Improving Girls’ Sense of Fit in Science

Abstract

A field experiment tested the effect of choosing a “favorite” role model on sense of fit in science among middle school girls. The girls participated in a 1-day science outreach event where they were exposed to multiple female role models. At the end of the event, participants were randomly assigned to choose and write about a favorite role model or to choose and write about their best friend. Girls in the role model choice condition experienced a significant increase in sense of fit in science, whereas girls in the control condition did not. Girls in the role model choice condition also tended to have stronger role model identification than girls in the control condition, and role model identification was correlated with sense of fit in science. Encouraging girls to actively choose and write about a favorite role model may help to maximize the impact of exposure to role models.

Keywords

stigma adolescent role model STEM sense of fit

Girls and women face many obstacles to pursuing careers in fields related to science, technology, engineering, and mathematics (STEM; Ceci, Ginther, Kahn, & Williams, 2014; Hill, Corbett, & St. Rose, 2010). Girls and women face pervasive negative stereotypes about their science abilities, may encounter sexism in STEM contexts, and may be subtly (and not so subtly) discouraged from pursuing careers in STEM by parents and teachers (Eccles, 2011; Milkman, Akinola, & Chugh, 2015; Moss-Racusin, Dovidio, Brescoll, Graham, & Handelsman, 2012; Nosek & Smyth, 2011; Williams, Phillips, & Hall, 2014). The presence of female role models, however, can serve as a buffer and improve girls’ and women’s interest, performance, and motivation in STEM fields (e.g., Dasgupta, 2011; Liben & Coyle, 2014; Marx & Roman, 2002).

The present research examines the impact of female role models on girls’ sense of fit in science. Although research has previously demonstrated the benefits of exposure to female role models for STEM outcomes, the present study tests the novel hypothesis that identifying and choosing a favorite role model and constructing a narrative about why she is a favorite can improve girls’ sense of fit above and beyond exposure to female role models. Drawing on recent research, we define sense of fit in science as a multifaceted construct that refers to a person’s overall sense that they belong in and can succeed in science (Gilbert, O’Brien, Garcia, & Marx, 2015; Walton & Cohen, 2007).

Sense of fit is critically important to success in STEM fields. Improving feelings of fit and belonging can increase interest in pursuing careers in STEM and math grades (Cheryan, Plaut, Davies, & Steele, 2009; Good, Rattan, & Dweck, 2012). While previous research has primarily examined sense of fit in college-aged samples, the present research sought to expand research on sense of fit to middle school girls. Female role models may increase girls’ sense of fit in science by providing tangible evidence that women can succeed in science. The present study examines the impact of a brief intervention that encourages girls to choose a female role model after participating in a science outreach event.

Early adolescence may be an ideal time to intervene and encourage girls to aspire to careers in the sciences (Betz & Sekaquaptewa, 2012; Plant, Baylor, Doerr, & Rosenberg-Kima, 2009). By the time they reach college, many women have already decided not to pursue careers in math-intensive STEM fields (Ceci et al., 2014). Science attitudes become more negative between seventh and ninth grades, and this decline is steeper for girls than boys (Barmby, Kind, & Jones, 2008). Despite having similar levels of achievement, girls as young as the fourth grade have weaker math and science self-concepts than boys (Kurtz-Costes, Rowley, Harris-Britt, & Woods, 2008). In addition, cultural stereotypes begin to exert greater influence on achievement, motivation, and performance during adolescence (Ambady, Shih, Kim, & Pittinsky, 2001; Eccles, Jacobs, & Harold, 1990; Kurtz-Costes, Copping, Rowley, & Kinlaw, 2014; Muzzatti & Agnoli, 2007). Thus, middle school may be an opportune time to intervene to improve girls’ sense of fit in science.

The Impact of Female STEM Role Models

Role models are an effective mechanism to improve women’s and girls’ outcomes in STEM. Exposure to competent female role models can improve college women’s performance on math tests, STEM identification, feelings of self-efficacy, and commitment to pursue STEM careers (Cheryan, Siy, Vichayapai, Drury, & Kim, 2011; Marx & Roman, 2002; Stout, Dasgupta, Hunsinger, & McManus, 2011). Role models have a positive impact on girls as well. For example, fifth-grade girls exposed to a female role model performed better on a math test than girls exposed to a male role model (Bages & Martinot, 2011). More relevant to the current research, participating in science outreach programs featuring female role models increases middle school girls’ sense of self-efficacy and interest in science (Liben & Coyle, 2014; Weisgram & Bigler, 2006).

In the present study, we examine whether choosing and briefly writing about a favorite female role model increases girls’ sense of fit in science above and beyond exposure to female role models. We test our hypothesis in the context of a science outreach program that exposes girls to multiple female scientists. There are at least two reasons to believe choosing a role model may be especially likely to benefit girls. First, not all role models are equally effective at eliciting positive changes. For example, sixth- and seventh-grade girls who were exposed to highly feminine female STEM role models were less interested in math than girls who were exposed to more gender-neutral female STEM role models (Betz & Sekaquaptewa, 2012). Role models who are perceived as “nerdy” can also reduce college women’s interest in pursuing science, although we are unaware of research on the impact of nerdy role models on younger girls (Cheryan et al., 2011; Rosenberg-Kim, Baylor, Plant, & Doerr, 2008). Encouraging girls to choose and write about a favorite role model may increase the chance that girls will select a role model who embodies qualities that they admire.

Second, believing that one has made a free choice is a powerful motivator of attitude change (e.g., Cooper, 2007). The act of “choosing” and writing about a favorite female role model may spark exploration and increase subjective identification with the role model. Subjective identification can increase feelings of belonging, self-efficacy, commitment to pursuing STEM careers, and a host of other positive outcomes (Dasgupta, 2011). Thus, by cueing girls in the science outreach program to “choose” a favorite female role model, we may be able to increase their sense of fit in science.

Overview

In the present study, we recruited middle school girls attending the Girls in STEM at Tulane (GiST) program, a 1-day scientific outreach program modeled on the popular Expanding Your Horizons (EYH) program (EYH Network, 2016; Liben & Coyle, 2014; Weisgram & Bigler, 2006; Yanowitz & Vanderpool, 2004). EYH conferences are typically 1-day events held on a university campus where middle school girls attend various workshops led by female scientists. The workshops are designed around different topics to provide broad exposure to content areas in science. Participating in EYH conferences increases girls’ science self-efficacy and interest in science (Weisgram & Bigler, 2006). As with any intervention, there are undoubtedly multiple aspects of EYH conferences that make them effective at increasing girls’ interest in science. However, one aspect that may make these types of conferences particularly effective is that girls are typically exposed to multiple potential role models, maximizing the likelihood that girls will find a role model that inspires them and with whom they can identify.

In the present study, we recruited girls in the fifth through eighth grades who were participating in the 1-day GiST outreach program where they attended three consecutive science workshops. Women who were advanced students, postdoctoral researchers, or faculty working at a local university led the workshops. Immediately prior to participating in GiST, the girls completed a measure of sense of fit in science. At the end of the event, the girls were randomly assigned to complete a writing exercise in which they either (1) identified their favorite workshop leader and wrote about why they chose her (role model choice condition) or (2) identified their best friend and wrote about why they chose her or him (control condition). After completing the writing exercise at the end of the event, the girls completed the sense of fit in science measure again and also completed a measure of role model identification.

We predicted that the role model choice manipulation would interact with time to impact sense of fit in science such that the increase in sense of fit from pre- to postparticipation would be bigger for participants in the role model choice condition as compared to the control condition. We also predicted that participants would have higher role model identification in the role model choice condition than the control condition. In addition, we conducted an exploratory examination of role model characteristics to determine whether some types of role models are more appealing to middle school girls. For example, high expertise role models may be more dissimilar to the girls either because they are older or their high status seems personally unattainable (e.g., Marx & Ko, 2012). Therefore, we examined natural variation in role model expertise and whether it was related to girls’ role model identification and sense of fit. Finally, we also coded participants’ open-ended responses to the writing prompt to examine the role model characteristics that are most salient to girls at this age. Although the analyses of role model expertise and the open-ended responses were exploratory, we hoped they might shed light on avenues for future research and interventions to improve girls’ STEM outcomes.

Method

Participants

One hundred and seventy-five girls (M _age = 11.77, SD = 1.06) from the New Orleans metropolitan area participated in the study while attending GiST, a 1-day science outreach event. Age was not correlated with any of the study variables and therefore is not discussed further. Although we did not collect data on participants’ ethnicity, participants were from over 40 local public and private schools that varied widely in their ethnic composition. GiST participants were recruited through numerous methods including disseminating information to local schools, radio announcements, Internet and social media, and word of mouth. The sample size was predetermined by the number of girls who participated in the outreach event, and therefore, we did not compute an a priori power analysis. Data were collected on two occasions (Spring 2014 and Fall 2014) in order to increase the sample size; there were no differences in any of the study variables across the two occasions. Eleven girls participated on both occasions; the responses of these 11 girls on their second occasion of participation were excluded from analyses. Degrees of freedom vary slightly across analyses due to missing data points.

Procedure

Consent forms were distributed to parents along with registration materials and were collected both in advance and on the day of the event. After participants arrived and registered for the event, they assembled for a short orientation. At the beginning of the orientation session, an experimenter briefly introduced the study, explained how to complete the questionnaires, and then administered the pretest measure of sense of fit in science.

After the orientation, girls were divided into groups and participated in three sequential science workshops. Workshop leaders were recruited through e-mails distributed to faculty, staff, and students in the Tulane University School of Science and Engineering and through word of mouth. Thirty percent of workshops were led by a leader with a PhD (i.e., a faculty member or postdoc), whereas 70% of the workshops were led by a leader without a PhD. (i.e., an advanced student or staff member). All workshop leaders were volunteers who designed their own age appropriate workshop in their area of expertise. Workshops included topics such as Using Your Brain (led by a neuroscientist); Fluids, Bubbles, and Slime (led by a physics graduate student); and Monkeys, Apes, and You (led by a biological anthropology graduate student). See Table 1 for brief descriptions of all workshops. Each workshop lasted approximately 1 hr and workshop leaders attempted to make the content fun and accessible, creating an informal learning environment. Workshop leaders were also asked to tell the girls how they came to be interested in science, what classes they took in high school to prepare them for their field, and what was their major as an undergraduate.

Table 1.

Workshop Descriptions.

Title	Description
Fluids, Bubbles, and Slime	Learn about the science of fluids. You will learn how to make a chunk of iron float in water and make a soap powered boat. We will prepare a bowl of liquid slime that becomes rock hard when you punch it!
Tiny Crabs, Big Ocean!	In this workshop, we will explore the effects of an oil spill on the food web of the Gulf of Mexico. You will also learn about the fascinating life cycle of the blue crab and look at baby crabs before they even look like crabs!
Using Your Brains	In this workshop, you will become brain scientists and discover how our brains allow us to feel, think, and remember. You will touch real brains and conduct experiments to learn how our nervous system works.
A Little Birdie Told Me	Come be a scientist-in-training: build your own local bird field guide, make your own sound recordings, and learn what it’s like to be a bird in the city.
Creative Computation	Have you ever wondered what goes into designing a computer game? Girls attending this workshop will learn basic computer programming concepts while creating an interactive talking character using Scratch.
The Missing Tooth	In this workshop, we will exploring a sample of 10-million-year-old shark teeth from Gainesville, FL. We will take a virtual field trip to the site of collection before playing with and sorting through a real-life sample!
Monkeys, Apes, and You	Fascinated by the monkeys and apes that you see at the Audubon Zoo? Ever wonder why they behave so much like your brother or sister? Come spend some time with us and channel your inner ape!
The Science of Bread	In this workshop, we will use air balloons to demonstrate how gas is released as a result of the process of fermentation. Finally, we will look at yeast that contains a glowing molecule under the microscope!
The Secret World of Fungi	Fungi are everywhere! Join us in looking at fungi in crazy places and grow your own pet fungi. Help us make new discoveries about the fungi on Tulane’s campus.
Follow the Signs!	Conformity includes wearing the latest fashions, following jaywalkers at a crosswalk, and only using restrooms designated for our own gender. We will conduct an experiment to observe conformity in action!
Beat The Heat	Join us as we explore how cold-blooded animals (ectotherms) manage to keep cool in a changing climate! We will experiment with gummy frogs to explore how ectotherms’ regulate body temperature.
Doctor, What Do I Have?	Be a doctor for a day and diagnose their patients’ kidney conditions. We will emphasize the importance of healthy choices to keep one’s kidneys functioning well for as long as possible.
Our Diverse Planet	Ever wondered what your house looked like from space? In this workshop, we’ll learn about Google Earth and how geologists use satellite images to learn about our world.
Tissue Engineering	Learn about the tissues in your body and how the properties of your tissues affect your body’s functions. In one of our lessons, we will look at the similarities between baking cookies and tissue engineering—yum!
Wilderness Explorers	Learn what it’s like to be a wildlife biologist! Join us on a scavenger hunt using global positioning system and learn how to radio track as we search for treasure around the Tulane campus.

After attending three workshops, participants attended a plenary session where an experimenter introduced the second part of the study. Research assistants distributed packets to participants containing the manipulation and study measures. Participants were randomly assigned to one of the two conditions. In the role model choice condition, participants were instructed to “Take a moment to think about your favorite workshop leader. Please tell us why she is your favorite leader. Please write a few sentences about why you picked her.” In the control condition, participants were instructed to “Take a moment to think about your favorite friend. Please tell us why she or he is your favorite friend. Please write a few sentences about why you picked her or him.” Once participants completed the manipulation, they completed measures of sense of fit in science and role model identification.¹ Once all participants completed the packet, they received a thorough debriefing that explained the goals of the study.

Measures

Sense of fit in science

Participants’ sense of fit in science was assessed with a composite measure consisting of two scales: science identification and sense of social fit in science (see Gilbert et al., 2015; Walton & Cohen, 2007). Science identification was assessed with 8 items adapted from the Brown and Josephs’s (1999) math identification questionnaire (e.g., “How important is it to you to do well in science?” and “I’m bad when it comes to science”). Sense of social fit in science was measured with 12 items adapted from Walton and Cohen’s (2007) sense of social fit in computer science measure (e.g., “I feel like I belong in science classes” and “I feel left out in science class”). Participants responded to all items on a Likert-type scale ranging from 1 to 7 where higher numbers indicated stronger endorsement. Science identification and sense of social fit in science were highly correlated at both pretest, r = .64, p < .001, and posttest, r = .66, p < .001, and were aggregated into a composite sense of fit in science variable consistent with past research (Gilbert et al., 2015; Walton & Cohen, 2007). The composite measure was reliable at both pretest (α = .88) and posttest (α = .91). Pretest sense of fit in science was highly correlated with posttest sense of fit in science, r = .83, p < .001.

Role model identification

During the posttest, participants completed a 4-item measure of role model identification (this measure was not completed at the pretest because participants had not yet met the role models). Before completing the measure of role model identification, all participants were asked to select their favorite workshop leader. Sample items include “How much are you similar to the workshop leader you chose?” and “How much do you look up to the workshop leader you chose?” Participants responded to all items on a Likert-type scale ranging from 1 to 7 (α = .82).

Role model expertise

We operationalized role model expertise as whether or not role models had a PhD. We were unable to determine the identity of the selected role models for 13 participants due to the nature of their descriptions (e.g., The reason she is my favorite workshop leader is because we have the same interests. Even though my goal is to become a physician, biology seems fun too!…I loved the way she talked so passionately about what she enjoys doing most). These individuals were omitted from analyses involving role model expertise.

Coding

We coded participants’ responses to the writing prompt with several categories theoretically derived from the role model literature. For example, similarity to a role model and role model competence can be critical for eliciting positive outcomes (e.g., Cheryan et al., 2011; Marx & Ko, 2012), and therefore, we coded for whether girls spontaneously mentioned similarity and competence. Moreover, girls typically view role models positively (Betz & Sekaquaptewa, 2012), and therefore, we also coded for whether participants expressed warmth and admiration in their description of the target. Because feeling supported and encouraged can be an important aspect of a role model relationship (Stanton-Salazar & Spina, 2003), we also coded for whether participants mentioned feeling supported.

In addition, we coded for general references to personality to examine whether girls in the role model condition were describing a person or an activity. We also coded for whether participants mentioned a science activity to rule out the possibility that girls in the control condition were writing about a friend who was a science role model. Finally, we also included two additional categories based on themes that emerged in the data. Many girls spontaneously mentioned that they enjoyed the activities and learned something new, so we added these two categories. See Table 2 for examples of each category.

Table 2.

Examples of Coding Categories.

Category	Examples of Responses Coded as “Yes”
Similarity to target	RM: My favorite workshop leader was Mrs. NAME. I like her because we both are interested in the same thing which was earth science Friend: My best friend’s name is NAME and I like her because she is just like me
Target’s competence	RM: NAME was my favorite workshop leader because she gave clear directions and know (sic) a lot about what she was teaching. Although all the workshops were fun, I learned the most in NAME’s Friend: My favorite friend is NAME. I chose her because she is in advanced classes. She is really good in science. She helps me with my science homework and a few projects
Target’s warmth	RM: She is my favorite leader because she is great and she is very nice. Also she is a sweet, kindhearted women Friend: NAME is my favorite friend because she is nice and sweet. She is also very creative
Admiration for target	RM: She told us one of her obstacles. The obstacle was being the only girl doing something and boys judging her. I can really relate to that. She inspired me to not be influenced by being the only girl Friend: My favorite friend is my favorite friend because he is just amazing. He is always there for me all the time. He encourages me in everything I do all the time. He is like my big brother. I always thank God he’s in my life. I love him
Supported by target	RM: She wasn’t afraid to be herself. I loved that she was willing to help me with scratch program. She even gave the website so I can do it at home (I also got a sheet about a computer program) Friend: She is my best friend because she is funny. She is also my best friend because she will always be there when I need her
Target’s personality	RM: I really liked the workshop leader at the birdwatching workshop. I chose her because she was nice and helped me have an enjoyable learning experience Friend: NAME because she is funny and nice. I always laugh at the things she says because they’re really funny
Science activity	RM: I liked this teacher because she kept us occupied with fascinating experiments. We mixed oil, water, and Dawn (dish soap) and we had an amazing outcome! Friend: My favorite friend is NAME. I chose her because she is in advanced classes. She is really good in science. She helps me with my science homework and a few projects
Enjoyment of activity	RM: My favorite workshop leader was the girl who was working with illusions and the concentration of the brain. She’s my favorite because I learned about the sensitivity of areas on the body. It was fun learning with her! Friend: My friend is my favorite because we both share a lot of the same interests, but we balance each other out when times get rough, I picked this one friend because she has known me since we were infants. We both like to do what is listed which includes dancing, singing, acting, drawing, and just having fun!
Learning	RM: Doctor NAME was my favorite because the way she explained kidneys was very easy to understand. Doctor NAME is a doctor and that is one of the career choseses [sic] I might consider. I felt her class was very informational Friend: NAME’s my favorite friend because we study together, we share stuff, I respect her and she respect me, and she is a nice and friendly friend

Note. “RM” denotes a sample response from the role model choice condition, whereas “Friend” denotes a response from the control condition. Coded categories are not mutually exclusive; for example, the same response may be coded as both referencing target personality and target warmth if appropriate.

One research assistant coded all written responses, and a second research assistant coded 27% of the responses. Interrater reliability ranged from κ = .70 to .97. We report the results for the primary coder.

Results

Sense of Fit

In order to examine the effect of the manipulation on changes in participants’ sense of fit, we estimated a linear mixed model to account for within-participant repeated measures. Time (0 = pretest, 1 = posttest), the manipulation (−1 = control condition, 1 = role model condition), and the time by manipulation interaction were designated as fixed effects. This analysis yielded a significant main effect of time, B = .13 (SE = .03), t(173) = 3.73, p < .001, 95% confidence interval [CI] = [.06, .19], d = .16.² The main effect of the manipulation was not significant, B = .01 (SE = .06), t(173) = .18, p = .85, 95% CI [−.10, .12], d = .03; however, there was a significant time by manipulation interaction, B = .07 (SE = .03), t(173) = 2.18, p = .03, 95% CI [.01, .14], d = .33 (see Table 3).

Table 3.

Means and Standard Deviations for Study Variables.

	Control (Friend)			Role Model Choice
Measure	Mean	SD	95% CI	Mean	SD	95% CI
Sense of fit (pretest)	5.69	.76	[5.54, 5.85]	5.72	0.75	[5.56, 5.88]
Sense of fit (posttest)	5.75	.88	[5.58, 5.91]	5.91	0.69	[5.75, 6.08]
Role model identification (posttest)	5.74	.75	[5.55, 5.92]	5.97	1.00	[5.78, 6.16]

Note. The potential range for all study variables is 1–7. CI = confidence interval.

We conducted simple effects tests to decompose the significant interaction. For participants in the role model choice condition, sense of fit increased significantly from the pretest to the posttest, B = .20 (SE = .05), t(173) = 4.14, p < .001, 95% CI [.10, .29], d = .20. However, for participants in the control condition, sense of fit did not change from the pretest to the posttest, B = .05 (SE = .05), t(173) = 1.100, p = .27, 95% CI [−.04, .15], d = .04.

Role Model Identification

Participants in the role model choice condition tended to have higher levels of role model identification than participants in the control condition; however, this difference was only marginally significant, t(171) = 1.76, p = .08, 95% CI [−.03, .50], d = .27. Role model identification was correlated with both the pretest measure of sense of fit, r = .29, p < .001, and the posttest measure of sense of fit, r = .34, p < .001.

Role Model Expertise

Thirty percent of workshops were led by role models with PhDs, and 25% of girls selected a role model with a PhD as their favorite, χ² < 1. This finding suggests that role model expertise did not influence girls’ choice of role models. The expertise of the selected role model was unrelated to girls’ role model identity, r = −.09, p = .44, and sense of fit in science, r = .04, p = .76.

Open-Ended Responses

We compared the role model choice and control conditions for each of the coded categories (see Table 4). Despite the fact that similarity to a role model has been identified as an important attribute of effective role models in the adult literature, very few girls spontaneously mentioned their similarity to the role model. In comparison, a relatively larger percentage of girls in the control condition mentioned similarity to their friend. A substantial minority of girls wrote about the role model’s competence, whereas comparatively few girls wrote about their friend’s competence. A sizeable minority of girls also mentioned feeling supported by the role model and described the role model’s warmth; however, perhaps not surprisingly, an even larger percentage of girls mentioned feeling supported by their friend and described their friend’s warmth. Almost half of the girls in the role model condition mentioned the role model’s personality and even more girls in the control condition mentioned their friend’s personality. Relatively few girls directly expressed admiration for either the role model or their friend. Whereas the vast majority of girls in the role model condition described a science activity, almost no one who wrote about a friend focused on a science activity. Finally, girls in the role model condition focused on enjoyment of an activity and learning at much higher rates than girls in the control condition.

Table 4.

Open-Ended Responses in the Role Model Choice and Control (Friend) Conditions.

	RM Choice		Control (Friend)		χ²
Category	%	95% CI	%	95% CI	χ²
Similarity to target	4.7	[.01, 9.2]	31.5	[21.6, 41.3]	21.0*
Target’s competence	31.4	[21.4, 41.4]	10.1	[3.7, 16.5]	12.1*
Supported by target	29.1	[19.3, 38.9]	60.7	[50.3, 71.0]	17.6*
Admiration for target	5.8	[.01, 10.9]	1.1	[−.01, 3.4]	**
Target’s warmth	26.7	[17.2, 36.3]	49.4	[38.8, 60.0]	9.5*
Target’s personality	44.2	[33.5, 54.9]	71.9	[62.4, 81.4]	13.8*
Science activity	65.1	[54.8, 75.4]	2.2	[−.01, 5.4]	78.0*
Enjoyment of an activity	41.9	[31.2, 52.5]	9.0	[2.9, 15.0]	25.1*
Learning	54.7	[43.9, 65.4]	2.2	[−.01, 5.4]	59.6*

Note. The table indicates the percentage of participants in each condition who wrote open-ended responses that were categorized in each category. RM = role model; CI = confidence interval.

*χ² significant at p < .01. **Two cells had expected cell counts <5 and therefore we did not report a χ² value.

Discussion

The opportunity to choose and write about a favorite role model improved girls’ science outcomes in an outreach program where they were exposed to multiple female role models. Girls who attended a science outreach event where they participated in consecutive workshops led by female scientists showed a significant increase in sense of fit in science when they were encouraged to choose and briefly write about a favorite workshop leader. In comparison, girls who participated in the outreach event, but who were not encouraged to choose and write about a favorite workshop leader, did not experience an increase in sense of fit.

The present study builds upon a growing body of research on the efficacy of exposure to role models for improving girls’ STEM outcomes (e.g., Betz & Sekaquaptewa, 2012; Liben & Coyle, 2014; Plant et al., 2009; Weisgram & Bigler, 2007). In the present research, girls were exposed to multiple female role models, offering them a chance to observe the diversity that exists among female scientists. In this context, choosing and writing about a favorite role model may have a positive impact above and beyond any effects associated with being exposed to multiple female role models. It is worth noting that girls in the present study were asked to choose their favorite “workshop leader” and were not specifically asked to choose a “role model.” This suggests that it may not be essential to apply the role model label to an individual in order to experience benefits.

The manipulation may have prompted girls to further explore their favorability toward their chosen role model. Exploration is an important developmental stage in identification that often occurs around this age (Phinney & Ong, 2007). Girls who chose and wrote about a favorite role model tended to have higher levels of role model identification; however, this effect did not quite achieve statistical significance. In general, role model identification was quite high, raising the possibility of ceiling effects on this measure. The relatively high scores on role model identification may have been due to the fact that all girls spent the day in a conference focused on women in science. Role model identification was positively related to sense of fit in science, consistent with our argument that role models play an important role in making girls feel like they belong in science.

In order to identify the characteristics of role models that are salient to middle school girls, we coded participants’ open-ended written responses. Compared to the control condition, girls who wrote about a role model were significantly more likely to focus on the role model’s competence. This finding complements the existing literature on the importance of role model competence for eliciting positive STEM outcomes (Marx & Roman, 2002). A substantial minority of girls in the role model condition also spontaneously mentioned feeling supported by the role model and commented on the role model’s warmth. The percentage of girls who wrote about their best friend’s support and warmth was even greater; however, this finding may reflect the importance of peer relations during adolescence.

Surprisingly, very few girls in the role model condition spontaneously wrote about their similarity to the role model. Past research suggests similarity to role models is important; however, this research examined college women (e.g., Cheryan et al., 2011; Marx & Ko, 2012). Interestingly, Betz and Sekaquaptewa (2012) found that middle school girls’ ratings of similarity to adult role models were rather modest despite overall positivity toward role models. Middle school girls may have difficulty seeing similarities between themselves and adult female role models. Future research should address whether the characteristics of effective role models are different for college women compared to younger girls. In the present study, many girls spontaneously wrote about how much they learned from the role model and how much they enjoyed the activities that the role model planned. Thus, role models who engender enthusiasm for learning and science may be especially important for middle school girls.

Strengths, Limitations, and Future Directions

One strength of the present study is that it demonstrated the impact of the role model choice manipulation in a field setting. The experiment occurred in the context of a meaningful, real-world event that sought to have a positive impact on the girls who participated. The present study provides evidence that the manipulation is robust enough to have an impact outside of a more constrained laboratory environment. Although we view the field nature of the present research as a strength, it is not without limitations. Because the goals of GiST are to provide an enrichment experience for as many girls as possible, it was not feasible to include an additional control condition in which a group of girls were randomly assigned to abstain from the event. Thus, the present study is not able to compare girls who participated in the outreach event with girls who did not participate to determine the impact of participation on sense of fit.

Another limitation of the field setting is that, because participants were aware that they were participating in a science outreach event designed for girls, they may have felt social desirability pressures when completing the study measures. We took precautions to reduce social desirability pressures by having research assistants who were not part of the outreach event itself and who were not wearing event T-shirts administer the study. In the future, researchers may be able to further reduce social desirability concerns by administering the study outside the context of the event—however, such procedures may lower participation rates.

Girls who participate in a science outreach event may differ from a randomly selected sample of middle school girls. Although some of the girls came with larger groups from their schools, other girls more actively sought out participation. It is not clear whether the benefit of identifying a female scientist as a role model on sense of fit in science would generalize to other samples.

To our knowledge, the present study is the first to examine sense of fit in science among middle school girls. To measure sense of fit in science among middle school girls, we adapted items used in previous research with college students (e.g., Gilbert et al., 2015; Walton & Cohen, 2007). For example, items that asked college students about the extent to which they felt like they belong in the computer science department were modified to ask girls about the extent to which they belong in science classes (Walton & Cohen, 2007). Although it may be surprising that choosing a female role model at a science outreach event would increase the extent to which girls feel they belong in their science classes, one interpretation is that the manipulation creates a general sense of “fitting in” that extends beyond the context of the outreach event. Research on the long-term effects of both participating in science outreach events featuring female role models and choosing favorite role models is needed to examine whether these effects truly carry over into other science contexts.

The present study was unable to rigorously examine whether the individual characteristics of the role model that the girls chose affected the results. With 86 girls in the role model choice condition and 15 different possible role models, we lacked statistical power to thoroughly examine individual role model characteristics. However, we were able to conduct exploratory analyses of the effects of role model expertise. There was no evidence that girls were more or less likely to choose role models with or without PhD. Furthermore, the expertise of girls’ chosen role model was unrelated to their role model identification and sense of fit in science. A thorough examination of the individual characteristics of effective role models is an important area for future research.

Conclusion

Middle school is an ideal time to intervene to increase girls’ interest in STEM fields, as this is a time when they may develop more negative attitudes toward math and science (e.g., Barmby et al., 2008). Many organizations have started outreach programs targeted at middle school girls to spark their interest in science and pursuing careers in STEM (e.g., Demetry et al., 2009; Laursen, Liston, Thiry, & Graf, 2007; Yanowitz & Vanderpool, 2004). These outreach programs offer girls exposure to multiple female role models and provide tangible evidence of the diversity among women in science. The present research suggests that encouraging girls to choose and write about a favorite role model, thus prompting them to examine and explore the reasons they favor a role model, may help to increase girls’ sense of fit in science.

Footnotes

Acknowledgments

The authors would like to thank the numerous individuals who contribute to the Girls in STEM at Tulane (GiST) outreach event. The authors would also like to thank several excellent undergraduate research assistants for their contributions to data collection.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was partially supported by a grant from the National Science Foundation: HRD0936722.

Notes

References

Ambady

Shih

Kim

Pittinsky

T. L.

(2001). Stereotype susceptibility in children: Effects of identity activation on quantitative performance. Psychological Science, 12, 385–390. doi:10.1111/1467-9280.00371a

Bages

Martinot

(2011). What is the best model for girls and boys faced with a standardized mathematics evaluation situation: A hardworking role model or a gifted role model? British Journal of Social Psychology, 50, 536–543. doi:10.1111/j.2044-8309.2010.02017.x

Barmby

Kind

P. M.

Jones

(2008). Examining changing attitudes in secondary school science. International Journal of Science Education, 30, 1075–1093. doi:10.1080/09500690701344966

Betz

D. E.

Sekaquaptewa

(2012). My fair physicist? Feminine math and science role models demotivate young girls. Social Psychological and Personality Science, 3, 738–746. doi:10.1177/1948550612440735

Brown

R. P.

Josephs

R. A.

(1999). A burden of proof: Stereotype relevance and gender differences in math performance. Journal of Personality and Social Psychology, 76, 246–257. doi:10.1037/0022-3514.76.2.246

Ceci

S. J.

Ginther

D. K.

Kahn

Williams

W. M.

(2014). Women in academic science a changing landscape. Psychological Science in the Public Interest, 15, 75–141. doi:10.1177/1529100614541236

Cheryan

Plaut

V. C.

Davies

P. G.

Steele

C. M.

(2009). Ambient belonging: How stereotypical cues impact gender participation in computer science. Journal of Personality and Social Psychology, 97, 1045–1060. doi:10.1037/a0016239

Cheryan

Siy

Vichayapai

Drury

B. J.

Kim

(2011). Do female and male role models who embody STEM stereotypes hinder women’s anticipated success in STEM? Social Psychological and Personality Science, 2, 656–664. doi:10.1177/1948550611405218

Cooper

(2007). Cognitive dissonance: 50 years of a classic theory. New York, NY: Sage.

10.

Dasgupta

(2011). Ingroup experts and peers as social vaccines who inoculate the self-concept: The stereotype inoculation model. Psychological Inquiry, 22, 231–246. doi:10.1080/1047840X.2011.607313

11.

Demetry

Hubelbank

Blaisdell

S. L.

Sontgerath

Nicholson

M. E.

Rosenthal

Quinn

(2009). Supporting young women to enter engineering: Long-term effects of a middle school engineering outreach program for girls. Journal of Women and Minorities in Science and Engineering, 15, 119–142.

12.

Dunlop

W. P.

Cortina

J. M.

Vaslow

J. B.

Burke

M. J.

(1996). Meta-analysis of experiments with matched groups or repeated measures designs. Psychological Methods, 1, 170–177.

13.

Eccles

J. S.

(2011). Gendered educational and occupational choices: Applying the Eccles et al. model of achievement-related choices. International Journal of Behavioral Development, 35, 195–201. doi:10.1177/0165025411398185

14.

Eccles

J. S.

Jacobs

J. E.

Harold

R. D.

(1990). Gender role stereotypes, expectancy effects, and parents’ socialization of gender differences. Journal of Social Issues, 46, 183–201. doi:10.1111/j.1540-4560.1990.tb01929.x

15.

Expanding Your Horizons Network. (2016). Expanding Your Horizons Network. Retrieved May 2, 2016, from http://www.eyhn.org/

16.

Gilbert

P. N.

O’Brien

L. T.

Garcia

D. M.

Marx

D. M.

(2015). Not the sum of its parts: Decomposing implicit academic stereotypes to understand sense of fit in Math and English. Sex Roles, 72, 25–39. doi:10.1007/s11199-014-0428-y

17.

Good

Rattan

Dweck

C. S.

(2012). Why do women opt out? Sense of belonging and women’s representation in mathematics. Journal of Personality and Social Psychology, 102, 700–717. doi:10.1037/a0026659

18.

Hill

Corbett

St. Rose

(2010). Why so few? Women in science, technology, engineering, and mathematics (American Association of University Women Report). Washington, DC: AAUW.

19.

Kurtz-Costes

Copping

K. E.

Rowley

S. J.

Kinlaw

C. R.

(2014). Gender and age differences in awareness and endorsement of gender stereotypes about academic abilities. European Journal of Psychology of Education, 29, 603–618.

20.

Kurtz-Costes

Rowley

S. J.

Harris-Britt

Woods

T. A.

(2008). Gender stereotypes about mathematics and science and self-perceptions of ability in late childhood and early adolescence. Merrill-Palmer Quarterly, 54, 386–409. doi:10.1353/mpq.0.0001

21.

Laursen

Liston

Thiry

Graf

(2007). What good is a scientist in the classroom? Participant outcomes and program design features for a short-duration science outreach intervention in K–12 classrooms. CBE-Life Sciences Education, 6, 49–64.

22.

Liben

L. S.

Coyle

E. F.

(2014). Chapter three-developmental interventions to address the STEM gender gap: Exploring intended and unintended consequences. Advances in Child Development and Behavior, 47, 77–115.

23.

Marx

D. M.

(2012). Superstars ‘like’ me: The effect of role model similarity on performance under threat. European Journal of Social Psychology, 42, 807–812. doi:10.1002/ejsp.1907

24.

Marx

D. M.

Roman

J. S.

(2002). Female role models: Protecting women’s math test performance. Personality and Social Psychology Bulletin, 28, 1183–1193. doi:10.1177/01461672022812004

25.

Milkman

K. L.

Akinola

Chugh

(2015). What happens before? A field experiment exploring how pay and representation differentially shape bias on the pathway into organizations. Journal of Applied Psychology,100, 1678–1712.

26.

Moss-Racusin

C. A.

Dovidio

J. F.

Brescoll

V. L.

Graham

M. J.

Handelsman

(2012). Science faculty’s subtle gender biases favor male students. Proceedings of the National Academy of Sciences, 109, 16474–16479.

27.

Muzzatti

Agnoli

(2007). Gender and mathematics: Attitudes and stereotype threat susceptibility in Italian children. Developmental Psychology, 43, 747–759. doi:10.1037/0012-1649.43.3.747

28.

Nosek

B. A.

Smyth

F. L.

(2011). Implicit social cognitions predict sex differences in math engagement and achievement. American Educational Research Journal, 48, 1125–1156. doi:10.3102/0002831211410683

29.

Phinney

J. S.

Ong

A. D.

(2007). Conceptualization and measurement of ethnic identity: Current status and future directions. Journal of Counseling Psychology, 54, 271–281. doi:10.1037/0022-0167.54.3.271

30.

Plant

Baylor

A. L.

Doerr

C. E.

Rosenberg-Kima

R. B.

(2009). Changing middle-school students’ attitudes and performance regarding engineering with computer-based social models. Computers & Education, 53, 209–215. doi:10.1016/j.compedu.2009.01.013

31.

Rosenberg-Kima

R. B.

Baylor

A. L.

Plant

E. A.

Doerr

C. E.

(2008). Interface agents as social models for female students: The effects of agent visual presence and appearance on female students’ attitudes and beliefs. Computers in Human Behavior, 24, 2741–2756. doi:10.1016/j.chb.2008.03.017

32.

Stanton-Salazar

R. D.

Spina

S. U.

(2003). Informal mentors and role models in the lives of urban Mexican-origin adolescents. Anthropology & Education Quarterly, 34, 231–254.

33.

Stout

J. G.

Dasgupta

Hunsinger

McManus

M. A.

(2011). STEMing the tide: Using ingroup experts to inoculate women’s self-concept in science, technology, engineering, and mathematics (STEM). Journal of Personality and Social Psychology, 100, 255–270. doi:10.1037/a0021385

34.

Walton

G. M.

Cohen

G. L.

(2007). A question of belonging: Race, social fit, and achievement. Journal of Personality and Social Psychology, 92, 82–96.

35.

Weisgram

E. S.

Bigler

R. S.

(2006). Girls and science careers: The role of altruistic values and attitudes about scientific tasks. Journal of Applied Developmental Psychology, 27, 326–348.

36.

Weisgram

E. S.

Bigler

R. S.

(2007). Effects of learning about gender discrimination on girls’ attitudes toward and interest in science. Psychology of Women Quarterly, 31, 262–269.

37.

Williams

J. C.

Phillips

K. W.

Hall

E. V.

(2014). Double jeopardy? Gender bias against women of color in science. Retrieved September 15, 2016, from http://www.toolsforchangeinstem.org/tools/double-jeopardy-report

38.

Yanowitz

L. K.

Vanderpool

S. S.

(2004). Assessing girls’ reaction to science workshops. Journal of Science Education and Technology, 3, 353–359.