Mixed-methods exploration of students’ written belonging explanations from general chemistry at a selective institution
Angela
Fink
a,
Jessica D.
Young
b,
Neil K.
Vuppala
c and
Regina F.
Frey
*d
aCenter for Integrative Research on Cognition, Learning, and Education, Washington University in St. Louis, St. Louis, MO 63130, USA
bDepartment of Chemistry, University of South Florida, Tampa, FL 33620, USA
cAlabama College of Osteopathic Medicine, Dothan, AL 36303, USA
dDepartment of Chemistry, University of Utah, Salt Lake City, UT 84112, USA. E-mail: gina.frey@utah.edu
First published on 19th October 2022
Abstract
This exploratory, mixed-methods study examines first-year general chemistry students' written responses on a belonging survey. Responses were thematically analyzed to identify students’ sources of belonging, which may help instructors choose effective strategies for enhancing belonging during the transition into college. Qualitative analysis generated a codebook containing 21 codes from 6 categories: Course Attributes, Interest, Perceptions, Social, Student Attributes, and Value. The qualitative coding data were transformed into quantitative frequency data, allowing identification of the most frequent themes across all participants on each of four surveys: early- and late-semester General Chemistry 1 and 2. Additional analyses explored how belonging explanations varied based on student characteristics that might influence their experience of this large introductory STEM course at a selective, high-income, predominantly White institution. Unique sources of belonging were expected to emerge for groups marginalized in STEM (i.e., Black and Hispanic students, women) and groups who might feel discouraged by a selective institutional and course culture (i.e., students with no credit-bearing AP scores, low course grades, or high belonging uncertainty). Results indicate the importance of interest for all participants' course-level belonging. Students' career goals, perceptions of the course content, and social dynamics with peers also proved universally influential. Some patterns were especially pronounced for marginalized or discouraged groups, who were disproportionately likely to discuss social comparisons and interactions, self-evaluate, and describe the utility-value of the course. These groups were also less likely to express positive cognitive and affective engagement in the course. Implications for supporting student belonging throughout the course sequence are discussed.
Mixed-methods exploration of general chemistry belonging at a selective institution
Feeling a sense of belonging – connectedness, inclusion, and relationship with others – is important for well-being (Baumeister and Leary, 1995; Gere and MacDonald, 2010) and predictive of academic success (Cohen and Garcia, 2008; Walton et al., 2012). Research shows the positive impact of institutional belonging on undergraduate outcomes like retention and achievement in college (O'Keeffe, 2013; Yorke, 2016; Slaten et al., 2018) and associations between institutional belonging and student perceptions, motivation, and engagement in class (Freeman et al., 2007; Zumbrunn et al., 2014). Recent research has extended these findings to belonging in STEM disciplines and courses (Veilleux et al., 2013; Zumbrunn et al., 2014; Wilson et al., 2015), with many studies focused on explaining and eliminating “gaps” between marginalized and privileged groups through belonging and belonging interventions (Good et al., 2012; Stout et al., 2012; Rosenthal et al., 2013; Walton et al., 2015; Master et al. 2016; Broda et al., 2018; Rainey et al., 2018; Sax et al., 2018; Daniels et al., 2019; Deiglmayr et al., 2019; Hoehne and Zander, 2019; Wilton et al., 2019; Binning et al., 2020; Fink et al., 2020; Rodriguez and Blaney, 2021; Edwards et al., 2022). The current investigation examines how students construct a sense of belonging in a high-enrollment, introductory-level general chemistry course, and how that construction might vary based on student identity and other characteristics.This exploratory, mixed-methods study sheds light on the sources of belonging in general chemistry through thematic analysis of first-year students’ written responses on a belonging survey. The study began with qualitative codebook development and coding of all available belonging explanations. The qualitative data were then transformed into quantitative frequency data. This transformation enabled identification of the most frequent themes among all participants and exploration of variation across student groups. Importantly, in response to recent calls in the literature (e.g., Ladson-Billings, 2007; Gutiérrez, 2013; Pearson et al., 2022) and constructive criticism from anonymous reviewers, this study strives to adopt a critical lens and consider how the learning environment reflects and reinforces systemic biases (e.g., racism, sexism) that marginalize some students and privilege others, thereby leading to inequities in classroom experiences, belonging, and other outcomes. This study also considers the exclusionary impact of a selective institutional culture that places a high value on grades. Most colleges are inherently “selective” in the sense that they rank and sort students through admissions and ongoing assessment (e.g., Jury et al., 2015), but this selectivity function is especially acute at elite universities and may negatively impact belonging. We did not undertake a critical approach until the writing phase of this research and have only begun to reckon with its implications. As a result, we focus narrowly on incorporating critical theory into our discussion of belonging, rather than delving deeply into the critical literature. Nonetheless, by examining classroom and institutional contexts as the targets for intervention, rather than placing the entire onus for change on students, we hope to advance the conversation about equity in undergraduate general chemistry.
Specifically, this study explores first-year students’ construction of course-level belonging in a general chemistry course sequence with high-stakes assessments, situated in a selective institution with a high-income, predominantly White student population. The analyses therefore probed the meaning of belonging for students immersed in the transition to college, surrounded by high-achieving and well-resourced peers, and confronted with challenging coursework that culminates in exams. Special attention was paid to groups marginalized in higher education, including Black or African American students and Hispanic or Latina/o/x students, and specifically in STEM, namely women. Attention was also paid to students who may feel devalued or discouraged by this performance-oriented learning environment, particularly students who have struggled in the past (on Advanced Placement exams) or are struggling during the transition to college to attain high course grades (in chemistry). Illuminating the factors that support and undermine belonging in this general chemistry course, especially for marginalized and discouraged groups, may help instructors reflect on what it means to belong in their course, discipline, and institutional context and how to support the belonging of diverse students.
Sources of belonging
Correlational research has examined the relationship between perceived belonging and a range of other variables. Belonging in STEM has been associated with behavioral and emotional engagement in class (Wilson et al., 2015), self-efficacy and perceived ability (Veilleux et al., 2013), and perceived compatibility between being a woman and a physician (Rosenthal et al., 2013). According to one model, belonging derives in part from a supportive classroom environment, and it contributes to greater self-efficacy, behavioral engagement, and ultimately performance (Zumbrunn et al., 2014). The relationship between course-level belonging and academic success (exam performance, retention between semesters) has been illustrated with two different populations of general chemistry students (Fink et al., 2020; Edwards et al., 2022). This connection between chemistry students’ belonging perceptions and course outcomes motivated the current investigation into the sources of course-level belonging.While correlational data illustrate how course belonging and other variables relate, they do not conclusively speak to the directionality of those relationships and the underpinnings of belonging. Limited qualitative research has solicited student ideas about the sources of belonging in STEM contexts. Some results echo the correlational research, like the finding that four key factors influenced the belonging of all STEM students: interpersonal relationships, perceived competence, personal interest, and science identity (Rainey et al., 2018). At the same time, qualitative work highlights how students’ attributes, like their demographic identities, influence their experiences of STEM learning environments and can lead them to rely on different sources of belonging. For instance, belonging may be threatened by under-representation of students’ gender, cultural, or socioeconomic background in the classroom (Rainey et al., 2018; Daniels et al., 2019; Rodriguez and Blaney, 2021). In such cases, students may lean on identity-based affinity groups (Rodriguez and Blaney, 2021) or a small network of supportive peers and mentors (Daniels et al., 2019) to bolster their belonging in STEM disciplines and classrooms.
Experimental research provides more causal evidence about the foundations of belonging. For instance, belonging interventions are typically designed to convince students that belonging uncertainty is a common experience shared by students of all backgrounds, and such uncertainty is temporary and will eventually resolve into a sense of belonging (Walton and Cohen, 2007; 2011). Several studies have demonstrated positive effects of student-targeted belonging interventions (vs. controls) on student outcomes, particularly among women (Walton et al., 2015), first-generation students and other groups identified as “disadvantaged” according to institutional data (Yeager et al., 2016; Murphy et al., 2020), and Asian, Black or African-American, and Hispanic or Latina/o/x students (Walton and Cohen, 2007; 2011; Binning et al., 2020; Murphy et al., 2020). While some studies have failed to detect significant effects (Broda et al., 2018; Fink et al., 2020), successful interventions suggest that students can gain belonging from finding similarities between themselves and peers, including the shared experience of doubting their belonging (Walton and Cohen, 2007; 2011). Another body of research focuses on growth mindset interventions, which have typically targeted students’ beliefs about intelligence, encouraging them to adopt a flexible (vs. fixed) view of academic ability (i.e., a growth mindset; Aronson et al., 2002; Good et al., 2003; Blackwell et al., 2007; Paunesku et al., 2015; Yeager et al., 2016a; 2016b; 2019; Fink et al., 2018; Binning et al., 2020; Fink et al., 2021). Recent work has shifted from intervening with students to intervening with faculty and the educational environment. Muenks et al. (2020) found students are sensitive to experimentally induced and actual STEM faculty mindsets, and crucially, students who encounter growth-minded faculty reported greater belonging in class. This result supports a model where sense of belonging derives from a supportive classroom environment (Zumbrunn et al., 2014).
Finally, other discipline-based education research points towards sources of belonging. For example, replacing a lecture-based biology course with a “high structure” course using multiple active-learning pedagogies increased students’ belonging; this increase was attributed to greater perceptions of faculty support and classroom comfort, plus an increased sense of peer support among women only (Wilton et al., 2019). These results align with a recent review by physics education researchers, which examined women's belonging in physics contexts (Lewis et al., 2016). Their review identified numerous factors that can influence women's belonging, including the quantity and quality of women peers and role models, ability beliefs and stereotypes, and outside influences like affirmation and support from non-academic relationships. The current study complements Lewis et al.'s (2016) review paper by exploring belonging in the context of general chemistry, where belonging concerns may be especially acute because a large proportion of students are first-year students transitioning into college.
Belonging inequities in higher education and STEM
As noted above, a prominent goal of belonging research (including our previous work) has been to “close the gap” between marginalized and privileged groups in terms of achievement, retention, and belonging in higher education and STEM. However, such “gap gazing” assumes that the ideal education system is one where marginalized and minoritized groups (e.g., low-income students, first-generation students, and students of color) attain the same outcomes as privileged groups (e.g., higher income students, continuing generation students, and White students) (Gutiérrez, 2008; Gutiérrez and Dixon-Román, 2010). In other words, this framework assumes the preservation of the current education system and strives to make diverse students successful within it. This approach fails to reckon with the exclusionary past and present of the education system and other structures in our society, and it measures success according to standards developed by and for groups already privileged within that system (Ladson-Billings, 2006; 2007; Gutiérrez and Dixon-Román, 2010; Pearson et al., 2022).In contrast, a critical perspective highlights how systemic educational inequities, combined with exclusionary institutional norms and practices, produce further inequities in student outcomes. Rather than narrowly defining academic excellence in terms of high grades, critical scholars take a more comprehensive, student-driven approach to conceptualizing success (e.g., as an ability to give back to one's community; Pearson et al., 2022). For instructors, a critical approach might involve reducing the weight of high-stakes assessments, which are known to produce biased results (Walton and Spencer, 2009; Walton et al., 2013), during the calculation of course grades. When students from all backgrounds are expected to navigate a university culture that may be incongruent with their own (e.g., Stephens, Fryberg et al., 2012; Stephens, Townsend et al., 2012; Jury et al., 2015; Boucher et al., 2017; Canning et al., 2020; Chang et al., 2020) in order to perform well on traditional measures of achievement (e.g., high-stakes testing), the definition and construction of belonging may look vastly different for marginalized vs. privileged groups.
The question of who belongs and how learners develop belonging is especially salient in STEM fields, where academic and professional inequities based on race, gender, and other aspects of identity (e.g., sexual orientation, ability status, socio-economic status) are persistent and well-documented (e.g., National Science Foundation, 2021). The culture of science reflects the norms of those most privileged with access to STEM education and professional opportunities, particularly White men, and it therefore marginalizes learners with one or more different identities, such as women of color (Carlone and Johnson, 2007; Rodriquez and Blaney, 2021). A recent study documented this systemic bias in the discipline of chemistry: introductory chemistry courses across 12 institutions (re)produced educational debts, failing to convey chemistry content knowledge to women and Black men as effectively as their peers (Van Dusen et al., 2021). Systemic bias in chemistry also leads to educational debts in terms of belonging, because general chemistry courses fail to instill the same sense of course-level belonging in women, particularly women of color, as men (Fink et al., 2020; Edwards et al., 2022). Thus, while all first-year students encounter obstacles and need time to develop belonging in general chemistry, marginalized students may experience additional challenges, e.g., an expectation that they must assimilate to or at least accommodate the dominant culture in order to belong. This study relies on students' own voices to illuminate how they define belonging in general chemistry and what aspects of the course context shore up or erode their sense of belonging.
Research objectives
The mixed-methods study described below explores the following research questions:(1) What themes are most prominent across all first-year students’ explanations of their belonging early in General Chemistry 1? Do those themes change over the course of General Chemistry 1?
(2) Once again looking at all first-year students, are the most prominent themes in General Chemistry 2 similar to or different from General Chemistry 1?
(3) How do first-year students marginalized (vs. privileged) in STEM education or discouraged by selectivity construct a sense of belonging in general chemistry? Specifically, this study examines the belonging of: (a) students with identities minoritized in STEM, including Black or African American students, Hispanic or Latina/o/x students, and women; (b) students who have struggled or are struggling in the college transition to perform well on traditional achievement measures; and (c) students who report a high level of belonging uncertainty in the course.
The first two questions emphasize the shared experience of students transitioning into college and adjusting to the environment of a large, introductory STEM course. The third research question considers how belonging in general chemistry might carry different meanings for students, depending on their unique identities and academic experiences.
Given the history of racism in higher education (Ladson-Billings, 2006; 2007; Gutiérrez and Dixon-Román, 2010) and the institutional context of this study (i.e., a predominantly White university), Black or African American and Hispanic or Latina/o/x students may question or even resist a sense of belonging in general chemistry, especially if such belonging requires assimilation to dominant cultural norms (e.g., independent vs. communal goal orientations; Stephens et al., 2012; Boucher et al., 2017; Chang et al., 2020). Moreover, a lack of racial and ethnic representation in the classroom may trigger identity threat for minoritized students (Purdie-Vaughns et al., 2008), potentially undermining belonging and requiring coping strategies to shore it up. Similar processes may influence the belonging of women. Although gender-related inequity in higher education and STEM is fundamentally different than race-related inequity, women may also question their course-level belonging and lean on coping strategies when faced with a lack of representation (Murphy et al., 2007) and/or masculine cultural norms in STEM (Boucher et al., 2017; Daniels et al., 2019). Women are well-represented on the instructor team and in the class composition of general chemistry at our institution, but students’ awareness of cultural stereotypes and professional disparities for women in STEM may contribute to identity threat.
Finally, the variation analyses examine the belonging of students who may feel devalued or discouraged by the selectivity intrinsic to an elite institution. Students are sensitive to whether admissions processes and course assessments aim to competitively select the “best” students (i.e., those with highest grades) or to support learning and skill development, and selectivity can undermine performance especially for marginalized students (Smeding et al., 2013; Jury et al., 2015). Given the association of belonging with self-efficacy and perceived ability (Veilleux et al., 2013; Zumbrunn et al., 2014; Wilson et al., 2015), students who struggle to attain high scores on traditional achievement metrics may also struggle to construct a sense of belonging. The current study explores this construction, differentiating students on the basis of Advanced Placement (AP) exam scores and course grades, as well as self-reported belonging uncertainty scores. Belonging uncertainty reflects students’ doubts about their belonging and their sense that belonging is performance dependent (Walton and Cohen, 2011; Deiglmayr et al., 2019; Hoehne and Zander, 2019; Fink et al., 2020; Edwards et al., 2022).
As noted earlier, standardized assessments like AP exams and other high-stakes tests are tools of an inequitable education system (Ladson-Billings, 2007; Gutiérrez and Dixon-Román, 2010; Klugman, 2013), and they are known to be systematically biased against minoritized groups (Walton and Spencer, 2009; Walton et al., 2013). Belonging uncertainty also appears sensitive to identity, with women reporting higher degrees of uncertainty in general chemistry than men (Fink et al., 2020; Edwards et al., 2022). As a result, these three variables (i.e., AP exam scores, course grades, belonging uncertainty) may covary with the identity variables (i.e., race/ethnicity and gender). Nonetheless, each analysis highlights the paradoxical challenge of convincing students that anyone can belong in general chemistry while also evaluating their achievement using traditional metrics of success like high-stakes testing.
Methods
Study setting
The study was conducted at a private, selective, Midwestern research university in the United States with a predominantly White, well-resourced student population. At the start of this study (Fall 2017), the university's acceptance rate was approximately 16%, and composite scores on the standardized ACT admissions test typically ranged from 33 (25th percentile at our institution) to 35 (75th percentile at our institution) out of 36. At this time, the undergraduate student body was 1% American Indian, Alaska Native, Native Hawaiian, or Pacific Islander, 21% Asian, 9% Black or African American, 9% Hispanic or Latina/o/x, 52% White only, and 10% White multiracial, with unreported race and ethnicity for 2% of students (percentages sum to more than 100% because students could report more than one identity). Using Pell Grant eligibility as a proxy for low income status, between 8% (2015–2016) and 14% (2018–2019) of students qualified as low income. Data from a 2017 study indicate that as of the graduating class of 2013, the institution enrolled more students from the top 1% income bracket than from the entire bottom 60% (Chetty et al., 2017, viaAisch et al., 2017).†Data were collected over two academic years (fall 2017 through spring 2019) in a two-semester general chemistry course sequence. Between 700 to 800 students enrolled in General Chemistry 1 (GC1) each fall, and approximately 550 to 650 enrolled in General Chemistry 2 (GC2) each spring. The decline in enrollment between semesters predominantly reflects the fact that some engineering majors (e.g., electrical engineering) only require GC1. Both courses were structured very similarly, including three one-hour lectures and a required recitation per week, plus mandatory enrollment in an associated but separate laboratory course. Although GC1 and GC2 were divided into two or three lecture sections each, the instructor team managed those sections as a unit. All instructors utilized the same assessments, students from different lecture sections were mixed together in recitations, and all assessments were graded through a unified procedure and used the same absolute grading scale.
In terms of assessment, students encountered a mix of low- and high-stakes evaluations, but the latter were more heavily weighted. On one hand, students completed ungraded clicker questions during lecture sessions and ungraded weekly homework sets; they completed weekly graded quizzes (and could drop the lowest two) and weekly graded homework sets (and could drop the lowest three); and they were allowed to drop the lowest of three midterm exam grades. On the other hand, the remaining two midterm exams and cumulative final exam comprised a large portion of their grade. For instance, exams accounted for over 75% of students’ course grades (330/430 possible points) in GC1 in Fall 2017. Thus, succeeding in the course sequence required success on high-stakes evaluations.
To help students succeed, they were strongly encouraged to participate in supplemental learning activities. For example, students could attend informal, daily help sessions with any instructor. In GC1, students had five instructors to choose from, including three or four women depending on the year, with at least one woman who lectured a section. In GC2, students had four instructors to choose from, including two women who had been members of the GC1 instructor team as well (but neither lectured during GC2). In addition, students were offered a formalized, department-sponsored Peer-Led Team Learning (PLTL) program (Hockings et al., 2008; Frey et al., 2018; Frey and Lewis, in press). Nearly 70% of GC1 students and about 50% of GC2 students participated in the PLTL program, joining two-hour weekly meetings to collaboratively solve new problems with 8–10 classmates facilitated by a trained peer leader. Finally, students received social-psychological interventions designed to support them in the transition into college. All GC1 students received a three-part growth mindset intervention embedded in their graded homework (Fink et al., 2018). The instructors also collaborated with the research team to pilot a social-belonging intervention (e.g., Walton and Cohen, 2007, 2011) in GC2 during spring 2018, but the experimental data showed no effect on students’ performance or self-reported belonging (Fink et al., 2020). Overall, the general chemistry instructors strive to provide a supportive environment where all students can feel comfortable and find resources to succeed.
Participants
All students enrolled in GC1 or GC2 from fall 2017 through spring 2019 (N = 1508‡) were invited to participate in this research, which was approved by the university's Institutional Review Board. Voluntary, informed consent was gathered in person during the associated general chemistry laboratory course, and a total of 1408 (93.4%) students consented to participate. By participating, students contributed their STEM course data, registrar data, and survey responses to the research team's educational data repository. Students received extra credit towards their laboratory course grades as compensation for completing surveys about general chemistry; students could also receive the extra credit for an alternative writing assignment, if they chose not to participate in the study. As part of the educational data repository, participants also may have been invited to complete surveys in other STEM courses, in exchange for extra credit in those courses.Following our previous research (Fink et al., 2018; Frey et al., 2018; Fink et al., 2020), this study focused on first-year students only (N = 1141), with analysis further limited to those who provided survey responses (N = 1078). This approach addresses research questions 1 and 2, illuminating the experience of students who complete the course during the transition into college. Given that first-year students’ early feelings of belonging predict performance and attrition from the course sequence (Fink et al., 2020; Edwards et al., 2022), it is critical to examine how they explain their course belonging.
Data
Race and ethnicity data were organized into four categories: Asian, Black or African American, Hispanic or Latina/o/x, and White. Students could also identify as American Indian, Alaska Native, Native Hawaiian, or Pacific Islander, but these groups were not included in the race and ethnicity analyses due to their small numbers (see below); however, these students were retained in the overall analyses and other variation analyses. Students could report more than one race or ethnicity, but to simplify analysis, they were placed in only one category. Because the university is predominantly White, priority was given to other identities less privileged by this institutional context, as follows: Asian identity took precedence over White identity (i.e., an Asian and White student was categorized as Asian), and Black or African American and Hispanic or Latina/o/x identity took precedence over Asian and White identities (e.g., a Black and White student was categorized as Black; a Hispanic and Asian student was categorized as Hispanic). If a student reported multiple marginalized identities (e.g., Black and Hispanic, American Indian and Hispanic), they were placed in a “multiple marginalized identity” category, which was not included in the four-category race and ethnicity analyses due to its small size (see below). This categorization scheme is a blunt tool for capturing students’ race and ethnicity, and students’ assigned categories may not accurately or completely reflect their own views of their identity. Nonetheless, the four-category race-and-ethnicity variable allows us to disaggregate the data and may illustrate meaningful variation in students’ construction of general chemistry belonging.
The final sample included 606 (56.2%) female and 472 (43.8%) male students. 384 (35.6%) participants were categorized as Asian, 101 (9.4%) as Black or African American (abbreviated as Black students in the results), 95 (8.8%) as Hispanic or Latina/o/x (abbreviated as Hispanic students), and 443 (41.1%) as White. A total of 29 (2.7%) students identified as American Indian, Alaska Native, Native Hawaiian, or Pacific Islander and/or reported multiple marginalized identities, and 26 (2.4%) students did not report race and ethnicity. Because the population of first-generation students remains small at this institution – only 61 (5.6%) students in this sample – that variable is not incorporated in this study.
The sense of belonging survey included six Likert-scale items about belonging in the course sequence, plus an open-ended question. The quantitative items were assessed on a six-point agreement scale (1 = strongly disagree to 6 = strongly agree), and their validity for this student population and course context was examined by prior research (Fink et al., 2020). Specifically, exploratory and confirmatory factor analyses showed the quantitative items statistically separated into two factors. The belonging factor (four items) explored students’ comfort with others in the course context (peers, instructors) and their overall feelings of fit, while the uncertainty factor (two items) examined the stability and performance contingency of perceived belonging. Fink et al. (2020) demonstrated predictive validity, replicating the finding that belonging can predict STEM course outcomes (e.g., Stout et al., 2012). Cronbach's alpha values for the current sample indicate acceptable internal consistency for both factors (0.7 < α < 0.8 = acceptable, 0.8 < α 0.9 = good), with scores ranging from 0.78–0.84 for the four belonging items and 0.72–0.84 for the two uncertainty items, depending on the survey time-point.
Crucially, one of the belonging items, “Setting aside my performance in the General Chemistry course, I feel like I belong,” was followed by the open-ended question, “Please explain why you agree or disagree with this statement.” The current investigation thematically analyzes students’ written responses to that open-ended question. All available first-year responses (N = 3255) were included in the analyses described below, grouped according to semester (GC1, GC2) and time (early, late). Mean response length ranged from 24.9 words (SD = 11.9) on the early GC1 survey to 16.8 words (SD = 11.3) on the late GC2 survey, and mean code count ranged from 2.54 codes (SD = 1.45) per early GC1 response to 2.24 codes (SD = 1.14) per late GC2 response. T-tests indicate women wrote slightly longer responses (p ≤ 0.001) and were assigned slightly more codes (p ≤ 0.004) than men on all surveys. ANOVAs show significant effects of race/ethnicity on response length for all surveys (p ≤ 0.0003) except early GC1. Pairwise tests show Asian students wrote shorter responses than all other groups (i.e., Black, Hispanic, and White students) on the late GC1 survey (p ≤ 0.02), while Black students wrote longer responses than Asian and White students on the early GC2 survey (p = 0.0002) and all other groups on the late GC2 survey (p ≤ 0.03). There was a significant effect of race/ethnicity on code counts for the late GC1 survey only (p = 0.03), but no pairwise comparisons proved significant (p ≥ 0.10).
This study utilized scores from four STEM-related AP exams – Biology, Calculus (AB or BC), Chemistry, and Physics (1, 2, C: Mechanics, or C: E & M) – which were combined to create a composite “AP proportion” score (Fink et al., 2018; Frey et al., 2018; Fink et al., 2020). For STEM subjects with multiple AP exams available, a student's highest reported exam score was used to calculate AP proportion. AP proportion scores ranged in value from 0 to 1, increasing 0.25 points for every AP STEM exam where a student earned a score of 4 or 5 (out of 5; 4 is the threshold for earning AP credit at this institution, but credit does not replace enrollment in General Chemistry). Therefore, students who (a) reported no AP STEM exam scores to the registrar or (b) received only AP exam scores ≤3 were assigned an AP proportion of 0, while students who earned scores ≥4 on all four AP STEM exams were assigned an AP proportion of 1. For the variation analyses, students with AP proportion scores of 0 (“no-AP group”) were compared to students with AP proportion scores of 0.75–1 (“high-AP group”).
General Chemistry course grades were obtained from the instructors after each semester. Like AP exam scores, course grades are imperfect measures of student learning, but they provide a gauge of students’ performance – and can have a large impact on student perceptions. As noted above, General Chemistry course grades incorporated multiple assessments, but a majority of points (approximately 75%) came from exams. As a result, these course grades are especially reflective of students’ ability to attain high scores on traditional metrics of academic success. The variation analyses compared the “low-grade group” who earned grades of C+ through D to the “high-grade group” who earned grades of A through B+. While a grade of C+ may not constitute low performance across institutions, it does in this course sequence. Only 21.3% of students received a C+ or lower in GC1 during this study, and only 15% received such a score in GC2. One percent or less of students in each course received an F, incomplete, or withdrew from class, and those students were not included in this analysis.
Finally, belonging uncertainty (BU) scores were calculated by averaging the two uncertainty items from the sense of belonging survey (Fink et al., 2020). BU scores ranged from 1 to 6, with low scores indicating confidence and high scores indicating uncertainty about one's belonging in General Chemistry. We used BU scores as a grouping variable in the current study, rather than belonging scores, for several reasons. First, BU scores among our students proved more variable than belonging scores; BU scores spanned the full 6-point range, while belonging scores demonstrated a strong negative skew (Fink et al., 2020). Second, the concept of “uncertainty” emerged in our thematic analyses (see Table 1), suggesting that BU scores would provide a meaningful link between the qualitative and quantitative data. Third, as noted in the introduction, the shared experience of belonging uncertainty is a key message of belonging interventions to improve student outcomes (e.g., Walton and Cohen, 2007, 2011). Finally, recent research specifically within general chemistry indicates that course-level belonging uncertainty, perhaps even more so than perceived belonging, varies meaningfully across groups and predicts course outcomes (Fink et al., 2020; Edwards et al., 2022; see also Deiglmayr et al., 2019, and Hoehne and Zander, 2019, for other STEM disciplines). Comparison BU groups were constructed based on the response scale (1 = strongly disagree, 2 = disagree, 3 = mildly disagree, 4 = mildly agree, 5 = agree, and 6 = strongly agree). Students who on average strongly disagreed or disagreed with the uncertainty items (i.e., BU scores ≤ 2) were placed in the “confident group,” and students who strongly agreed or agreed with those items (i.e., BU scores ≥ 5) were placed in the “uncertain group.”
| Category | Code | Coding | Short descriptions | Illustrative examples |
|---|---|---|---|---|
| Note: coding column indicates whether individual codes reflect only presence/absence of a theme or also polarity. The majority of codes are marked present (1) or absent (0), but seven are marked present and positive (1), present and negative/neutral (−1), or absent (0). | ||||
| Course attributes | Content/skills | 0, 1 | Comments on topics or skills covered within the course | 1. I really enjoy the material covered in the course |
| 2. I like the problem solving thinking in the class and the analytical thinking | ||||
| Difficulty | 0, 1 | Describes challenges, demands, or appropriateness of difficulty level. Focuses on perceptions of the course, not self-perceptions. | 3. [the class] is a good level for me/gives me a good challenge | |
| 4. [the class] assumed people have been exposed to those concepts | ||||
| Environment | −1, 0, 1 | Describes overall atmosphere of course and how that atmosphere affects self or others | 5. I feel like everyone is working together | |
| 6. The instructor is very inviting and does not make me feel inferior or ignorant when I ask questions | ||||
| Structure | 0, 1 | Evaluates element(s) of the course design, talking about the value or quality of specific policies, procedures, teaching practices, or resources | 7. I find the presentation of material partial to my learning style | |
| 8. There are so many resource available | ||||
| Interest | Behavioral engagement | −1, 0, 1 | Describes personal willingness to put forth or perform specific behaviors for a goal | 9. I feel like I have the work ethic to learn the material deeply and well |
| 10. I don't necessarily feel like I need to (or should…) kill myself to get an A | ||||
| Cognitive/affective engagement | −1, 0, 1 | Describes personal feelings about the course, conveying (lack of) fondness, enthusiasm, or curiosity. −1 can indicate dislike neutrality (“I’m indifferent”) | 11. I enjoy learning the material | |
| 12. I'm not interested in it at all (−1) | ||||
| Individual interest | −1, 0, 1 | Describes (lack of) “passion” for course, reflecting persistent, well-developed (vs. situational, emergent) interest. −1 indicates dispassion/apathy | 13. I have always enjoyed [the course] regardless of how… grades go | |
| 14. I didn't really feel passion about it (−1) | ||||
| Perceptions | Peer comparison | 0, 1 | Compares themselves and their peers, highlighting a difference or similarity | 15. I feel like other students are similarly motivated to me |
| 16. This common goal makes me feel like I belong | ||||
| Self-efficacy | −1, 0, 1 | Explicitly conveys (lack of) belief or confidence in their ability to succeed | 17. I feel like I can… keep up with the material | |
| 18. I don't feel as though my calc 2 skills are up to par for the class | ||||
| Uncertainty | 0, 1 | Conveys being unsure of belonging in course or unsure of major/career | 19. I still have doubts about my prehealth field | |
| 20. Sometimes I feel a bit uneasy if I should belong in prehealth | ||||
| Metacognition | −1, 0, 1 | Reflects on their own comprehension or learning, demonstrating self-awareness of their understanding or study process | 20. Even if my grades may or may not reflect it… I feel as if I have learned | |
| 21. Sometimes the concepts and equations don't connect for me | ||||
| Social | Peers | 0, 1 | Mentions other students in the course, including friends, groups of students, or the whole class | 22. I belong with my other MechEs |
| 23. I think everyone struggles to transition | ||||
| Professors | 0, 1 | Mentions instructors in the course | 24. The professors aim to help me find long-term success | |
| 25. I find the people/professors enjoyable to be around | ||||
| Student attributes | Extracurricular | 0, 1 | Describes non-course experiences that enrich learning or foster inclusion in the course | 26. I love the research I do and am strongly interested in it |
| 27. I volunteer in clinical setting, and I have shadowed professionals in healthcare | ||||
| Identity | 0, 1 | References demographic identity (e.g., own or others’ gender, race, or ethnicity) | 28. There are obvious barriers of being a female | |
| 29. As a possible computer science transfer and black woman | ||||
| Major | 0, 1 | Talks about own or others’ (possible) majors | 30. I feel like a pretty typical STEM major | |
| 31. I am not majoring in a science | ||||
| Preparation | 0, 1 | Describes readiness for the course, e.g., based on high school or previous college classes | 32. I took accel chem and AP chem | |
| 33. I have taken a lot of STEM classes | ||||
| Value | Career | 0, 1 | Discusses long-term professional plans, both general and concrete. Applies to mentions of pre-health and pre-med tracks | 34. I want to be a doctor |
| 35. It is the only field I can imagine myself being in | ||||
| Performance | 0, 1 | Talks about course grades, performance, success, or results | 36. When I don't do well it's hard to take | |
| 37. My performance in class has no effect [on belonging] | ||||
| Requirement | 0, 1 | Mentions the course fulfills a requirement or is necessary for long-term goals | 38. I need the class to get where I want | |
| 39. … into a field that requires this course | ||||
| Utility | −1, 0, 1 | Expresses belief or feeling the course will be applicable or useful for future endeavors | 40. … greatly applies to what I plan to study in the coming years | |
| 41. I don't think it is beneficial to my future avenues of research | ||||
The grouping variables are correlated with one another. For example, prior research indicates that women are more likely to experience high levels of belonging uncertainty (Deiglmayr et al., 2019; Hoehne and Zander, 2019; Fink et al., 2020; Edwards et al., 2022), and in the current study, women were disproportionately assigned to the uncertain group at each time point (χ2(1) values ≥ 13.85, p ≤ 0.0002). Nonetheless, supplemental analyses suggest the grouping variables are not fully redundant with one another (Appendix 1), and variation across all variables was explored despite their inter-correlations.
Analysis
Each team member contributed unique insights to the coding process. J. D. Y. was an undergraduate chemistry major and research assistant trained in qualitative coding. N. V. was also an undergraduate chemistry major, Physics PLTL leader, and research assistant trained in qualitative coding. A. F. is a research scientist with a PhD in linguistics, who specializes in cognitive science and researches student perceptions. A. Y. is a research scientist with a PhD in biology, who specializes in evidence-based, inclusive teaching practices. M. J. C. is a data analyst with a PhD in psychology. R. F. F. is a STEM faculty member (PhD in chemistry) and discipline-based researcher with expertise in qualitative coding. As part of our effort to adopt a critical lens, we offer a brief positionality statement to acknowledge that our lived experiences may differ from participants in this research and influence our interpretation of the data (Pearson et al., 2022). All members of the coding team are cisgender men (2) and women (4) in STEM; one of us identifies as Asian and the rest as White. We share a commitment to supporting the success and well-being of all students, but we also recognize the need to learn from the perspectives of those marginalized in higher education and STEM based on identities we do not share (e.g., gender, race/ethnicity, ability status, sexual orientation, socioeconomic status).
The final codebook included 21 codes grouped into 6 categories (see Table 1 or Appendix 3 for streamlined and detailed codebooks, respectively). To establish the validity of our content analysis, we assessed inter-rater reliability (IRR), which indicates agreement between coders after accounting for chance (Kirppendorff, 2004). Because J. D. Y. and N. V. were tasked with applying the final codebook to the full data sample, we tested their IRR. They independently coded 360 belonging responses, coding each response for the presence or absence of each code (0, 1). Multiple codes could be marked present for an individual response (see Fig. 1). In addition, some codes could be marked not only present, but also positive or negative/neutral (1, −1) in polarity (e.g., describing how a course does or does not have utility for future endeavors; see Table 1). As a result, we calculated IRR using Krippendorff's alpha (α), which is appropriate for ordinal data, among other types, with values ≥0.80 indicating robust reliability and values ≥0.67 sufficing for “tentative conclusions” (Kirppendorff, 2004).
 | ||
| Fig. 1 Complete example response with qualitative coding applied. Each code represented by a different color. Multiple codes could apply to the same phrase (e.g., Difficulty and Structure), and a subset of codes could be coded as positive or negative (e.g., Self-Efficacy). | ||
Across the entire IRR sample (i.e., simultaneously comparing each coder's decisions across all 21 codes), we observed an overall α of 0.71, suggesting a moderate degree of chance-adjusted IRR. We then analyzed each individual code separately, to identify specific points of disagreement between the two coders. The majority of codes met the threshold of α ≥ 0.67. Codes with lower levels of agreement (Environment, Identity, Metacognition, Self-Efficacy, Structure) tended to be low in frequency (e.g., Environment and Identity appeared in less than 1% of responses), so each instance of disagreement had a strong influence on α. We retained these infrequent codes because of their potential usefulness for instructor feedback and program evaluation. Given the exploratory nature of this study, we decided there was sufficient agreement to proceed with coding, which was supported by validity tests (Appendix 4). We also counterbalanced coding to mitigate the effect of any remaining coder variation.
Results
Overall view of General Chemistry 1 and 2 themes (findings 1 and 2)
In the overall results, we highlight the top five themes on each survey for the full first-year sample. While there is not a sharp decline in frequency after the fifth most frequent theme, we chose this top-five focus for the sake of clarity and not overwhelming the reader. Also, narrowing the analysis to the top themes still demonstrates the diversity and stability of students’ belonging explanations. | ||
| Fig. 2 Key findings and representative quotes. Findings 1 and 2 (orange) stem from the overall analysis of all first-year participants, and top-five codes are color-coded according to category. Findings 3–6 reflect the variation analyses and are grouped into two components of belonging: perceptions of self and others (blue), and motivation (green). Each variation finding is supported by two codes (e.g., Peers and Peer Comparison for finding 3). Most quotes were selected to illustrate significant differences in frequency between marginalized or discouraged (bold) vs. privileged groups. | ||
 | ||
| Fig. 3 Top ten (of 21) codes assigned to General Chemistry belonging explanations at each survey time point. Codes are color-coded, with frequency data (points) connected by lines to help readers visualize the codes’ relative prominence over time. Ns for each survey are listed on the x-axis. | ||
Variation linked to student characteristics
Next, we examined how the construction of belonging varied based on student characteristics (i.e., gender, race/ethnicity, AP proportion, course grade, and BU score). Significant results from each variation analysis are presented in Appendix 6, and the following sections integrate these analyses. Rather than consecutively describing each analysis, we present key findings (3–6) that emerged across multiple groups. Findings 3–6 are further paired into two overarching elements: perception of others and self, and motivation. Excerpts from student responses are provided in summary Fig. 2 to guide interpretation of the quantitative data and represent student perspectives in their own words.Perception of others and self (findings 3 and 4)
For instance, women were assigned the Peers code significantly more often than men at the end of GC1 (χ2(1) = 10.30, p = 0.001) and early in GC2 (χ2(1) = 5.49, p = 0.02); this gender difference shrunk over time and became marginally significant by the end of GC2 (χ2(1) = 3.53, p = 0.06). A parallel pattern emerged for the Peer Comparison code, which was also assigned to women's responses significantly more often than men's (late GC1: χ2(1) = 12.77, p < 0.001; early GC2: χ2(1) = 5.36, p = 0.02). Similarly, the no-AP group was assigned the Peers code more frequently than the high-AP group on the late GC1 survey (χ2(1) = 9.53, p = 0.002), and they were assigned the Peer Comparison code more often at the end of each semester (late GC1: χ2(1) = 9.37, p = 0.002; late GC2: χ2(1) = 4.03, p = 0.045). This social focus was less pervasive for other groups, but the low-grade and uncertain groups were assigned the Peer Comparison code significantly more often than their counterparts on the early GC1 survey (low-grade group: χ2(1) = 6.08, p = 0.01; uncertain group: χ2(1) = 3.92, p = 0.048).
One contrasting pattern that emerged was a significant effect of race/ethnicity on Peers at the end of GC2 (χ2(3) = 8.36, p = 0.04); this is the only instance where a marginalized (vs. privileged) group focused less on other students. This main effect reflects a marginally significant tendency for Hispanic students (4.8%) to receive the Peers code less often than Asian students (17.3%; z = −2.30, p = 0.09) and White students (17.3%; z = 2.31, p = 0.09). Except for this one pattern, the majority of Peer-related results demonstrate how social aspects of the classroom environment can especially influence marginalized or discouraged students’ feelings of belonging in a course.
Second, we also observed a robust tendency for the low-grade group to receive the Self-Efficacy code more frequently than the high-grade group, with significant differences emerging on every survey except early GC1, when students had yet to receive much feedback (late GC1: χ2(1) = 10.75, p = 0.001; early GC2: χ2(1) = 7.16, p = 0.007; late GC2: χ2(1) = 6.98, p = 0.008). In this case, the polarity results were more mixed. On the late GC1 survey, the difference lay in negative Self-Efficacy, which was assigned to 13.9% of low-grade group responses but only 1.4% of high-grade group responses. However, throughout GC2, the grade-group difference in Self-Efficacy reflected both positive and negative self-efficacy beliefs. The low-grade group was assigned both positive (early GC2: 16.1%; late GC2: 15.6%) and negative Self-Efficacy (early GC2: 6.5%; late GC2: 6.3%) more often than the high-grade group (positive: 9.7% early GC2, 8.6% late GC2; negative: 1.4% early GC2, 1.0% late GC2). Thus, students struggling to earn high grades in GC2 were more likely to be both self-affirming and self-critical regarding ability (Fig. 2).
While it is unsurprising that students who struggle to succeed on exams (AP or course-based) would reflect on their ability to succeed, these results illustrate the connection between student achievement and belonging. Even though the open-response prompt asked students to “set aside your performance in class,” they struggled to do so. The impact of grades on belonging was clearest at the end of GC1, where the belonging explanations of the no-AP group (χ2(1) = 15.77, p < 0.001), the low-grade group (χ2(1) = 30.37, p < 0.001), and also women (χ2(1) = 5.01, p = 0.03) were assigned the Performance code significantly more often than their comparison groups.
While discouraged groups focused especially on self-perceptions of ability, some marginalized students explained their belonging via self-evaluation of learning, which was indexed by Metacognition and could also be marked for polarity. This code was not a source of variation across multiple grouping variables; instead, only a significant effect of race/ethnicity was observed on the late GC1 (χ2(2) = 8.44, p = 0.04) and late GC2 surveys (χ2(2) = 8.76, p = 0.03). Post hoc tests show the belonging responses of Black students were assigned Metacognition significantly more often than White students (z = −2.64, p = 0.04) on the late GC1 survey and marginally more often than Hispanic (z = −2.41, p = 0.07) and White students (z = −2.28, p = 0.09) on the late GC2 survey. Frequency tables reveal these differences are driven by positive (vs. negative) Metacognition. On the late GC1 survey, 21.1% of Black students were assigned the positive Metacognition code, compared to 9.1% of White students; on the late GC2 survey, 15.6% of Black students were ascribed positive Metacognition, compared to 3.2% of Hispanic students and 8.5% of White students. The quotes in Fig. 2 illustrates how positive self-perceptions about learning may bolster belonging for groups marginalized in STEM.
Motivation (findings 5 and 6)
Variation in the Career code was often accompanied by variation in the Requirement code, which also belongs to the Value category, as students mentioned the required steps to pursue their target field. The low-grade group was ascribed Requirement more often than the high-grade group at the beginning and end of the year (early GC1: χ2(1) = 4.28, p = 0.04; late GC2: χ2(1) = 5.66, p = 0.02; Appendix 6). In addition, a marginal main effect of race/ethnicity emerged at the start of the course (early GC1: χ2(3) = 7.28, p = 0.06), with Black students (21.8%) assigned the Requirement code significantly more often than Asian students (10.8%; z = 2.68, p = 0.04). The quotes in Fig. 2 show how marginalized groups may derive belonging from completing steps towards their career goals.
Crucially, marginalized or discouraged groups were assigned the positive Cognitive and Affective Engagement code less often than privileged groups. A main effect of race and ethnicity proved significant on the early GC2 survey (χ2(3) = 7.85, p = 0.049) and marginal on the late GC2 survey (χ2(3) = 7.36, p = 0.06), in both cases reflecting a difference between Black and Asian students (early GC2: z = −2.29, p = 0.095; late GC2: z = −2.63, p = 0.04). Frequency tables confirmed Black students (early GC2: 36.9%; late GC2: 25.0%) were ascribed positive Cognitive and Affective Engagement less often than Asian students (early GC2: 52.2%; late GC2: 48.5%). The low-grade group also received this code less frequently then the high-grade group on most surveys (early GC1: χ2(1) = 3.76, p = 0.053; early GC2: χ2(1) = 4.48, p = 0.03; late GC2: χ2(1) = 5.47, p = 0.02), reflecting less positive Engagement among low scorers (early GC1: 35.2%; early GC2: 32.3%; late GC2: 23.4%) compared to high scorers (early GC1: 43.5%; early GC2: 55.5%; late GC2: 52.9%). Last, Cognitive and Affective Engagement was marked less often for the uncertain group compared to the confident group at the beginning and end of the year (early GC1: χ2(1) = 4.59, p = 0.03; late GC2: χ2(1) = 4.82, p = 0.03), due to less positive engagement among uncertain (early GC1: 30.5%; late GC2: 26.6%) compared to confident students (early GC1: 44.6%; late GC2: 55.8%).
The Individual Interest code is similar to Cognitive and Affective Engagement, but it indexes a more well-developed, persistent passion about a subject, rather than emergent or situational interest (Harackiewicz et al., 2016a). Fig. 2 contains positive and negative example quotes, reflecting intense fondness or disdain for chemistry, respectively. Individual Interest was seldom marked, averaging less than 5% of survey responses at any time (Appendix 5), and it showed variation only at the end of the course (late GC2). A marginal main effect of race and ethnicity was found (late GC2: χ2(3) = 7.43, p = 0.059), with post hoc tests showing a significant difference in frequency between Hispanic and White students (z = −2.64, p = 0.04). Frequency tables revealed Hispanic students (6.5%) were coded for positive Individual Interest more frequently than White students (1.4%). Discouraged (vs. privileged) groups were also assigned Individual Interest more often, but their responses were mixed in polarity. Specifically, the low-grade group (Fisher's exact: p = 0.005) and the uncertain group (Fisher's exact: p = 0.001) were assigned Individual Interest more often than their counterparts on the late GC2 survey, reflecting more frequent coding of both positive and negative Individual Interest. Thus, marginalized or discouraged (vs. privileged) groups were coded for positive situational interest and enjoyment (Cognitive and Affective Engagement) less often, but feelings of passion for or against the subject (Individual Interest) more often at the end of the year.
Discussion and implications for instructors
Overall patterns
The only code that failed to return to the top five after the first survey was Preparation, which dropped in frequency from 20.6% to 5.5% over the course of GC1. While extensive research has established that academic “preparation” measures (e.g., ACT/SAT scores, AP exam scores, high school course grades) continue to predict exam performance throughout general chemistry and beyond (Tai et al., 2005; Sadler and Tai, 2007; Xu et al., 2013; Frey et al., 2017; Frey et al., 2018; Fink et al., 2020; Edwards et al., 2022), this result indicates that the influence of high-school academic preparation on students’ belonging perceptions dwindles over the first-semester course. Given this pattern of results, instructors may wish to address student concerns about academic preparation early in GC1, e.g., by promoting a growth-mindset environment and affirming all students’ potential to succeed (Dweck and Leggett 1988; Fink et al., 2018; Canning et al., 2019; Muenks et al., 2020), but focus on the more stable belonging themes as the course continues.
Perception of self and others
Finding 3 suggests these groups may experience stronger “social-comparison concerns” than others, i.e., a stronger tendency to evaluate themselves and their abilities relative to other students (Festinger, 1954). Students who report more social-comparison concerns tend to feel less comfortable expressing ideas in group settings, and they exhibit poorer STEM outcomes, including grades, persistence, and self-efficacy (Micari and Drane, 2011). However, such concerns can be reduced through targeted intervention. One experimental study among STEM peer-learning groups showed that peer learning with a growth mindset intervention significantly reduced students’ social-comparison concerns compared to standard peer learning, with a larger effect among disadvantaged students (per ACT/SAT scores; Micari and Pazos, 2014). Given the current finding that marginalized students frequently mentioned peers in their belonging explanations, instructors may wish to proactively address social-comparisons concerns in order to support student belonging. Besides social-psychological interventions, instructors can also structure group-work activities in ways that promote equitable participation and carefully train peer leaders to ensure awareness of potential student concerns (Eddy et al., 2015). Reducing comparison concerns will likely bear dividends if it facilitates student contributions to group work, because collaborative learning is known to improve STEM achievement and retention (e.g., Hockings et al., 2008; Frey et al., 2018; Micari and Pazos, 2019; Frey and Lewis, in press), as well as self-efficacy and self-regulation (Micari and Pazos, 2021).
Instructors can also leverage the importance of Peers to help cultivate an equitable and inclusive classroom environment. White et al. (2020) recently reviewed evidence-based practices for fostering equity and inclusion in chemistry courses, and two practices focus heavily on social relationships in the classroom. As above, the authors highlight the value of active learning and group work, describing a “Teach Your Peers” group quiz activity that “rewards students for learning, encourages students to assist their peers, and enhances the classroom community” (p. 334). Such group work creates the opportunity for students to collaborate and succeed together rather than competing. Second, White et al. (2020) discuss cultivating student–student (and student–instructor) relationships. They suggest relationship-building strategies that may already be familiar, including think-pair-share activities, as well as novel ideas like having students create an online classroom presence (e.g., using tools like Padlet). Students can also benefit from relationships with upper-year Learning Assistants (LAs) or other peer leaders; one recent study in introductory biology showed the presence of LAs increased students’ belonging and confidence and decreased feelings of isolation (Clements et al., 2022). Explicitly creating opportunities for students from all identity groups and academic backgrounds to connect and learn together may convey to marginalized students that belonging in STEM, at least in this general chemistry course, does not necessarily require assimilation or conformity to one mode of success.
These results align with prior research demonstrating the correlation between belonging and self-efficacy. Surveys and focus groups among STEM majors in quantitative fields like chemistry found that students’ perceived belonging was significantly correlated with self-efficacy and perceived ability (expected GPA in major), though not actual ability (achieved GPA) (Veilleux et al., 2013). In general chemistry, performance has been shown to predict belonging in the course (Edwards et al., 2022). Another study in psychology showed a significant correlation between self-efficacy and belonging, which predicted intent to persist with a psychology major (Lewis and Hodges, 2015). Belonging undoubtedly has social components, but the current study supports the idea that perceived ability is an influential factor for course-level belonging (Rainey et al., 2018).
At the same time, the impact of perceived ability may be counterbalanced by a metacognitive focus on learning. Current results showed that Black students were significantly more likely than White (late GC1, late GC2) and Hispanic (late GC2) students to positively reflect on their learning and understanding in the course (indexed by the Metacognition code). Thus, marginalized students’ belonging may be bolstered through self-awareness of and confidence in their conceptual learning. This interpretation is strengthened by research showing that metacognitive awareness supports better student outcomes, including improved course performance (Akyol et al., 2010; Mutambuki et al., 2020) and more positive perceptions like increased chemistry self-efficacy (Kirbulut, 2014). Therefore, another potential strategy to support belonging is to explicitly teach students metacognitive strategies for planning, monitoring and evaluating their own learning (Tanner, 2012), so they are equipped to discern their level of understanding and feel more confident in belonging when they successfully learn. The benefit of positive self-perceptions and a belief in one's potential to learn may prove especially strong in classes where students’ peers also endorse the growth-mindset message (Yeager et al., 2019).
Motivation
We interpret these Career- and Requirement-focused belonging explanations as evidence that marginalized students may gain belonging – or if not increased belonging, at least resilience – from the knowledge that completing general chemistry will serve their long-term professional goals, e.g., becoming a doctor, engineer, or scientist. This explanation aligns with previous research demonstrating correlations between belonging and utility value (Master et al., 2016; Stout et al., 2012) and the positive effects of utility-value interventions on student interest and success (Harackiewicz et al., 2016a). For groups marginalized on the basis of identity in particular, this future orientation and emphasis on utility value may serve as a coping strategy for dealing with identity threat. Under Cohen and Garcia's (2008) Identity Engagement Model, when students detect cues of potential stereotyping, they engage in a threat appraisal process of evaluating their resources and desire to cope with that threat. The authors emphasized the potential for self-affirmation interventions to help threatened students succeed by reinforcing their self-worth, but utility-value interventions can also have a protective effect for groups marginalized in STEM (Harackiewicz et al., 2016b). Given that utility-value strategies help students most when they discover value for themselves (Canning and Harackiewicz et al., 2015), instructors might support belonging by building into the curriculum opportunities for students to reflect on the value of a course for their own personal goals.
The connection between belonging and interest, particularly for marginalized groups, is consistent with prior research. One survey-based study found a robust relationship between course belonging and positive emotional engagement for all students in STEM courses (Wilson et al., 2015), while another interview-based study showed personal interest as a key influence on students’ STEM belonging (Rainey et al., 2018). Although women did not show an outsized emphasis on interest in the current study, a recent gender-focused experimental study provides insight into the link between belonging and interest for marginalized groups. Master et al. (2016) found that lower belonging, triggered by gender stereotypical vs. non-stereotypical classrooms, contributed to lower interest in computer science among high school girls than boys. Thus, stereotype threat can adversely affect both belonging and interest. The current data cannot speak to the causality of the relationship – whether belonging begets interest, or vice versa – but the studies above suggest a bidirectional relationship.
Finally, the Interest-related results among the low-grade group are unsurprising, given the body of research establishing the role of student interest in learning and achievement (Hidi and Harackiewicz, 2000; Hidi and Renninger, 2006). At the same time, the emergence of this connection in students’ belonging responses strengthens the hypothesis that cultivating interest may enhance not only performance, but also belonging in the course. The results in this section, along with the importance of the Cognitive and Affective Engagement code among all study participants (i.e., it appears in 40–50% of responses), underscores the potential for interest interventions to benefit student outcomes (Harackiewicz et al., 2016a), including course belonging. Harackiewicz and colleagues offer numerous suggestions for how instructors can cultivate student interest, including novel and hands-on activities, personalization of content/examples to actual student interests, and problem-based instruction.
Limitations
As with any single-institution study, our results may be limited in generalizability, specific to similar research universities and quantitative STEM courses (i.e., high enrollment, lecture-based, and introductory level). Because belonging is at least partially context-dependent, instructors should formally (e.g., with surveys) or informally (e.g., through class discussion) explore what sources of belonging are most salient for their own students. Despite this limitation, the results of this study illuminate a complex array of reasons why students perceive themselves as belonging in general chemistry – or not. They also demonstrate the unique belonging perceptions of students potentially marginalized or discouraged by their learning environment, offering strategies to support belonging by creating a more inclusive and learning- (vs. performance) oriented course context.Another common concern is participants’ self-selection. However, over 90% of general chemistry students consented to contribute their course data to the education research repository. Among first-year students who provided consent, 93% of those enrolled in GC1 provided at least one belonging survey response, as did 86% of those enrolled in GC2. Therefore, the current findings may not fully represent the student population, but the volume of data in this study (over 3000 belonging responses) increases confidence that the results reflect much of the variation in first-years’ perceptions.
Last, this study is subject to some methodological limitations. First, the open-response prompt that solicited students’ belonging explanations appeared after the quantitative belonging items, which may have influenced the language students used. Another potential issue was the difficulty of establishing IRR; lower reliability for a handful of infrequent codes is a threat to the validity of the qualitative codebook. However, as noted in the Methods, we took steps to mitigate this threat, including counterbalancing the response coding and testing predictions for how the codebook should function (Appendix 4). Next, the demographic categories used in the variation analyses were coarse-grained measures, which may have oversimplified or overlooked aspects of students’ gender, racial, and ethnic identities. The demographics were also limited in scope and did not address other aspects of identity that could influence belonging in STEM, e.g., sexual orientation, ability status, and socio-economic status. Furthermore, the demographic analyses examined each variable in isolation rather than looking at their interactions, which may have obscured variation across marginalized groups (e.g., Van Dusen et al., 2021). A final limitation is the exploratory nature of this study. Although significance tests were employed to compare groups, the tests were not driven by a priori predictions. Instead, the threshold of α = 0.05 was used to narrow the discussion and hone in on patterns worth further investigation. The results of this study should therefore be treated as tentative, generating hypotheses about what factors most strongly influence students’ course-level belonging in general chemistry and what strategies might enhance the belonging of marginalized or discouraged groups.
Conclusions
Through mixed-methods analysis of two-years’ worth of written belonging responses from first-year general chemistry students, this study explored what factors underlie students’ sense of belonging in the course. Overall, students transitioning into college reported a wide range of belonging influences, and the top factors remained quite stable throughout the year. Anywhere from 40–50% of responses at any given time mentioned the theme of cognitive and affective engagement; career goals, course content, and peers all remained top-five themes throughout the year. Variation analyses revealed that certain themes were particularly prominent among marginalized or discouraged groups, including those who face systemic bias in STEM education (e.g., Black and Hispanic students, women) and those who may feel devalued by a selective institutional culture focused on grade-based achievement (e.g., students with no credit-bearing AP scores, low course grades, or high belonging uncertainty). The belonging explanations of marginalized students were disproportionately likely, relative to privileged comparison groups, to make social comparisons and discuss social interactions with peers, self-evaluate their ability and learning in the course, and mention the utility-value of the course for their long-term goals. Marginalized groups were also less likely to describe positive feelings of situational interest and enjoyment in the course. These findings suggest that instructors might strengthen students’ belonging in general chemistry through inclusive strategies that cultivate positive perceptions of others and self and interventions to stoke student interest and engagement. Given the stability of the top sources of belonging across the two-semester sequence, these strategies may remain relevant and continue to support student belonging and success throughout the year.Conflicts of interest
There are no conflicts to declare.Appendix 1. Relationships among grouping variables
Fig. 4 and 5 below illustrate the associations among several grouping variables, including gender, BU group, grade group, and AP group. The number of participants assigned to any give group differed across the four surveys (i.e., early- and late-semester in GC1 and GC2), depending how many students responded at each time. Given these different sub-samples, caution should be exercised in making direct longitudinal comparisons. Nonetheless, these figures show students’ designation as marginalized/discouraged or privileged group members depends on the specific student characteristic under investigation and how these characteristics can interact with one another. | ||
| Fig. 4 Composition of belonging uncertainty (BU) groups in terms of gender and performance. Confident and uncertain students (x-axis) had mean BU scores ≤2 and ≥5, respectively, on each survey (panels = course, rows = time). Grade groups distinguish low course grades C+ through D (red) from high course grades A through B+ (green). Distributions also differ by gender (columns). | ||
 | ||
| Fig. 5 Composition of belonging uncertainty (BU) groups in terms of gender and Advanced Placement (AP) scores. Confident and uncertain students (x-axis) had mean BU scores ≤2 and ≥5, respectively, on each survey (panels = course, rows = time). AP groups distinguish students with no AP exam scores ≥4 (red) vs. with 3–4 AP scores ≥4 (green). Distributions also differ by gender (columns). | ||
For instance, one might expect performance and belonging uncertainty to be confounded, such that the low-grade group and the uncertain group comprise the same students. However, Fig. 4 indicates the two variables are not redundant with one another: some students with high grades struggle with high belonging uncertainty, especially among women. In the same vein, one might also expect a confound between AP experience/success and belonging uncertainty. Yet once again, we observe crossover between these variables, such that some students with no AP exam scores nonetheless report confidence and, to a lesser extent, students with several high AP exam scores report uncertainty (Fig. 5). These complex relationships motivated the retention of all these grouping variables in this investigation, and their interaction may be an informative topic for future research.
Appendix 2. Codebook development
On the belonging survey, students were asked not only about belonging in General Chemistry, but also about their “STEM or prehealth field.” Field-level results are not discussed in this study; however, the field-level responses were included during thematic codebook development, because we aimed to create a single manual that could encompass both levels of belonging. In addition, codebook development incorporated responses from students in a second STEM course, Introductory Physics. Like the field-level responses from General Chemistry, the physics responses are not reported in this study.Codebook development began with J. D. Y. examining a preliminary sample of 80 written responses from General Chemistry. This sample represented 10 randomly chosen participants, who responded to the course- and field-level prompts (2 items) at the beginning and end of both General Chemistry courses (4 times). J. D. Y. generated an exhaustive list of over 100 themes found in those responses. J. D. Y. and A. F. then met to discuss the themes, eliminating or combining ideas that seemed redundant or overly specific. Next, J. D. Y. analyzed another sample of 80 chemistry responses from 10 new random participants. She applied the existing themes to the new responses, noting any that did not reappear and creating new themes as necessary. Once again, J. D. Y. and A. F. discussed the themes, how to streamline them, and how to formalize them into codes, producing an initial set of 33 codes.
At this point, the coding team began testing agreement when multiple individuals independently coded additional, randomly-selected responses for the presence or absence of each code. Importantly, responses could be assigned several codes, because students often discussed multiple themes in their belonging explanations. At the same time, each code was marked only once for a given response; multiple mentions of the same theme were not tallied up. Besides checking for coding discrepancies, the team also examined the frequency of codes, eliminating some that were uninformative.
First, J. D. Y. and A. F. coded a round of approximately 300 course- and field-level responses from General Chemistry. Next, A. Y. was trained on the codebook and joined J. D. Y. and A. F. in coding another chemistry-only sample of 150 responses. Finally, a team of five coders (the three above plus N. V. and M. J. C.) reviewed the codebook together and independently coded a total of 300 Introductory Physics responses over three rounds. After each iteration of coding, the team members discussed their coding decisions, resolved any discrepancies, and debated whether the codebook needed further refinement. While R. F. F. did not participate in coding, she consulted with the team during codebook development, lending her perspective as a General Chemistry instructor and qualitative researcher to the discussions. At the end of this process, a final list of 21 codes grouped into 6 categories had been created (see Table 1 or Appendix 3 for streamlined and detailed codebooks, respectively).
Appendix 3. Detailed codebook
All responses should be interpreted and coded relative to the question prompt:Course-Level Question
1a. _____ Setting aside my performance in class, I feel like I belong in the [insert course name] course. 〈1 = strongly disagree, 6 = strongly agree〉
1b. Please explain why you agree or disagree with this statement.
Field-level Question
2a. _____ Setting aside my performance in class, I feel like I belong in my STEM or prehealth field. 〈1 = strongly disagree, 6 = strongly agree〉 [Note: In Physics, students who reported a non-STEM major rated their belonging in “my field of interest”]
2b. Please explain why you agree or disagree with this statement.
| Category | Code | Description | Example(s) |
|---|---|---|---|
| Note: within the code column, the numbers in parentheses (x,x) indicate whether individual codes reflect only presence/absence of a theme or also polarity. The majority of codes are marked present (1) or absent (0), but seven are marked present and positive (1), present and negative/neutral (−1), or absent (0). | |||
| Course and field attributes | Content/skills (0, 1) | Comments on the topics or skills (e.g., scientific thinking) covered within the target STEM course(s) or courses within the target field. Exclude statements that might refer to learning outside of courses, in research or other experiences (e.g., exclude “everything I learn in Education is something I care so much about”) | • “I really enjoy the material covered in the course” |
| • “The subjects covered in STEM appeal to me” | |||
| • “…the technical knowledge I gain from classes” | |||
| • “I like the problem solving thinking in the class and the analytical thinking.” | |||
| Difficulty (0, 1) | States whether the target STEM course(s) or target field presents a challenge, or describes how the course/field is demanding or at an appropriate level of difficulty. This code focuses on student perceptions of the course or field, not perceptions of themselves. However, some statements may also be coded for Self-Efficacy or Metacognition. | • “General Chemistry 1 really pushes me to study harder… and gives me a good challenge” | |
| • “I definitely find this class very interesting even though its hard” | |||
| • “these classes are draining” | |||
| • “[this class] is a good level for me” | |||
| • “[the class] assumed people have been exposed to these concepts… so I feel like I am playing catch up” | |||
| • “…this new material in Physics 198 is really difficult for me to understand” | |||
| Environment (−1, 0, 1) | Describes the overall atmosphere of the target STEM course(s) or field and how that atmosphere affects themselves or others. Conveys a positive (e.g., inclusive, supportive) or negative climate (e.g., inequitable, competitive), which may be linked to surrounding peers, instructors, etc. Avoid vague examples where it is unclear if the environment is responsible for students’ perceptions (e.g., exclude “I feel comfortable in lectures”). | • “…I feel that the class is supportive…” | |
| • “The course is very welcoming and makes all feel involved” | |||
| • “I feel like everyone is working together” | |||
| • “I like being surrounded by people who share similar goals and interests as me” | |||
| • “The instructor is very inviting and does not make me feel inferior or ignorant when I ask questions” | |||
| Structure (0, 1) | Evaluates element(s) of the course or program design. Talks about the value or quality of specific policies (e.g., course goals, attendance), procedures (e.g., grading), teaching practices (e.g., clickers), or resources (e.g., help sessions, PLTL) in the target STEM course(s) or field. Merely noting the type of course (e.g., intro) or mentioning an activity is insufficient. | • “courses with specific focuses and in-depth coverage of a limited number of ideas or concepts” | |
| • “there are so many resources available” | |||
| • “I find the presentation of material partial to my learning style” | |||
| Interpersonal | Peers (0, 1) | Mentions other students in the target STEM course(s) or field, including friends, groups of students, or the whole class. | • “I have good friends in the classes” |
| • “I belong with my other MechEs” | |||
| • “I think everyone struggles to transition” | |||
| Professors (0, 1) | Mentions instructors in the target STEM course(s) or field. | • “the professors aim to help me find long-term success” | |
| • “I find the people/professors enjoyable to be around/learn from” | |||
| Student attributes | Extracurricular (0, 1) | Describes non-course experiences like volunteering or research that enrich learning or foster feelings of inclusion in the target STEM course(s) or field. | • “I love the research I do and am strongly interested in it” |
| • “I volunteer in clinical settings, and I have shadowed professionals in healthcare.” | |||
| Identity (0, 1) | References demographic identity (e.g., their own or others’ gender or race or ethnicity). | • “there are obvious barriers of being a female…” | |
| • “As a possible computer science transfer and black woman…” | |||
| • “…and also really fear being the stereotypical stupid girl in my classes” | |||
| Major (0, 1) | Talks about their own or others’ (possible) majors. This code applies more often to course-level than field-level responses. | • “I am not majoring in a science” | |
| • “many pre-meds major in Anthro…” | |||
| • “I need to take it for my major” | |||
| • “I feel like a pretty typical STEM major” | |||
| • “the core material of my Biology degree… is very interesting to me” | |||
| Preparation (0, 1) | Describes readiness for the target STEM course(s) or field, potentially referencing high school or previous college classes. Avoid vague examples where the nature of preparation is unclear (e.g., exclude “I have been doing this for many years”) | • “I took accel chem and AP chem” | |
| • “Chemistry in high school wasn't challenging” | |||
| • “I have taken a lot of STEM classes” | |||
| Interest | Behavioral engagement (−1, 0, 1) | Describes a personal willingness to put forth effort for a goal or to actively perform specific behaviors to achieve that goal. Mention of goal without effort is not sufficient (e.g., exclude “Med school… is an important goal of mine”), nor is mention of effort without a goal (e.g., exclude “I have worked hard and feel I belong”). | • “I worked hard to get to this school” |
| • “I feel I have the dedication and mental stamina to continue my path in [STEM]” | |||
| • “We all strive for our personal best in academic achievement” | |||
| • “I feel like I have the work ethic to learn the material deeply and well” | |||
| • “I have taken the right steps towards becoming [a doctor]” | |||
| • “I don’t necessarily feel like I need to (or should…) kill myself to get an A” | |||
| Cognitive/affective engagement (−1, 0, 1) | Describes personal feelings about the target STEM course(s) or field and conveys (lack of) fondness, enthusiasm, or curiosity. Avoid statements about wants, especially those relating to future outcomes rather than current experiences (e.g., exclude “I want to be in STEM”, “I want to be a doctor”). Neutral examples (e.g., “I’m pretty indifferent”) coded negatively. | • “I enjoy learning the material…” | |
| • “It's the most exciting field to me” | |||
| • “one of the most intriguing and fascinating topics” | |||
| • “I really do love science and the content is interesting” | |||
| • “I do not like physics” | |||
| • “I’m not interested in it at all” | |||
| Individual interest (−1, 0, 1) | Describes (lack of) “passion” for the target STEM course(s) or field. Reflects persistent, well-developed individual interest, rather than situational or emergent interest that may not continue beyond the current context. Negatively coded examples convey neutrality, dispassion, or apathy, rather than strong negative emotions. | • “I've always been passionate about STEM…” | |
| • “I have also dreamed of being a pediatric surgeon” | |||
| • “I have always enjoyed physics regardless of how…grades go” | |||
| • “My interests in science, patient experience, and general healthcare push me to pursue the prehealth field” | |||
| • “I think it was a good challenge but I didn't really feel passion about it.” | |||
| Value | Career (0, 1) | Discusses long-term, professional plans, including general aims (e.g., helping others/society) and more concrete ones (e.g., becoming a doctor). This code applies to mentions of pre-health and pre-med tracks or other “fields,” which indicate students are thinking about their post-college plans. Include vague responses (e.g., include “it's the only thing I can see myself doing”) only for STEM/field-level responses. | • “I want to be a doctor” |
| • “This is what I want to do with my life” | |||
| • “Because I am passionate about my prehealth field” | |||
| • “It is the only field I can imagine myself being in” | |||
| Performance (0, 1) | Talks about course grades, performance, success or results. This code can apply whether the student is grade-oriented or not. | • “When I don’t do well it's hard to take” | |
| • “My grades are not as high as I want” | |||
| • “my performance in class has no effect” | |||
| • “I am not struggling to pass” | |||
| Requirement (0, 1) | Mentions that the course fulfills a requirement or is necessary for long-term goals. | • “I need the class to get where I want” | |
| • “I want to pursue a premed career and part of that process is Gen Chem” | |||
| • “…into a field that requires this course” | |||
| Utility (−1, 0, 1) | Expresses a belief or feeling that the target STEM course(s) and/or its contents will be applicable, relevant, or useful for future endeavors. | • “…greatly applies to what I plan to study in the coming years” | |
| • “I don’t’ think it is beneficial to my future avenues of research” | |||
| Perceptions | Peer comparison (0, 1) | Makes a comparison between themselves and their peers, highlighting a difference or similarity. | • “…people with similar goals and plans” |
| • “I feel like other students are similarly motivated to me” | |||
| • “this common goal makes me feel like I belong” | |||
| Self-efficacy (−1, 0, 1) | Explicitly conveys (lack of) belief or confidence in their ability to succeed in the target STEM course(s) or field. Student judges whether they have the competence to perform the tasks required for successful completion of the course or entrance into the field. | • “I know I have the skills to perform well in class” | |
| • “I feel like I can… keep up with material” | |||
| • “I feel lost in lecture… I don’t feel as though my calc 2 skills are up to par for the class” | |||
| • “I am great at STEM” | |||
| • “[this class] is a good level for me” | |||
| Uncertainty (0, 1) | Conveys that the student is currently unsure of their belonging in the target STEM course(s) or field, or about their decision to pursue this path (e.g., currently unsure about major or career). Exclude cases where students convey performance anxiety (e.g., “I am nervous for the next exam”). | • “I still have doubts about my prehealth field” | |
| • “Sometimes I feel a bit uneasy if I should belong in prehealth, but I do feel fairly comfortable amongst my STEM peers” | |||
| • “A shared interest makes me feel like I belong. Sometimes my performance makes me feel like I don’t belong” | |||
| Metacognition (−1, 0, 1) | Reflects on their own comprehension or learning in the target STEM course(s) or field. Demonstrates self-awareness of their understanding or learning, potentially identifying strengths and weaknesses. May convey that learning is more important than performance. May also be coded for Self-Efficacy, if understanding is linked to success in the course or field. | • “The teachers helped me a lot to understand new material” | |
| • “Even if my grades may or may not reflect it…I feel as if I have learned” | |||
| • “That said, I do comprehend it” | |||
| • “…sometimes the concepts and equations don’t connect for me” | |||
| • “No matter how much I study this information does not click with me” | |||
| • “…this new material in Physics 198 is really difficult for me to understand” | |||
| • “I’m not a very math oriented person” | |||
| • “I feel lost in lecture… I don’t feel as though my calc 2 skills are up to par for the class” | |||
Appendix 4. Validation of coding scheme
Prior to conducting the primary analyses for this study, the research team used the early-semester survey data from General Chemistry 1 to examine whether the coding scheme behaved as expected, i.e., whether the data supported predictions that certain codes would relate to certain student characteristics. All early-semester, course-level belonging responses from first-year students enrolled in General Chemistry 1 during the study period were incorporated into this brief validity assessment (N = 990). The analyses addressed the following two questions:A. Does the frequency of the Preparation and Self-Efficacy codes vary according to students’ prior success with college-level coursework, indexed by AP exam scores?
B. Does the Uncertainty code increase in frequency as belonging uncertainty scores increase?
4A. Associations of codes with AP success
In descriptive terms, the Preparation code was more frequent among students with high AP proportion scores (0.75–1) compared to those with low or moderate scores (0–0.5; see Table 2). Logistic regression confirmed an overall effect of AP proportion on the presence vs. absence of the Preparation code (likelihood ratio test: χ2(4) = 9.91, p = 0.04), though Tukey tests revealed no significant pairwise comparisons between students with different AP proportion scores. The data suggest that high-school academic preparation is a more salient cue to belonging for students with more success on AP exams, while students with less AP success may derive belonging from other sources. The Preparation code was especially prominent among the small contingent of students (N = 48) with AP proportion scores of 1, indicating exemplary scores (4–5 out of 5) on all 4 STEM AP exams (Biology, Calculus, Chemistry, Physics). Almost one third (31.2%) of students with the maximum level of AP exam success were assigned the Preparation code.| AP proportion score | N | Preparation frequency (%) |
|---|---|---|
| Note: AP proportion scores ranged in value from 0 to 1, increasing 0.25 points for every AP STEM exam where a student earned a score of 4 or 5 (out of 5). | ||
| 0 | 317 | 21.1 |
| .25 | 190 | 17.4 |
| .5 | 267 | 17.2 |
| .75 | 168 | 26.8 |
| 1 | 48 | 31.2 |
In terms of the Self-Efficacy code, which incorporated positive and negative polarity coding, the validity analysis explored whether positive mentions of the Self-Efficacy code corresponded with more AP exam success and negative mentions with less AP exam success. Fig. 6 represents the early-semester GC1 responses of all first-year participants whose responses were tagged for Self-Efficacy (N = 177/17.9% of responses). As predicted, the proportion of responses marked for positive Self-Efficacy increased in tandem with AP proportion score. Among students who reported no high scores on AP exams (AP proportion = 0), more than one quarter of Self-Efficacy responses included negative comments (28.8%; negative = 23.7%, both positive and negative = 5.1%; Fig. 6). In contrast, students who reported extensive AP exam success almost exclusively expressed positive self-efficacy: 94.4% and 100% of the Self-Efficacy responses from students with AP proportion scores of 0.75 or 1, respectively, were positive.
 | ||
| Fig. 6 Polarity of early GC1 Self-Efficacy coding by AP exam success. AP proportion (x-axis) = 0 indicates no STEM AP exam scores of 4–5 (of 5); AP proportion = 1 indicates exemplary scores on all 4 STEM AP exams. Sub-sample sizes (Ns) above the bars show how many participants in each AP bin were assigned the Self-Efficacy code on the early GC1 survey. Proportion (y-axis) shows how many instances of Self-Efficacy were positive or negative in direction, or both (see legend for color coding). | ||
4B. Uncertainty association
The Uncertainty code applied to belonging explanations where students described feeling unsure about their belonging in the course, their major, or field. It was relatively infrequent, occurring in only 5.2% of early-semester GC1 responses. Nonetheless, this qualitative code was still expected to show a meaningful relationship with student characteristics. In particular, we expected the Uncertainty code to increase in frequency as students’ quantitative BU scores increased. That positive association was confirmed by logistic regression (likelihood ratio test: χ2(1) = 18.78, p < 0.001) and is illustrated by the descriptive data in Table 3. While the overall frequency of the Uncertainty code was low (5.2%), the table shows the code occurred more frequently among students with BU scores at or above the mid-point of the scale (3.5), peaking at 22.6% of responses from students near the top of the BU scale (5.5). This pattern supports the convergent validity of the qualitative codebook.| Belonging uncertainty score | N | Uncertainty frequency (%) |
|---|---|---|
| 1 | 40 | 0 |
| 1.5 | 61 | 1.6 |
| 2 | 139 | 2.2 |
| 2.5 | 117 | 4.3 |
| 3 | 160 | 3.1 |
| 3.5 | 146 | 6.2 |
| 4 | 132 | 6.8 |
| 4.5 | 81 | 7.4 |
| 5 | 64 | 9.4 |
| 5.5 | 31 | 22.6 |
| 6 | 10 | 0 |
Appendix 5. Complete list of overall frequencies from GC1 and GC2
| Code | General Chemistry 1 | General Chemistry 2 | ||
|---|---|---|---|---|
| Early | Late | Early | Late | |
| Note: codes listed in alphabetical order.a Top five codes at each time point. | ||||
| N | 990 | 833 | 753 | 679 |
| Behavioral engagement | 5.2 | 5.2 | 6.0 | 6.5 |
| Career | 23.3 | 14.4 | 19.7 | 19.0 |
| Cognitive/affective engagement | 41.6 | 51.7 | 51.5 | 49.9 |
| Content | 20.1 | 34.6 | 30.1 | 29.3 |
| Difficulty | 9.4 | 15.5 | 8.6 | 10.2 |
| Environment | 6.9 | 8.5 | 4.1 | 4.4 |
| Extracurriculars | 0.3 | 0.2 | 0.0 | 0.0 |
| Identity | 0.5 | 0.2 | 0.5 | 0.1 |
| Individual interest | 2.5 | 2.0 | 3.2 | 2.9 |
| Major | 9.9 | 7.2 | 7.7 | 7.2 |
| Metacognition | 12.4 | 14.2 | 8.9 | 11.5 |
| Peer comparison | 9.9 | 9.7 | 8.8 | 7.1 |
| Peers | 19.7 | 18.0 | 16.1 | 15.8 |
| Performance | 12.7 | 12.4 | 12.0 | 12.1 |
| Preparation | 20.8 | 5.5 | 5.2 | 2.5 |
| Professors | 6.9 | 5.9 | 2.8 | 5.6 |
| Requirement | 14.2 | 11.0 | 13.1 | 12.4 |
| Self-efficacy | 17.9 | 14.9 | 11.7 | 10.3 |
| Structure | 7.2 | 10.1 | 5.3 | 7.5 |
| Uncertainty | 5.2 | 4.4 | 3.3 | 4.0 |
| Utility | 7.7 | 5.8 | 4.9 | 5.7 |
Appendix 6. Significant results from the variation analyses
For brevity, the tables below include only codes that showed at least one significant difference for the grouping variable under investigation. For the race/ethnicity analysis, codes with marginal main effects are also included, because post hoc pairwise comparisons could reveal significant pairwise differences (Tables 5–9).| Code | General Chemistry 1 | General Chemistry 2 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Early | Late | Early | Late | ||||||
| F | M | F | M | F | M | F | M | ||
| Note: * p < 0.05 for comparison by sex. | |||||||||
| N | 559 | 431 | 477 | 356 | 441 | 312 | 398 | 281 | |
| Cognitive/affective engagement | 42.8 | 40.1 | 52.6 | 50.6 | 54.9 | 46.8* | 51.0 | 48.4 | |
| Major | 10.4 | 9.3 | 9.2 | 4.5* | 8.6 | 6.4 | 7.8 | 6.4 | |
| Peer comparison | 11.3 | 8.1 | 13.0 | 5.3* | 10.9 | 5.8* | 8.3 | 5.3 | |
| Peers | 21.3 | 17.6 | 21.8 | 12.9* | 18.8 | 12.2* | 18.1 | 12.5 | |
| Performance | 12.2 | 13.5 | 14.7 | 9.3* | 12.5 | 11.2 | 13.3 | 10.3 | |
| Structure | 8.4 | 5.6 | 12.4 | 7.0* | 4.3 | 6.7 | 7.3 | 7.8 | |
| Code | General Chemistry 1 | General Chemistry 2 | ||||||
|---|---|---|---|---|---|---|---|---|
| Early | Late | Early | Late | |||||
| No AP | High AP | No AP | High AP | No AP | High AP | No AP | High AP | |
| Note: * p < 0.05 for comparisons by AP group. | ||||||||
| N | 317 | 216 | 254 | 194 | 220 | 182 | 202 | 160 |
| Career | 23.3 | 19.9 | 18.5 | 9.3* | 22.7 | 18.1 | 25.7 | 13.1* |
| Difficulty | 10.4 | 9.7 | 18.9 | 13.9 | 5.5 | 8.8 | 11.4 | 5.0* |
| Peer comparison | 11.4 | 8.8 | 13.0 | 4.1* | 9.1 | 4.4 | 6.9 | 1.9* |
| Peers | 19.2 | 19.0 | 22.4 | 10.8* | 14.5 | 12.1 | 13.4 | 10.0 |
| Self-efficacy | 18.6 | 22.2 | 19.3 | 11.3* | 10.0 | 11.5 | 8.9 | 10.0 |
| Code | General Chemistry 1 | General Chemistry 2 | ||||||
|---|---|---|---|---|---|---|---|---|
| Early | Late | Early | Late | |||||
| Low | High | Low | High | Low | High | Low | High | |
| Note: * p < 0.05 for comparisons by course grade group. | ||||||||
| N | 167 | 494 | 135 | 443 | 64 | 441 | 65 | 394 |
| Behavioral engagement | 4.9 | 5.0 | 7.7 | 2.8* | 11.3 | 5.8 | 10.9 | 6.0 |
| Career | 30.5 | 20.9 | 22.2 | 12.2* | 32.8 | 17.0* | 26.2 | 14.5* |
| Cognitive/affective engagement | 35.3 | 45.1* | 45.9 | 54.9 | 40.6 | 57.1* | 38.5 | 54.8* |
| Difficulty | 6.6 | 9.3 | 20.7 | 14.0* | 6.3 | 8.8 | 9.2 | 7.4 |
| Individual interest | 4.3 | 2.3 | 3.1 | 1.6 | 6.5 | 2.3 | 9.4 | 1.8* |
| Peers | 21.6 | 19.0 | 17.0 | 20.1 | 10.9 | 19.0 | 7.7 | 19.0* |
| Performance | 12.6 | 13.2 | 23.7 | 6.3* | 14.1 | 10.2 | 16.9 | 11.7 |
| Preparation | 23.4 | 21.7 | 8.1 | 5.0 | 12.5 | 4.3* | 3.1 | 2.5 |
| Requirement | 18.6 | 12.6 | 11.9 | 9.3 | 18.8 | 11.8 | 21.5 | 9.1* |
| Self-efficacy | 18.0 | 16.6 | 23.0 | 11.5* | 25.0 | 11.1* | 21.5 | 9.4* |
| Code | General Chemistry 1 | General Chemistry 2 | ||||||
|---|---|---|---|---|---|---|---|---|
| Early | Late | Early | Late | |||||
| Low BU | High BU | Low BU | High BU | Low BU | High BU | Low BU | High BU | |
| Note: * p < 0.05 for comparisons by BU group. | ||||||||
| N | 240 | 105 | 254 | 114 | 243 | 56 | 231 | 94 |
| Behavioral engagement | 4.6 | 1.0 | 3.9 | 5.3 | 6.2 | 3.6* | 3.0 | 10.6 |
| Career | 22.5 | 36.2* | 14.2 | 16.7 | 23.9 | 19.6 | 14.3 | 26.6* |
| Cognitive/affective engagement | 45.4 | 32.4* | 55.9 | 46.5 | 49.8 | 37.5 | 56.7 | 42.6* |
| Difficulty | 10.4 | 2.9* | 13.8 | 16.7 | 8.2 | 10.7 | 8.7 | 11.7 |
| Individual interest | 2.5 | 1.9 | 2.4 | 3.5 | 2.1 | 3.6 | 0.9 | 8.5* |
| Peer comparison | 8.3 | 16.2* | 7.9 | 11.4 | 7.4 | 7.1 | 5.6 | 10.6 |
| Performance | 10.0 | 20.0* | 5.9 | 23.7* | 9.9 | 14.3 | 10.4 | 12.8 |
| Structure | 5.4 | 3.8 | 9.8 | 5.3 | 6.6 | 5.4 | 7.4 | 1.1* |
| Uncertainty | 1.7 | 12.4* | 1.2 | 14.0* | 2.5 | 10.7* | 1.3 | 11.7* |
| General Chemistry 1 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Early | Late | |||||||||
| Code | Asian | Black | Hispanic | White | Pairwise Comparisons | Asian | Black | Hispanic | White | Pairwise Comparisons |
| Note: logistic regressions modeled the effect of race/ethnicity on each code's frequency, with type III likelihood-ratio chi-squared tests to assess the significance of the main effect. Bold font indicates significant (p < 0.05) main effects of race/ethnicity; italic font indicates marginal (0.5 > p > 0.10) main effects. Tukey-adjusted post hoc tests assessed the significance of each pairwise comparison. * p < 0.05, † 0.5 > p > 0.10. | ||||||||||
| N | 351 | 87 | 84 | 418 | 301 | 76 | 73 | 340 | ||
| Career | †19.7 | 33.3 | 26.2 | 23.4 | A & B* | 12.6 | 21.1 | 19.2 | 14.4 | |
| Cognitive/affective engagement | 42.2 | 35.6 | 44.0 | 41.4 | 50.8 | 43.4 | 45.2 | 53.8 | ||
| Content | 23.9 | 17.2 | 22.6 | 18.2 | 36.5 | 30.3 | 34.2 | 33.8 | ||
| Environment | *9.7 | 9.2 | 4.8 | 4.5 | A & W* | 8.6 | 9.2 | 9.6 | 7.9 | |
| Individual interest | 2.0 | 3.4 | 3.6 | 2.9 | 2.0 | 1.3 | 4.1 | 2.1 | ||
| Major | 7.4 | 9.2 | 9.5 | 12.4 | *4.3 | 5.3 | 8.2 | 10.3 | A & W* | |
| Metacognition | 12.0 | 16.1 | 16.7 | 12.4 | *15.0 | 22.4 | 19.2 | 10.9 | B & W* | |
| Peers | 22.5 | 18.4 | 15.5 | 18.4 | 16.6 | 19.7 | 20.5 | 17.6 | ||
| Professors | *12.0 | 5.7 | 4.8 | 3.6 | A & W* | 8.0 | 2.6 | 5.5 | 5.3 | |
| Requirement | †10.8 | 21.8 | 14.3 | 14.8 | A & B* | 9.0 | 13.2 | 13.7 | 12.6 | |
| General Chemistry 2 | ||||||||||
| N | 270 | 65 | 67 | 316 | 237 | 64 | 62 | 283 | ||
| Career | 17.4 | 26.2 | 20.9 | 19.9 | 13.9 | 23.4 | 25.8 | 20.5 | ||
| Cognitive/affective engagement | *57.4 | 41.5 | 52.2 | 48.1 | A & B† | †53.2 | 34.4 | 51.6 | 49.1 | A & B* |
| Content | 30.4 | 18.5 | 26.9 | 32.9 | *32.1 | 29.7 | 14.5 | 29.3 | A & H*, H & W† | |
| Environment | 6.3 | 3.1 | 1.5 | 2.8 | 3.8 | 4.7 | 1.6 | 5.7 | ||
| Individual interest | 3.3 | 4.6 | 4.5 | 2.8 | †3.4 | 4.7 | 8.1 | 1.4 | H & W* | |
| Major | *4.8 | 4.6 | 7.5 | 11.7 | A & W* | 6.3 | 1.6 | 9.7 | 9.2 | |
| Metacognition | 8.9 | 13.8 | 6.0 | 8.2 | *13.1 | 20.3 | 4.8 | 9.9 | B & H†, B & W† | |
| Peers | 14.8 | 20.0 | 14.9 | 16.5 | *17.3 | 14.1 | 4.8 | 17.3 | A & H†, H & W† | |
| Professors | 4.1 | 0.0 | 1.5 | 2.5 | 6.8 | 6.3 | 4.8 | 4.2 | ||
| Requirement | 10.4 | 13.8 | 17.9 | 14.6 | 11.0 | 9.4 | 19.4 | 12.7 | ||
Acknowledgements
This research was funded by an Inclusive Excellence grant from the Howard Hughes Medical Institute (Award No. 42008723; PI: Regina Frey) to Washington University in St. Louis. We thank Jia Luo, Rick Schneider, and the General Chemistry Assistants in Instruction for their help in administering the belonging survey, and we are grateful to the General Chemistry instructor team for their support of this research. We also thank Mike Cahill and Ali York for their help developing the qualitative codebook and establishing inter-rater reliability.References
- Aisch G., Buchanan L., Cox A. and Quealy K., (2017), Some colleges have more students form the top 1 percent than from the bottom 60. Find yours, New York Times. [Online] 18th January. Available from https://nyti.ms/2jRcqJs. [Accessed 4/26/2022].
- Akyol G., Sungur S. and Tekkaya C., (2010), The contribution of cognitive and metacognitive strategy use to students' science achievement, Educ. Res. Eval., 16(1), 1–21.
- Aronson J., Fried C. B. and Good C., (2002), Reducing the effects of stereotype threat on African American college students by shaping theories of intelligence, J. Exp. Soc. Psychol., 38(2), 113–125.
- Baumeister R. F. and Leary M. R., (1995), The need to belong: Desire for interpersonal attachments as a fundamental human motivation, Psychol. Bull., 117(3), 497–529.
- Binning K. R., Kaufman N., McGreevy E. M., Fotuhi O., Chen S., Marshman E., Kalendar Z. Y., Limeri L. B., Betancur L. and Singh C., (2020), Changing social contexts to foster equity in college science courses: An ecological-belonging intervention, Psychol. Sci., 31(9), 1059–1070.
- Blackwell L. S., Trzesniewski K. H. and Dweck C. S., (2007), Implicit theories of intelligence predict achievement across an adolescent transition: A longitudinal study and an intervention, Child Dev., 78(1), 246–263.
- Boucher K. L., Fuesting M. A., Diekman A. B. and Murphy M. C., (2017), Can I work with and help others in this field? How communal goals influence interest and participation in STEM fields, Front. Psychol., 8, 901.
- Broda M., Yun J., Schneider B., Yeager D. S., Walton G. M. and Diemer M., (2018), Reducing inequality in academic success for incoming college students: A randomized trial of growth mindset and belonging interventions, J. Res. Educ. Effect., 11(3), 317–338.
- Canning E. A. and Harackiewicz J. M., (2015), Teach it, don’t preach it: The differential effects of directly-communicated and self-generated utility–value information, Motiv. Sci., 1(1), 47–71.
- Canning E. A., Muenks K., Green D. J. and Murphy M. C., (2019), STEM faculty who believe ability is fixed have larger racial achievement gaps and inspire less student motivation in their classes, Sci. Adv., 5(2), eaau4734.
- Canning E. A., LaCosse J., Kroeper K. M. and Murphy M. C., (2020), Feeling like an imposter: The effect of perceived classroom competition on the daily psychological experiences of first-generation college students, Soc. Psychol. Pers. Sci., 11(5), 647–657.
- Carlone H. B. and Johnson A., (2007), Understanding the science experiences of successful women of color: Science identity as an analytic lens, J. Res. Sci. Teach., 44(8), 1187–1218.
- Chang J., Wang S. W., Mancini C., McGrath-Mahrer B. and Orama de Jesus S., (2020), The complexity of cultural mismatch in higher education: Norms affecting first-generation college students’ coping and help-seeking behaviors, Cultur. Div. Ethnic Minor. Psychol., 26(3), 280.
- Chetty R., Friedman J. N., Saez E., Turner N. and Yagan D., (2017), Mobility report cards: The role of colleges in intergenerational mobility (No. w23618), National Bureau of Economic Research, pp. 1–61 DOI:10.3386/w23618.
- Clements T. P., Friedman K. L., Johnson H. J., Meier C. J., Watkins J., Brockman A. J. and Brame C. J., (2022), “It made me feel like a bigger part of the STEM community”: Incorporation of Learning Assistants Enhances Students’ Sense of Belonging in a Large Introductory Biology Course, CBE – Life Sci. Educ., 21(2), ar26.
- Cohen G. L. and Garcia J., (2008), Identity, belonging, and achievement, Curr. Dir. in Psychol. Sci., 17(6), 365–369.
- Cracolice M. S., Deming J. C. and Ehlert B., (2008), Concept learning versus problem solving: A cognitive difference, J. Chem. Educ., 85(6), 873–878.
- Daniels H. A., Grineski S. E., Collins T. W. and Frederick A. H., (2019), Navigating social relationships with mentors and peers: Comfort and belonging among men and women in STEM summer research programs, CBE – Life Sci. Educ., 18(2), ar17.
- Deiglmayr A., Stern E. and Schubert R., (2019), Beliefs in “brilliance” and belonging uncertainty in male and female STEM students, Front. Psychol., 10, 1114.
- Dougherty C., Mellor L. and Jian S., (2006), The Relationship between Advanced Placement and College Graduation. 2005 AP Study Series, Report 1, Texas: National Center for Educational Accountability, pp. 1–15https://eric.ed.gov/?id=ED519365.
- Dweck C. S. and Leggett E. L., (1988), A social-cognitive approach to motivation and personality, Psychol. Rev., 95(2), 256–273.
- Eccles J., (2009), Who am I and what am I going to do with my life? Personal and collective identities as motivators of action, Educ. Psychol., 42(2), 78–89.
- Eddy S. L., Brownell S. E., Thummaphan P., Lan M.-C. and Wenderoth M. P., (2015), Caution, student experience may vary: Social identities impact a student's experience in peer discussions, CBE - Life Sci. Educ., 14(4), 1–17.
- Edwards J. D., Barthelemy R. S. and Frey R. F., (2022), The relationships between course-level belonging (i.e., sense of belonging and belonging uncertainty) and academic performance in General Chemistry 1, J. Chem. Educ., 99(1), 71–82.
- Festinger L., (1954), A theory of social comparison processes, Human Rel., 7(2), 117–140.
- Fink A. F., Cahill M. J., McDaniel M. A., Hoffman A. and Frey R. F., (2018), Improving general chemistry performance through a growth mindset intervention: Selective effects on underrepresented minorities, Chem. Educ. Res. Pract., 19(3), 783–806.
- Fink A., Frey R. F. and Solomon E. D., (2020), Belonging in general chemistry predicts first-year undergraduates' performance and attrition, Chem. Educ. Res. Pract., 21(4), 1042–1062.
- Fink A., Mattson P. S., Cahill M. J., Frey R. F. and McDaniel M. A., (2021), Growth mindset intervention improves performance of highly nontraditional introductory psychology students at a community college, Teach. Psychol., 00986283211053806.
- Freeman T. M., Anderman L. H. and Jensen J. M., (2007), Sense of belonging in college freshmen at the classroom and campus levels, J. Exp. Educ., 75(3), 203–220.
- Frey R. F. and Lewis S. E., (in press, 2022). Implementing Cooperative Learning in Large Undergraduate Classes via PLTL, in Overson C. E., Hakala C. M., Kordonowy L. L. and Benassi V. A., (ed.), In their own words: What scholars want you to know about why and how to apply the science of learning in your academic setting, Society for the Teaching of Psychology (APA, Division 2).
- Frey R. F., Cahill M. J. and McDaniel M. A., (2017), Students' concept-building approaches: A novel predictor of success in chemistry courses, J. Chem. Educ., 94(9), 1185–1194.
- Frey R. F., Fink A., Cahill M. J., McDaniel M. A. and Solomon E. D., (2018), Peer-led team learning in general chemistry I: Interactions with identity, academic preparation, and a course-based intervention, J. Chem. Educ., 95(12), 2103–2113.
- Gere J. and MacDonald G., (2010), An update of the empirical case for the need to belong, J. Indiv. Psychol., 66(1), 93–115.
- Good C., Aronson J. and Inzlicht M., (2003), Improving adolescents' standardized test performance: An intervention to reduce the effects of stereotype threat, J. Appl. Dev. Psychol., 24(6), 645–662.
- Good C., Rattan A. and Dweck C. S., (2012), Why do women opt out? Sense of belonging and women's representation in mathematics, J. Pers. Soc. Psychol., 102(4), 700–717.
- Gutiérrez R., (2008), Research commentary: A gap-gazing fetish in mathematics education? Problematizing research on the achievement gap, J. Res. Math. Educ., 39(4), 357–364.
- Gutiérrez R., (2013), The sociopolitical turn in mathematics education, J. Res. Math. Educ., 44(1), 37–68.
- Gutiérrez R. and Dixon-Román E., (2010), Beyond gap gazing: How can thinking about education comprehensively help us (re) envision mathematics education? Mapping Equity and Quality in Mathematics Education, Dordrecht: Springer, pp. 21–34.
- Harackiewicz J. M., Smith J. L. and Priniski S. J., (2016a). Interest matters: The importance of promoting interest in education, Pol. Insights Behav. Brain Sci., 3(2), 220–227.
- Harackiewicz J. M., Canning E. A., Tibbetts Y., Priniski S. J. and Hyde J. S., (2016b). Closing achievement gaps with a utility-value intervention: Disentangling race and social class, J. Pers. Soc. Psychol., 111(5), 745–765.
- Hidi S. and Harackiewicz J. M., (2000), Motivating the academically unmotivated: A critical issue for the 21st century, Rev. Educ. Res., 70(2), 151–179.
- Hidi S. and Renninger K. A., (2006), The four-phase model of interest development, Educ. Psychol., 41(2), 111–127.
- Hirsch J. L. and Clark M. S., (2019), Multiple paths to belonging that we should study together, Persp. Psychol. Sci., 14(2), 238–255.
- Hockings S. C., DeAngelis K. J. and Frey R. F., (2008), Peer-led team learning in general chemistry: Implementation and evaluation, J. Chem. Educ., 85(7), 990–996.
- Hoehne E. and Zander L., (2019), Sources of male and female students' belonging uncertainty in the computer sciences, Front. Psychol., 10, 1740.
- Jury M., Smeding A. and Darnon C., (2015), First-generation students’ underperformance at university: The impact of the function of selection, Front. Psychol., 6, 710.
- Kirbulut Z. D., (2014), Modeling the relationship between high school students' chemistry self-efficacy and metacognitive awareness, Int. J. Environ. Sci. Educ., 9(2), 177–196.
- Kirppendorff K., (2004), Reliability in content analysis: Some common misconceptions and recommendations, Human Commun. Res., 30(3), 411–433.
- Klugman J., (2013), The Advanced Placement arms race and the reproduction of educational inequality, Teach. Coll. Record, 115(5), 1–34.
- Ladson-Billings G., (2006), From the achievement gap to the education debt: Understanding achievement in US schools, Educ. Res., 35(7), 3–12.
- Ladson-Billings G., (2007), Pushing past the achievement gap: An essay on the language of deficit, J. Negro Educ., 316–323.
- Lewis K. L. and Hodges S. D., (2015), Expanding the concept of belonging in academic domains: Development and validation of the Ability Uncertainty Scale, Learn. Individ. Differ., 37, 197–202.
- Lewis K. L., Stout J. G., Pollock S. J., Finkelstein N. D. and Ito T. A., (2016), Fitting in or optin out: A review of key social-psychological factors influencing a sense of belonging for women in physics, Phys. Rev. Phys. Educ. Res., 12(2), 020110.
- Master A., Cheryan S. and Meltzoff A. N., (2016), Computing whether she belongs: Stereotype undermine girls' interest and sense of belonging in computer science, J. Educ. Psychol., 108(3), 424–437.
- Merriam S. B., (2009), Qualitative research: A guide to design and implementation, San Francisco: Jossey-Bass.
- Micari M. and Drane D., (2011), Intimidation in small learning groups: The roles of social-comparison concern, comfort, and individual characteristics in student academic outcomes, Active Learn. Higher Educ., 12(3), 175–187.
- Micari M. and Pazos P., (2014), Worrying about what others think: A social-comparison concern intervention in small learning groups, Active Learn. Higher Educ., 15(3), 249–262.
- Micari M. and Pazos P., (2019), Small fish in a small pond: The impact of collaborative learning on academic success for less-prepared students in a highly selective STEM environment, Higher Educ. Res. Dev., 38(2), 294–306.
- Micari M. and Pazos P., (2021), Beyond grades: Improving college students' social-cognitive outcomes in STEM through a collaborative learning environment, Learn. Environ. Res., 24(1), 123–136.
- Miller D. M. and Czegan D. A. C., (2016), Integrating the liberal arts and chemistry: A series of general chemistry assignments to develop science literacy, J. Chem. Educ., 93(5), 864–869.
- Muenks K., Canning E. A., LaCosse J., Green D. J., Zirkel S., Garcia J. A. and Murphy M. C., (2020), Does my professor think my ability can change? Students’ perceptions of their STEM professors’ mindset beliefs predict their psychological vulnerability, engagement, and performance in class, J. Exp. Psychol., Gen., 149(11), 2119–2144.
- Murphy M. C., Steele C. M. and Gross J. J., (2007), Signaling threat: How situational cues affect women in math, science, and engineering settings, Psychol. Sci., 18(10), 879–885.
- Murphy M. C., Gopalan M., Carter E. R., Emerson K. T. U., Bottoms B. L. and Walton G. M., (2020), A customized belonging intervention improves retention of socially disadvantaged students at a broad-access university, Sci. Adv., 6(29), eaba4677.
- Mutambuki J. M., Mwavita M., Muteti C. Z., Jacob B. I. and Mohanty S., (2020), Metacognition and active learning combination reveals better performance on cognitively demanding general chemistry concepts than active learning alone, J. Chem. Educ., 97(7), 1832–1940.
- National Center for Science and Engineering Statistics, (2021), Women, Minorities, and Persons with Disabilities in Science and Engineering: 2021, Special Report NSF 21-321, Alexandria, VA: National Science Foundation, available at https://ncses.nsf.gov/wmpd.
- Ogunde J. C., Overton T. L., Thompson C. D., Mewis R. and Boniface S., (2017), Beyond graduation: motivations and career aspirations of undergraduate chemistry students, Chem. Educ. Res. Pract., 18(3), 457–471.
- O'Keeffe P., (2013), A sense of belonging: Improving student retention, Coll. Stud. J., 47(4), 605–613.
- Paunesku D., Walton G. M., Romero C., Smith E. N., Yeager D. S. and Dweck C. S., (2015), Mind-set interventions are a scalable treatment for academic underachievement, Psychol. Sci., 26(6), 784–793.
- Pearson M. I., Castle S. D., Matz R. L., Koester B. P. and Byrd W. C., (2022), Integrating critical approaches into quantitative STEM equity work, CBE – Life Sci. Educ., 21(1), es1.
- Purdie-Vaughns V., Steele C. M., Davies P. G., Ditlmann R. and Crosby J. R., (2008), Social identity contingencies: how diversity cues signal threat or safety for African Americans in mainstream institutions, J. Pers. Soc. Psychol., 94(4), 615.
- Rainey K., Dancy M., Mickelson R., Stearns E. and Moller S., (2018), Race and gender differences in how sense of belonging influence decisions to major in STEM. Intl. J. STEM Educ., 5(1), 10.
- Rodriguez S. L. and Blaney J. M., (2021), “We're the unicorns in STEM”: Understanding how academic and social experiences influence sense of belonging for Latina undergraduate students, J. Div. Higher Educ, 14(3), 441–455.
- Rosenthal L., Levy S. R., London B., Lobel M. and Bazile C., (2013), In pursuit of the MD: The impact of role models, identity compatibility, and belonging among undergraduate women, Sex Roles, 68(7), 464–473.
- Sadler P. M. and Tai R. H., 2007. Advanced placement exam scores as a predictor of performance in introductory college biology, chemistry and physics courses, Sci. Educ., 16(2), 1–19.
- Sax L. J., Blaney J. M., Lehman K. J., Rodriguez S. L., George K. L. and Zavala C., (2018), Sense of belonging in computing: The role of introductory courses for women and underrepresented minority students, Soc. Sci., 7(8), 122.
- Slaten C. D., Christopher D., Elison Z. M., Deemer E. D., Hughes H. A. and Shemwell D. A., (2018), The development and validation of the university belonging questionnaire, J. Exp. Educ., 86(4), 633–651.
- Smeding A., Darnon C., Souchal C., Toczek-Capelle M. C. and Butera F., (2013), Reducing the socio-economic status achievement gap at university by promoting mastery-oriented assessment, PLoS One, 8(8), e71678.
- Stephens N. M., Fryberg S. A., Markus H. R., Johnson C. S. and Covarrubias R., (2012), Unseen disadvantage: How American universities' focus on independence undermines the academic performance of first-generation college students, J. Pers. Soc. Psychol., 102(6), 1178.
- Stephens N. M., Townsend S. S., Markus H. R. and Phillips L. T., (2012), A cultural mismatch: Independent cultural norms produce greater increases in cortisol and more negative emotions among first-generation college students, J. Exp. Soc. Psychol., 48(6), 1389–1393.
- Stephens N. M., Hamedani M. G. and Townsend S. S. M., (2019), Difference matters: Teaching students a contextual theory of difference can help them succeed, Persp. Psychol. Sci., 14(2), 156–174.
- Stout J. G., Ito T. A., Finkelstein N. D. and Pollock S. J., (2012), How a gender gap in belonging contributes to the gender gap in physics participation, AIP Conf. Proc., 1513(1), 402–405.
- Tai R. H., Sadtler P. M. and Loehr J. F., (2005), Factors influencing success in introductory college chemistry, J. Res. Sci. Teach., 42(9), 987–1012.
- Tanner K. D., (2012), Promoting student metacognition, CBE – Life Sci. Educ., 11(2), 113–120.
- The College Board, (2021), AP Program, viewed February, 2021, https://ap.collegeboard.org/.
- Van Dusen B., Nissen J., Talbot R. M., Huvard H. and Shultz M., (2021), A QuantCrit investigation of society's educational debts due to racism and sexism in chemistry student learning, J. Chem. Educ., 99(1), 25–34.
- Veilleux N., Bates R., Allendoerfer C., Jones D., Crawford J. and Floyd-Smith T., (2013), ‘The relationship between belonging and ability in computer science’, 44th ACM Technical Symposium on Computer Science Education, 6–9 March, 2013, pp. 65–70.
- Walton G. M. and Cohen G. L., (2007), A question of belonging: Race, social fit, and achievement, J. Pers. Soc. Psychol., 92(1), 82–96.
- Walton G. M. and Cohen G. L., (2011), A brief social-belonging intervention improves academic and health outcomes of minority students, Science, 331, 1447–1451.
- Walton G. M. and Spencer S. J., (2009), Latent ability: Grades and test scores systematically underestimate the intellectual ability of negatively stereotyped students, Psychol. Sci., 20(9), 1132–1139.
- Walton G. M., Cohen G. L., Cwir D. and Spencer S. J., (2012), Mere belonging: The power of social connections, J. Pers. Soc. Psychol., 102(3), 513–532.
- Walton G. M., Spencer S. J. and Erman S., (2013), Affirmative meritocracy, Soc. Issues Policy Rev., 7(1), 1–35.
- Walton G. M., Logel C., Peach J. M., Spencer S. J. and Zanna M. P., (2015), Two brief interventions to mitigate a “chilly climate” transform women's experience, relationships, and achievement in Engineering, J. Educ. Psychol., 107(2), 468–485.
- Wang Y., Rocabado G. A., Lewis J. E. and Lewis S. E., (2021), Prompts to promote success: Evaluating utility value and growth mindset interventions on general chemistry students' attitude and academic performance, J. Chem. Educ., 98(5), 1476–1488.
- White K. N., Vincent-Layton K. and Villarreal B., (2020), Equitable and inclusive practices designed to reduce equity gaps in undergraduate chemistry courses, J. Chem. Educ., 98(2), 330–339.
- Wilson D., Jones D., Bocell F., Crawford J., Kim M. J., Veilleux N., Floyd-Smith T., Bates R. and Plett M., (2015), Belonging and academic engagement among undergraduate STEM students: A multi-institutional study, Res. Higher Educ., 56, 750–776.
- Wilton M., Gonzalez-Nino E., McPartlan P., Terner Z., Christoffersen R. E. and Rothman J. H., (2019), Improving academic performance, belonging, and retention through increasing structure of an introductory biology course, CBE – Life Sci. Educ., 18(4), ar53.
- Xu X., Villafane S. M. and Lewis J. E., (2013), College students' attitudes toward chemistry, conceptual knowledge and achievement: structural equation model analysis, Chem. Educ. Res. Pract., 14(2), 188–200.
- Yeager D. S., Romero C., Paunesku D., Hulleman C. S., Schneider B., Hinojosa C., Lee H. Y., O’Brien J., Flint K., Roberts A., Trott J., Greene D., Walton G. M. and Dweck C. S., (2016), Using design thinking to improve psychological interventions: The case of the growth mindset during the transition to high school, J. Educ. Psychol., 108(3), 374.
- Yeager D. S., Walton G. M., Brady S. T., Akcinar E. N., Paunesku D., Keane L., Kamentz D., Ritter G., Duckworth A. L., Urstein R., Gomez E. M., Markus H. R., Cohen G. L. and Dweck C. S., (2016), Teaching a lay theory before college narrows achievement gaps at scale, Proc. Natl. Acad. Sci. U. S. A., 113(24), E3341–E3348.
- Yeager D. S., Hanselman P., Walton G. M., Murray J. S., Crosnoe R., Muller C., Tipton E., Schneider B., Hulleman C. S., Hinojosa C. P., Paunesku D., Romero C., Flint K., Roberts A., Trott J., Iachan R., Buontempo J., Yang S. M., Carvalho C. M., Hahn P. R., Gopalan M., Mhatre P., Ferguson R., Duckworth A. L. and Dweck C. S., (2019), A national experiment reveals where a growth mindset improves achievement, Nature, 573, 364–369.
- Yorke M., (2016), The development and initial use of a survey of student 'belongingness', engagement and self-confidence in UK higher education, Assess. Eval. Higher Educ., 41(1), 154–166.
- Zumbrunn S., McKim C., Buhs E. and Hawley L. R., (2014), Support, belonging, motivation, and engagement in the college classroom: A mixed method study, Instr. Sci., 42(5), 661–684.
Footnotes |
| † To begin addressing this inequity, the university recently implemented need-blind admissions. This change took place after the current study was completed. |
| ‡ Ns deviate slightly from Fink et al. (2020), which spanned the same semesters, for two reasons. First, Fink et al. (2020) selected their sample only from students who enrolled in GC1 and then continued (or not) to GC2, because they investigated attrition from the sequence. The current sample selection also included some students who entered the course sequence at GC2. Second, students participated in this research by opting into an educational data repository. Repository participants are recruited from many STEM courses and consent to share data from all their STEM courses. By the time this analysis occurred, a small number of students not included in Fink et al. (2020) had opted into the data repository during a subsequent course, making their GC data newly available. |
| This journal is © The Royal Society of Chemistry 2023 |