’s gender stereotypes about four different STEM fields. Dev. Psychol. 60, 1109–1130. doi: 10.1037/dev0001612

PubMed Abstract | Crossref Full Text | Google Scholar

Tobler, S., Sinha, T., Köhler, K., Hafen, E., and Kapur, M. (2024). Impact of narrative versus expository instruction in science education on recall, understanding, and transfer: A meta-analysis. ArXiv [Preprint] doi: 10.3929/ethz-b-000696792

Crossref Full Text | Google Scholar

Trelease, J. (2013). The read-aloud handbook, 7th Edn. London: Penguin.

Google Scholar

U.S. Department of Education, Institute of Education Sciences, and National Center for Education Statistics (2024). School Pulse Panel 2021–22, 2022–23, 2023–24, and 2024–25. Alexandria, VA: National Science Foundation.

Google Scholar

United Nations Children International Children’s Fund [UNICEF] (2020). Towards an equal future: Reimagining girls’ education through STEM. New York, NY: UNICEF.

Google Scholar

University of Texas Health Science Center at Houston (2024). Afterschool STEM programs. Houston, TX: University of Texas Health Science Center at Houston.

Google Scholar

van den Hurk, A., Meelissen, M., and van Langen, A. (2019). Interventions in education to prevent STEM pipeline leakage. Int. J. Sci. Educ. 41, 150–164. doi: 10.1080/09500693.2018.1549372

Crossref Full Text | Google Scholar

VanMeter-Adams, A., Frankenfeld, C. L., Bases, J., Espina, V., and Liotta, L. A. (2014). Students who demonstrate strong talent and interest in STEM are initially attracted to STEM through extracurricular experiences. CBE—Life Sci. Educ. 13, 687–697. doi: 10.1187/cbe.14-04-0066

PubMed Abstract | Crossref Full Text | Google Scholar

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes, Vol. 86. Harvard university press.

Google Scholar

Walan, S. (2019). Teaching children science through storytelling combined with hands-on activities–a successful instructional strategy? Education 47, 34–46. doi: 10.1080/03004279.2017.1386228

Crossref Full Text | Google Scholar

Wang, M. M., Cardarelli, A., Brenner, J., Leslie, S. J., and Rhodes, M. (2025). Maladaptive but malleable: Gender-science stereotypes emerge early but are modifiable by language. Child Dev. 96, 865–880. doi: 10.1111/cdev.14089

PubMed Abstract | Crossref Full Text | Google Scholar

Wang, M. T., and Degol, J. (2013). Motivational pathways to STEM career choices: Using expectancy–value perspective to understand individual and gender differences in STEM fields. Dev. Rev. 33, 304–340. doi: 10.1016/j.dr.2013.08.001

PubMed Abstract | Crossref Full Text | Google Scholar

Weible, J. L., and Zimmerman, H. T. (2016). Science curiosity in learning environments: Developing an attitudinal scale for research in schools, homes, museums, and the community. Int. J. Sci. Educ. 38, 1235–1255. doi: 10.1080/09500693.2016.1186853

Crossref Full Text | Google Scholar

Weisgram, E. S., and Bigler, R. S. (2006). Girls and science careers: The role of altruistic values and attitudes about scientific tasks. J. Appl. Dev. Psychol. 27, 326–348. doi: 10.1016/j.appdev.2006.04.004

Crossref Full Text | Google Scholar

West, J., Tarullo, L., Aikens, N., Malone, L., and Carlson, B. L. (2007). FACES 2009 study design (no. b14256669cb34ed2a584d459442fe076). Washington, DC: Mathematica Policy Research.

Google Scholar

What Works Clearinghouse (2015). WWC guidance for assessing attrition bias (WWC Brief No. 080715). U.S. Department of Education, Institute of Education Sciences, National Center for Education Evaluation and Regional Assistance. Washington, DC: Institute of Education Sciences [IES].

Google Scholar

White, G. W., Stepney, C. T., Hatchimonji, D. R., Moceri, D. C., Linsky, A. V., Reyes-Portillo, J. A., et al. (2016). The increasing impact of socioeconomics and race on standardized academic test scores across elementary, middle, and high school. Am. J. Orthopsychiatry 86, 10–21. doi: 10.1037/ort0000137

PubMed Abstract | Crossref Full Text | Google Scholar

Wright, T. S., Cervetti, G. N., Wise, C., and McClung, N. A. (2022). The impact of knowledge-building through conceptually-coherent read alouds on vocabulary and comprehension. Read. Psychol. 43, 70–84. doi: 10.1080/02702711.2021.1974392

Crossref Full Text | Google Scholar

Xia, X., Bentley, L. R., Fan, X., and Tai, R. H. (2025). STEM outside of school: A meta-analysis of the effects of informal science education on students’ interests and attitudes for STEM. Int. J. Sci. Mathemat. Educ. 23, 1153–1181. doi: 10.1007/s10763-024-10504-z

Crossref Full Text | Google Scholar

Young, J., Ortiz, N., and Young, J. (2017). STEMulating interest: A meta-analysis of the effects of out-of-school time on student STEM interest. Int. J. Educ. Mathemat. Sci. Technol. 5, 62–74. doi: 10.18404/ijemst.61149

Crossref Full Text | Google Scholar

Zhou, Y., and Shirazi, S. (2025). Factors influencing young people’s STEM career aspirations and career choices: A systematic literature review. Int. J. Sci. Math. Educ. doi: 10.1007/s10763-025-10552-z

Crossref Full Text | Google Scholar

Zucker, T., Mesa, M. P., DeMaster, D., Oh, Y., Assel, M., McCallum, C., et al. (2024). Evaluation of a community-based, hybrid STEM family engagement program at pre-kindergarten entry. Front. Educ. 9:1281161. doi: 10.3389/feduc.2024.1281161

Crossref Full Text | Google Scholar

Zucker, T. A., Maldonado, G. Y., Assel, M., McCallum, C., Elias, C., Swint, J. M., et al. (2022). Informal science, technology, engineering and math learning conditions to increase parent involvement with young children experiencing poverty. Front. Psychol. 13:1015590. doi: 10.3389/fpsyg.2022.1015590

PubMed Abstract | Crossref Full Text | Google Scholar