Exploring How Holistic Teaching and Institutional Support Relate to Community College STEM Students’ Momentum and Self-Efficacy in Career-Relevant Competencies

Abstract

This study investigates how holistic teaching practices and institutional support at community colleges shape science, technology, engineering, and mathematics (STEM) students’ momentum and self-efficacy in career-relevant competencies. Using survey data from three community colleges, we apply structural equation modeling (SEM) to assess these relationships while accounting for institutional variation using multi-group analysis. Our findings demonstrate that holistic teaching practices are positively associated with students’ curricular, cognitive, and meta-cognitive momentum, indicating that integrated, supportive classroom instruction contributes to sustained engagement and self-regulated learning in STEM pathways. Holistic teaching practices also show a marginal positive relationship with career readiness self-efficacy and professional and interpersonal self-efficacy, with cognitive and meta-cognitive momentum mediating these associations. In contrast, institutional support is not related to students’ momentum but is positively associated with professional and interpersonal self-efficacy, which may point to its role in shaping broader skill development independent of short-term academic engagement. These findings suggest that holistic teaching practices and institutional support differentially contribute to students’ academic momentum and career-related self-efficacy, which highlights the importance of coordinated efforts across classroom and institutional levels within the broader STEM ecosystem in fostering both short-term engagement and long-term professional competencies among diverse community college STEM learners.

1. Introduction

Community colleges play a critical role in the science, technology, engineering, and mathematics (STEM) ecosystem by providing accessible entry points into educational and career pathways for diverse populations, especially students from historically underserved groups across racial, gender, and economic lines (Kisker et al., 2023; Wang, 2020). Yet students pursuing community college STEM pathways continue to encounter intertwined academic, institutional, and structural barriers that influence not only their academic momentum—students’ sustained progress through coursework, engagement with learning, and development of self-regulation skills (Núñez, 2022; Wang, 2017, 2024b)—but also their self-efficacy and development in career-relevant competencies, including professional and interpersonal skills and career self-efficacy that enable community college STEM students to prepare for further education and employment (Deil-Amen, 2006; Myran & Sylvester, 2021; Núñez, 2023). Understanding how community colleges can foster these outcomes is therefore essential for supporting student persistence, transfer, and career advancement in STEM fields (Bahr et al., 2017).

For community college students, the classroom represents the primary and most sustained site of engagement with their educational experiences (Deil-Amen, 2011). Given the purported disciplinary rigor and gatekeeping norms that often characterize STEM instruction (Gasiewski et al., 2012; Seymour & Hewitt, 1997), classroom experiences play a pivotal role in shaping students’ academic momentum and the development of beliefs and competencies necessary for persistence, transfer, and career advancement (Deil-Amen, 2011; Komarraju et al., 2010; Wang, 2020). In particular, holistic, human-centered teaching practices, including those that attend to students’ academic learning while also recognizing the social, emotional, and structural dimensions of their educational experiences, have been shown to positively influence student motivation, persistence, and skill development (Felder & Brent, 1996; Freeman et al., 2014; Wang et al., 2017). Such practices shape how students experience course expectations, assessment, and instructional relationships, and signal whether classrooms are designed to support learning alongside students’ broader life circumstances. In STEM classrooms, where rigid norms and high-stakes evaluation can disproportionately undermine students from historically underserved groups, holistic teaching practices may be especially consequential for sustaining momentum and supporting the development of transferable competencies. As a result, classroom instruction engaging holistic teaching practices represents a critical and proximal lever for creating more inclusive cultures of opportunity in community college STEM.

Beyond the classroom, students encounter a broader institutional context that shapes their sense of belonging, safety, connection, and access to support across the college (Coburn, 2004; Nora, 1990; Wang, 2024a). Institutional support captures how students experience this wider organizational environment, including the extent to which institutional structures, policies, and everyday interactions are perceived as accessible, equitable, and responsive to students’ academic, social, and emotional needs (Drake, 2011; Grillo & Leist, 2013). Importantly, these dimensions reflect students’ lived experiences of institutional culture and structure, that is, how systems operate in practice and are felt by students, rather than the mere presence of discrete programs or services. In this way, institutional support functions as a contextual condition that can either reinforce or undermine students’ engagement, persistence, and well-being, particularly for those navigating structural barriers in community college STEM pathways.

Despite substantial investments in STEM student success initiatives, limited research has examined how classroom teaching practices and broader institutional supports jointly function as part of an integrated ecosystem. Existing studies frequently examine instruction or institutional structures in isolation, which offers valuable but limited understanding of how teaching practices and broader institutional supports interact to shape students’ forward momentum and their self-efficacy in career-relevant competencies within community college STEM pathways. To address this gap, the present study investigates how these two levels of the STEM ecosystem (classroom-level instruction and institution-level supports) together shape students’ momentum and the development of self-efficacy in competencies critical for future educational and career pathways. Using survey data from students at three community colleges, we employ structural equation modeling (SEM) to assess these relationships while accounting for institutional variation using multi-group analysis. This study thus maps the STEM support ecosystem from students’ perspectives with the goal of contributing to efforts to design more coherent, equitable cultures and structures of opportunity in community college STEM education. Guided by this larger focus, we seek to answer the following main research question: How do holistic teaching practices and institutional support shape community college STEM students’ momentum and self-efficacy in career-relevant competencies?

More specifically, we address the following two sub-questions:

RQ1. How are holistic teaching practices and institutional support associated with community college STEM students’ curricular, cognitive, and meta-cognitive momentum, as well as their self-efficacy in career-related competencies?

RQ2. How are different forms of momentum associated with students’ self-efficacy in career-relevant competencies, and to what extent do they mediate the relationships between holistic teaching practices/institutional support and self-efficacy?

2. Relevant Literature

This review broadly examines research on community college STEM education with a focus on student momentum and career-relevant competencies. It also explores how classroom teaching practices and institutional supports influence students’ engagement, skill development, and persistence. In addition, the review describes how these proximal and distal factors interact to form a STEM student support ecosystem and identifies gaps that motivate the current study.

2.1. Student Momentum and Career-Relevant Competencies in Community College STEM Contexts

Research on community college STEM education increasingly emphasizes both forward momentum and the development of career-relevant competencies as key indicators of student success. Momentum has been defined as sustained progress and engagement across curricular, cognitive, and meta-cognitive domains (Adelman, 1999; Attewell et al., 2012; Wang, 2017). In community college contexts, where students often navigate complex course sequences, transfer requirements, and balancing academic and life responsibilities, sustained momentum is strongly associated with persistence, transfer, and degree completion in STEM (Núñez, 2022; Wang, 2015), which makes it a proximal predictor of long-term educational outcomes.

Closely related to momentum, career-relevant competencies, including professional and interpersonal skills and confidence in applying these skills in educational and workplace settings, is increasingly recognized as an essential outcome for community college STEM students preparing for further education and employment (Hora, 2019; Núñez, 2023). These competencies support students’ ability to communicate effectively, collaborate with others, and navigate professional and academic environments, all of which are linked to long-term career readiness and persistence in STEM fields (Herrera et al., 2022; Núñez, 2023). Evidence suggests that students who maintain high levels of cognitive and meta-cognitive engagement are more likely to develop these competencies and the accompanying self-efficacy (Wang, 2024a). Accordingly, momentum functions not only as a marker of academic progress but also as a mechanism through which students develop competencies critical to their educational and professional trajectories (Wang, 2017).

Despite the growing recognition of momentum and career-relevant competencies as critical outcomes, research on these topics and their relationships in community college STEM contexts remains limited. Existing literature focuses on retention, course completion, or transfer rates in isolation, rather than examining how students’ learning, engagement, and skill development are interrelated or shaped by proximal and distal supports. Studies addressing momentum and competencies often focus on four-year institutions or general postsecondary populations, leaving gaps in understanding how these outcomes manifest for community college STEM students, who are disproportionately first-generation, underrepresented, or balancing multiple responsibilities (Concha et al., 2024; Doran, 2023; Núñez, 2022).

This limited evidence constrains the development of interventions that simultaneously support students’ academic progress and the cultivation of career-relevant competencies. Examining self-efficacy in these competencies is particularly important, as confidence in applying such skills is closely linked to persistence, engagement, and long-term career readiness in STEM (e.g., Hora, 2019; Lent et al., 1994). Together, these gaps underscore the need to examine how holistic teaching practices and institutional support are associated with multiple dimensions of momentum and with self-efficacy in career-relevant competencies among community college STEM students.

2.2. Classroom Teaching Practices and Student Outcomes in STEM

Community college students primarily engage with their institutions through the classroom, which makes teaching practices a central driver of academic momentum, personal growth, and long-term success. Research has consistently shown that effective pedagogical practices, such as active learning, student-centered instruction, and holistic approaches that attend to students’ cognitive, social, and emotional needs, can enhance student engagement, persistence, and learning outcomes (Felder & Brent, 1996; Wang, 2020; Wang et al., 2017).

In STEM classrooms, where stringent expectations and evaluation mechanisms can amplify barriers for historically underserved students, holistic, human-centered teaching practices are particularly promising (Cohen & Kelly, 2018; Reyes, 2011; Roksa et al., 2009). Culturally responsive and humanizing pedagogies validate students’ identities and lived experiences, foster belonging, and strengthen motivation, supporting persistence and confidence in developing competencies essential for future academic and career pathways (Bartolomé, 1994; Gay, 2018; Ladson-Billings, 1995; Villegas & Lucas, 2002).

Yet in practice, students in community college STEM classrooms experience instruction not as isolated pedagogical strategies or discrete philosophical approaches, but rather as an integrated set of policies, practices, and relational dynamics that together signal whether the learning environment supports their success as whole people (Deil-Amen, 2011; Wang, 2020). Viewed this way, holistic teaching practices build upon insights from student-centered, culturally responsive, and humanizing pedagogical traditions (e.g., Bartolomé, 1994; Felder & Brent, 1996; Gay, 2018; Ladson-Billings, 1995), but reflect a broader conceptualization that emphasizes how multiple dimensions of classroom support converge in students’ lived experiences: the structure and flexibility of course policies, the extent to which instructors proactively connect students to resources and support services, and the quality of instructor-student relationships in conveying care, respect, and recognition of students’ complex lives beyond the classroom (Wang, 2017, 2020).

Taken together, this body of evidence suggests that holistic teaching practices may influence momentum by promoting sustained engagement across curricular, cognitive, and meta-cognitive domains. These practices may also shape students’ self-efficacy in career-relevant competencies through both direct and indirect pathways: directly by fostering the skills and confidence to navigate professional and academic contexts, and indirectly by supporting momentum. Overall, the literature suggests that holistic teaching practices constitute a critical classroom context in community college STEM that may influence students’ engagement and academic momentum while also contributing to the development of self-efficacy in career-relevant competencies through students’ lived instructional experiences.

2.3. Institutional Support and Student Outcomes in STEM

While the literature on classroom teaching practices in community colleges is robust, comparatively less is known about how students experience institutional support, including structures, policies, and relational practices designed to foster belonging, access, and equity across campus. Evidence suggests that advising, tutoring, financial aid, and other support services can promote retention and academic progress (Drake, 2011; Grillo & Leist, 2013; Nora, 1990; Wang, 2017, 2020), but less clear is how these supports function in relation to classroom experiences, particularly in STEM contexts. Recent research indicates that students’ engagement with institutional support is shaped not only by the availability of resources but also by the coherence and relational quality of those supports. When institutional structures are misaligned with classroom instruction or student needs, they may fail to reinforce learning or even constrain students’ progress and growth (Wang, 2024a).

The reviewed evidence suggests that classroom teaching and institutional supports can operate as interconnected components of a STEM student support ecosystem. Holistic teaching practices provide proximal conditions for engagement and self-regulation, generating momentum and fostering self-efficacy in career-relevant competencies, while institutional supports reinforce these processes by removing barriers, offering guidance, and providing opportunities for applied skill development. Together, these proximal and distal supports may help students navigate barriers, sustain engagement, and build the professional, interpersonal, and career-oriented skills and confidence necessary for persistence, transfer, and long-term success in STEM pathways.

Despite growing recognition of the importance of this ecosystem, most studies examine instruction or institutional support in isolation, which offers limited insight into how these components jointly influence momentum and self-efficacy in career-relevant competencies in community college STEM contexts (Perez, 2023). Accordingly, the current study models the combined influence of holistic teaching and institutional support on students’ momentum and self-efficacy in career-relevant competencies in community college STEM contexts.

3. Theoretical Framework

This study is guided by Wang’s (2017) Theoretical Model of Momentum for Community College Student Success, which examines how systemic, institutional, and individual factors shape students’ progress toward long-term educational goals. The model identifies three domains of momentum: curricular, teaching and learning (including cognitive and meta-cognitive components), and motivational. In this study, we focus primarily on curricular and teaching and learning momentum as key mechanisms influenced by holistic teaching and institutional support.

Teaching and learning momentum overall refer to the essential forms of progress that community college students need to develop and sustain to advance academically and professionally. Specifically, cognitive momentum reflects the cumulative progress toward learning and mastery of subject matter, while meta-cognitive momentum refers to self-regulatory capacity planning, monitoring progress, and learning from experience. Curriculum momentum captures forward motion through course-taking patterns (clarity about sequences, knowledge of requirements, intent to persist), whether following a structured program of study (such as a certificate or associate degree) or pursuing a specific educational objective (such as preparing for transfer, engaging in non-credit education, or taking courses for personal enrichment). These domains capture students’ engagement in STEM coursework, their academic progress, and their ability to plan, monitor, and regulate learning.

We also expand the motivational momentum domain by treating self-efficacy in career-relevant competencies as proxy motivational outcomes, which represent concrete indicators of STEM-related aspirations, agency, and readiness for future educational and career pathways. Specifically, we distinguish career readiness self-efficacy (confidence in preparing for and advancing within one’s STEM field) from professional and interpersonal self-efficacy (confidence in collaboration, communication, and problem-solving). These represent conceptually distinct domains of workplace competence, likely shaped by different developmental experiences and mechanisms (Hora, 2019; Lent et al., 1994; Núñez, 2023). The framework also incorporates contextual and demographic factors, including pre-college experiences, financial challenges, race, gender, first-generation status, employment, among others, to account for structural barriers and subgroup differences. By situating classroom and institutional factors within a broader ecosystem of student experience, this model provides a conceptual foundation for examining how holistic teaching and institutional support work together to shape STEM student momentum and development within community college settings.

4. Methods

4.1. Study Context and Research Design

This study is part of a large, transformative mixed methods project examining holistic support and student success across community colleges. The broader project targeted all degree- and credit-seeking students at three large institutions in two states: two in the Midwest and one in the Southeast. The three institutions vary in size and student demographics, serving between 25,000 and 70,000 students, with 30–40% low-income, 49–58% first-generation, and 35–75% students of color. All three colleges offer STEM programs, aligned with the United States Bureau of Labor Statistics definitions, alongside general education and vocational pathways. These programs lead to certificates, technical diplomas, or associate degrees and are designed to provide students with technical skills, foundational knowledge, and preparation for transfer to four-year institutions or entry into STEM careers. A key goal shared across all three research sites is to support the development of self-efficacy in career-relevant competencies that equip students for further education and professional success in STEM fields.

4.2. Participants and Data Collection

The target population for the larger project included all 69,209 degree- or credit-seeking students enrolled in Fall 2023. For this study, we focused on the subset of 1258 students enrolled in STEM programs. Data were collected via a survey developed based on our theoretical framework and relevant literature. The survey included questions assessing students’ experiences with teaching practices and institutional support, dimensions of student momentum, and self-efficacy beliefs in their competencies related to career development, including interpersonal, professional capabilities (Hora & Lee, 2021; Lent et al., 1994, 2000; Wang, 2017), along with demographic information. Face and content validity were established through review by experts and students similar to those at the research sites, and construct validity was assessed using confirmatory factor analyses (CFA), described further in Section 4.4.

The survey was distributed via Qualtrics in October 2023, with weekly reminders, and closed at the end of November. A total of 7874 students completed the survey, for a response rate of 11.4%. While a response rate of 11.4% is not considered high in conventional survey research, the literature on college student surveys indicates that response rates tend to be lower, with women and older students being more likely to respond (Porter & Umbach, 2006; Sax et al., 2008). College survey research also suggests a response rate of 5–10% is considered reliable from a minimum sample of 500 students (Fosnacht et al., 2017). Given our target population of 69,209 students and a final sample of 7874 with proportionately more non-traditional age students and women completing the survey, these results align with college survey research and represent a sufficient sample size with quality responses for robust analysis. After data cleaning, 6953 valid responses remained, of which 1258 STEM students were included in this study (see

Table 1. Characteristics of survey respondents.

		Total		Institution A		Institution B		Institution C
		N a	%	N a	%	N a	%	N a	%
Total		1258		247		141		870
Race/ethnicity	Students of Color	853	67.81	104	42.11	85	60.28	664	76.32
	White Students	405	32.19	143	57.89	56	39.72	206	23.68
Gender	Man	641	50.95	137	55.47	76	53.90	428	49.20
	Woman	545	43.32	97	39.27	55	39.01	393	45.17
	Non-Binary	72	5.72	13	5.26	10	7.09	49	5.63
Transgender		76	6.04	11	4.45	11	7.80	54	6.21
International Students		131	10.41	20	8.10	18	12.77	93	10.69
DACA Recipients		50	3.97	5	2.02	9	6.38	36	4.14
Financially Dependent		685	54.45	114	46.15	56	39.72	515	59.20
Low-Income Students b		462	36.72	95	38.46	61	43.26	306	35.17
First-Generation Students		776	61.69	146	59.11	97	68.79	533	61.26
Adult Students		431	34.26	101	40.89	72	51.06	258	29.66
Full-Time Students		809	64.31	160	64.78	82	58.16	567	65.17
Employed		776	61.69	179	72.47	93	65.96	504	57.93
Served in U.S. Armed Forces, Reserves, or National Guard		56	4.45	6	2.43	10	7.09	40	4.60
Been in Foster Care		25	1.99	5	2.02	1	0.71	19	2.18
Single Parents		70	5.56	15	6.07	21	14.89	34	3.91
Have Diagnosed Disability		247	19.63	72	29.15	23	16.31	152	17.47
English as First Language		928	73.77	198	80.16	115	81.56	615	70.69
Been Incarcerated in Jail or Prison or a Juvenile Facility		34	2.70	8	3.24	8	5.67	18	2.07
Access to Transportation		1169	92.93	236	95.55	131	92.91	802	92.18

^a Due to the missing data, columns may not sum to the total sample size. ^b Low-income is defined as students’ total household income in 2022 below $30,000.

for demographic details).

4.3. Measures

4.3.1. Key Independent Variables

Given our study focus, our key exogenous (independent) variables include two latent factors: holistic teaching practices and institutional support. Holistic teaching practices capture the extent to which instructors create supportive, student-centered learning environments, including the helpfulness of grading and attendance policies, guidance toward academic resources, and the degree to which students feel supported as whole individuals in their classes. Consistent with the literature reviewed above, these indicators reflect an integrated instructional orientation that draws on student-centered, humanizing, and culturally responsive pedagogical traditions, while emphasizing how these elements function together in students’ lived classroom experiences. Rather than measuring discrete instructional strategies or singular pedagogical philosophies, this construct captures the combined influence of course policies, resource connections, and relational support that collectively signal an instructor’s commitment to students’ academic success and overall well-being. Institutional support reflects students’ experiences of the broader campus environment, including feelings of welcome, safety, connection, access to supportive resources, and perceptions of faculty fairness and understanding of students’ unique life circumstances. This factor captures students’ holistic college experiences that extend both within and outside the classroom. Together, these latent factors represent proximal (classroom) and distal (institutional) conditions that may influence students’ momentum and self-efficacy in career-relevant competencies.

4.3.2. Main Dependent Variables

The study includes two main endogenous (dependent) latent variables capturing self-efficacy in career-relevant competencies. The first, career readiness self-efficacy, reflects students’ confidence in their ability to prepare for, enter, and advance in their chosen field, including goal setting, skill acquisition, navigating hiring processes, and pursuing further education. The second, professional and interpersonal self-efficacy, captures students’ confidence in applying professional and interpersonal skills in academic and workplace contexts, such as collaborating with others, generating creative solutions, effectively communicating ideas in writing and orally, and responding constructively to feedback. Together, these variables represent the key proximal outcomes that holistic teaching practices and institutional supports are hypothesized to influence.

4.3.3. Mediating Variables

Students’ curricular, cognitive, and meta-cognitive momentum were included as mediators in the model, hypothesized to be directly influenced by holistic teaching practices and institutional support, and to directly predict self-efficacy in career-relevant competencies. Curricular momentum was measured using items assessing students’ clarity about course sequences, knowledge of required courses each semester, and intention to continue enrolling until achieving their academic goals. Cognitive momentum was assessed via items capturing the degree to which coursework supported understanding of content, development of knowledge and expertise, problem-solving, critical thinking, and preparation for their intended field or occupation. Meta-cognitive momentum was measured using items reflecting self-regulated learning behaviors, including planning and preparation for class, monitoring progress, prioritizing tasks, and learning from prior experiences.

A set of individual background characteristics, such as race/ethnicity, gender, employment, first-generation status, full-time enrollment, disability status, and single-parent status were included as covariates.

Table 2. List of variables.

Variable Name	Description
Exogenous Variables
Holistic Teaching Practices	How helpful to your learning are your instructors’ grading practices or policies? a How helpful to your learning are your instructors’ attendance expectations or policies? a How helpful are your instructors in pointing you to resources or services that support your success in classes? a How supported do you feel as a whole person in your classes? a
Institutional Support	Reflecting on your experience at the college, how much do you feel …   welcomed on this campus? b   supported at your college? b   connected to your college? b   safe at your college? b   that you have a person or place to turn to if you have questions or concerns? b   that faculty are understanding of students’ unique life circumstances? b   that faculty are fair and unbiased in their treatment of individual students? b
Mediating Variable
Curriculum Momentum	How clear are you about the next course(s) you should take? a How informed do you feel about what course(s) you need to take each semester? a How likely are you to continue enrolling until you reach your main goal? a
Cognitive Momentum	How much do your learning experiences through coursework help you …   have a clear understanding of course content? a   develop knowledge and expertise in the topics covered? a   develop your problem-solving skills? a   refine your critical thinking? a   prepare you for your field or occupation of interest? a
Meta-Cognitive Momentum	How often do you use the following approaches?   Preparing for class by reading assigned text and other learning materials ahead of time c   Planning your activities in order to do your best work (e.g., identifying best places to study; managing time to ensure timely completion of academic tasks) c   Monitoring your academic progress and making corrections/adjustments when needed c   Determining and prioritizing important tasks c   Learning from prior experiences, knowledge, and errors c
Endogenous Variables: Self-Efficacy in Career-Relevant Competencies
Career Readiness Self-Efficacy	In light of your future career and education, how confident are you about each of the following?   Be well prepared for a job in your intended field a   Go through a hiring process and secure a job in your field a   Meet the needs of an employer in your intended field a   Have gained the necessary skills to be successful in your field a   Develop a clear career plan to promote your professional advancement a   Know if/what further education may be required to meet long-term goals a   Successfully pursue further education (including transfer to a bachelor’s degree program) a
Professional and Interpersonal Self-Efficacy	In light of your future career and education, how confident are you about each of the following?   Work collaboratively with others to complete a given task a   Come up with creative solutions to problems or challenges a   Know who and where to go if you need help to solve complex problems a   Effectively give a presentation individually and collaboratively a   Effectively share your ideas out loud a   Write clearly and effectively to communicate information a   Effectively handle criticism and constructive feedback a
Covariates
Race/Ethnicity	What race/ethnicity best describes you? 1 = Students of Color   ● African American or Black   ● Middle Eastern or North African or Arab or Arab American   ● Southeast Asian   ● American Indian or Alaskan Native   ● Indigenous   ● Hispanic or Latinx/Latina/Latino or Chicanx/Chicana/Chicano   ● Pacific Islander or Native Hawaiian   ● Other Asian or Asian American   ● Other (please specify)   ● Prefer not to answer 2 = White or Caucasian
DACA	Are you a recipient of the DACA (Deferred Action for Childhood Arrivals) program? d
Gender	What is your gender? 1 = Man; 2 = Woman; 3 = Non-Binary
Transgender	Are you transgender? d
Financially Dependent	Are you considered financially dependent on your parent(s) or legal guardian(s)?
Low-Income	Considering income from all sources, are you considered as low-income? d e
First-Generation	Are you a first-generation student? d
Adult Students	Are you an adult student? d
Full-Time Status	Are you full-time student? d
Employment Status	Are you employed? d
Veteran	Are you currently or have you ever served in the U.S. Armed Forces, Reserves, or National Guard? d
Foster	Have you ever been in foster care? d
Single Parents	Are you a single parent with at least one dependent child? d
Disability	Do you have a documented or diagnosed disability? d
English as the First Language	Is English your first or primary language? d
Been Incarcerated in Jail or Prison or a Juvenile Facility	Have you ever been incarcerated in jail or prison or a juvenile facility? [Incarceration means serving a term in prison or jail.] d
Access to Transportation	Do you have access to transportation (via own car/carpooling, public transportation, riding a bike, etc.) to and from your college? d

^a Response options for the items are 1 = Not at all, 2 = A little, 3 = Somewhat 4 = Very, 5 = Extremely. ^b Response options for the items are 1 = None, 2 = A little, 3 = Some, 4 = A lot, 5 = A great deal. ^c Response options for the items are 1 = Never, 2 = Rarely, 3 = Sometimes, 4 = Often, 5 = Very often. ^d Response options for the items are 1 = Yes, 0 = No. ^e Low-income is defined as students’ total household income in 2022 below $30,000.

provides a complete detailed description of all study variables.

4.4. Data Analysis

We conducted multi-group structural equation modeling (SEM), using institution as the grouping variable, to examine the relationships among holistic teaching practices, institutional support, student momentum, and self-efficacy in career-relevant competencies among STEM students. SEM simultaneously estimates measurement and structural components, linking observed survey items to latent constructs and evaluating relationships among exogenous, mediating, and endogenous variables (Anderson & Gerbing, 1988; Hatcher & O’Rourke, 2013; Jöreskog & Sörbom, 1996; Kaplan, 2009). Multi-group SEM allows testing whether these relationships operate equivalently across institutions, with institution specified as the grouping variable (Evermann, 2010). Analyses were conducted in R (Version 4.5.1) using the lavaan package (R Core Team, 2020; Rosseel, 2012).

4.4.1. Preparation and Descriptive Analysis

Descriptive statistics were computed for the full sample and disaggregated by institution. Non-normality in observed variables was addressed using maximum likelihood estimation with robust standard errors (MLR), and missing data were handled with full information maximum likelihood (FIML; Enders & Bandalos, 2001; Li, 2016).

4.4.2. Confirmatory Factor Analysis

We specified a full measurement model (Figure 1) including all seven latent constructs: two exogenous variables (holistic teaching practices and institutional support), three mediating variables (curricular, cognitive, and meta-cognitive momentum), and two endogenous variables measuring self-efficacy in career-relevant competencies. CFA using MLR estimation in R (lavaan package) confirmed that the factor structure fit the data well for the full STEM sample and within each institution. This validated measurement model provided a robust foundation for the subsequent multi-group SEM analyses.

4.4.3. Multi-Group SEM

With the validated measurement model, we specified the structural model (Figure 2) to examine how holistic teaching practices and institutional support influenced student momentum and core competencies, both directly and indirectly through momentum. Institution was treated as the grouping variable to test whether the model operated equivalently across the three colleges.

Measurement invariance was assessed sequentially. Configural invariance tested whether the same factor structure held across institutions with freely estimated parameters. Metric invariance then constrained factor loadings to examine whether constructs were measured on the same scale, and scalar invariance further constrained item intercepts to allow comparison of latent means. Changes in robust fit indices (ΔCFI ≤ 0.01, ΔRMSEA ≤ 0.015) were used alongside chi-square difference tests to evaluate whether deviations from invariance were functionally trivial (Chen, 2007; Cheung & Rensvold, 2000).

After establishing measurement invariance, the structural model was fitted for the full STEM sample and separately for each institution to assess model fit. Multi-group SEM was then used to evaluate structural invariance, first by freely estimating structural paths across institutions (configural model) and then constraining all paths to be equal. If the constrained model fit the data adequately, structural invariance was assumed; if not, individual paths were examined to identify sources of non-invariance.

The structural model included four sets of regressions: the effects of holistic teaching and institutional support on momentum, the effects of momentum on career-related self-efficacy, direct effects of holistic teaching and institutional support on career-related self-efficacy, and their indirect effects on self-efficacy via momentum. Indirect effects were evaluated to test whether momentum mediated the relationships between exogenous variables (holistic teaching practices and institutional support) and endogenous variables (career-related self-efficacy). All SEM analyses were conducted in R using the lavaan package (Rosseel et al., 2025), with MLR to account for non-normality (Li, 2016), and FIML to handle missing data (Enders & Bandalos, 2001). Model fit was evaluated using multiple indices, including the comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR). Values indicating good fit were CFI and TLI ≥ 0.95, RMSEA ≤ 0.06, and SRMR ≤ 0.08 (Hu & Bentler, 1999). Fit indices were examined for the full sample and separately within each institution to ensure acceptable model fit prior to multi-group comparisons.

5. Findings

Descriptive statistics for STEM students from the three institutions are presented in Table 1. The results are organized in four sections: CFA measurement model, multi-group SEM measurement invariance, multi-group SEM structural invariance, and final multi-group SEM with institution as the grouping variable.

5.1. Confirmatory Factor Analysis Findings

The full seven-factor CFA model fit the data well (χ²(635, N = 1258) = 1993.16, p < 0.001; CFI = 0.956; TLI = 0.951; RMSEA = 0.038; SRMR = 0.042). Although the chi-square was significant, which is a common issue in large samples such as ours (Kline, 2023), other indices suggested a strong fit. Standardized factor loadings were all above 0.5 and significant (

Table 3. Unstandardized and standardized factor loadings of the measurement model.

Factors and Indicators	Unstd.		SE	Std.All
Holistic Teaching Practices
How helpful to your learning are your instructors’ grading practices or policies? a	1.000		0.627	0.670
How helpful to your learning are your instructors’ attendance expectations or policies? a	0.893	***	0.047	0.555
How helpful are your instructors in pointing you to resources or services that support your success in classes? a	1.073	***	0.058	0.713
How supported do you feel as a whole person in your classes? a	1.339	***	0.066	0.824
Institutional Support
Reflecting on your experience at the college, how much do you feel …
welcomed on this campus? b	1.000		0.755	0.774
supported at your college? b	1.168	***	0.033	0.862
connected to your college? b	1.256	***	0.043	0.788
safe at your college? b	0.739	***	0.033	0.616
that you have a person or place to turn to if you have questions or concerns? b	1.154	***	0.049	0.737
that faculty are understanding of students’ unique life circumstances? b	1.083	***	0.043	0.759
that faculty are fair and unbiased in their treatment of individual students? b	0.912	***	0.041	0.688
Curriculum Momentum		***
How clear are you about the next course(s) you should take? a	1.000	***	0.608	0.592
How informed do you feel about what course(s) you need to take each semester? a	1.075	***	0.054	0.632
How likely are you to continue enrolling until you reach your main goal? a	0.714	***	0.056	0.574
Cognitive Momentum
How much do your learning experiences through coursework help you …
have a clear understanding of course content? a	1.000		0.571	0.775
develop knowledge and expertise in the topics covered? a	1.150	***	0.038	0.822
develop your problem-solving skills? a	1.093	***	0.049	0.727
refine your critical thinking? a	1.071	***	0.052	0.687
prepare you for your field or occupation of interest? a	1.171	***	0.062	0.681
Meta-Cognitive Momentum
How often do you use the following approaches?
Preparing for class by reading assigned text and other learning materials ahead of time c	1.000		0.572	0.561
Planning your activities in order to do your best work (e.g., identifying best places to study; managing time to ensure timely completion of academic tasks) c	1.207	***	0.059	0.734
Monitoring your academic progress and making corrections/adjustments when needed c	0.984	***	0.058	0.680
Determining and prioritizing important tasks c	0.949	***	0.064	0.706
Learning from prior experiences, knowledge, and errors c	0.821	***	0.056	0.644
Career Readiness Self-Efficacy
In light of your future career and education, how confident are you about each of the following?
Be well prepared for a job in your intended field a	1.000		0.848	0.833
Go through a hiring process and secure a job in your field a	1.173	***	0.033	0.855
Meet the needs of an employer in your intended field a	1.112	***	0.031	0.882
Have gained the necessary skills to be successful in your field a	1.080	***	0.030	0.846
Develop a clear career plan to promote your professional advancement a	1.018	***	0.035	0.781
Know if/what further education may be required to meet long-term goals a	0.767	***	0.034	0.635
Successfully pursue further education (including transfer to a bachelor’s degree program) a	0.659	***	0.033	0.544
Professional and Interpersonal Self-Efficacy
In light of your future career and education, how confident are you about each of the following?
Work collaboratively with others to complete a given task a	1.000		0.627	0.672
Come up with creative solutions to problems or challenges a	0.924	***	0.040	0.704
Know who and where to go if you need help to solve complex problems a	1.076	***	0.053	0.721
Effectively give a presentation individually and collaboratively a	1.189	***	0.050	0.688
Effectively share your ideas out loud a	1.152	***	0.051	0.694
Write clearly and effectively to communicate information a	0.897	***	0.047	0.661
Effectively handle criticism and constructive feedback a	0.856	***	0.049	0.629

Note. *** p < 0.001. ^a Response options for the items are 1 = Not at all, 2 = A little, 3 = Somewhat 4 = Very, 5 = Extremely. ^b Response options for the items are 1 = None, 2 = A little, 3 = Some, 4 = A lot, 5 = A great deal. ^c Response options for the items are 1 = Never, 2 = Rarely, 3 = Sometimes, 4 = Often, 5 = Very often.

). CFA by institution also confirmed acceptable fit across our research sites (

Table 4. Fit indices of measurement models.

Model	Chi-Square	df	Δχ2 (Δdf)	p	CFI	ΔCFI	TLI	RMSEA	ΔRMSEA	SRMR
All three institutions (N = 1258)	1993.164	635			0.956		0.951	0.038		0.042
Institution A (N = 247)	1190.868	635			0.909		0.899	0.052		0.064
Institution B (N = 141)	931.552	635			0.926		0.918	0.050		0.068
Institution C (N = 870)	1689.195	635			0.954		0.949	0.040		0.044
Multi-Group Analysis: Institution
Configural model	3811.615	1905			0.943		0.937	0.044		0.051
Metric invariance model	3888.985	1967	67.7 (62)	0.290	0.942	0.001	0.938	0.043	0.001	0.053
Scalar invariance model	4041.961	2029	154.6 (62)	0.000	0.939	0.003	0.937	0.044	0.001	0.053

).

5.2. Results of Measurement Invariance Tests

Building on the CFA results, we tested measurement invariance across institutions (Table 4). The configural model, which imposed no equality constraints, demonstrated good fit (χ²(1905, N = 1258) = 3811.615, p < 0.001, CFI = 0.943, TLI = 0.937, RMSEA = 0.044, SRMR = 0.051). Constraints were then added by fixing factor loadings to be equal across institutions, which did not significantly differ from the configural model (p = 0.290) and showed minimal changes in fit indices (ΔCFI and ΔRMSEA), supporting the equivalence of loadings. Further constraining item intercepts for scalar invariance resulted in a significant chi-square difference (p < 0.001), but changes in CFI (ΔCFI = 0.003) and RMSEA (ΔRMSEA = 0.001) were trivial. These results indicate that factor loadings and item intercepts are equivalently measured across institutions, allowing valid comparisons of structural relationships and latent means.

5.3. Results of Structural Invariance Tests

After establishing measurement invariance, we tested structural invariance to examine whether the hypothesized relationships among constructs were consistent across the three institutions (

Table 5. SEM model fit statistics and results of structural invariance tests.

Model	Chi-Square	df	Δχ2 (Δdf)	p	CFI	ΔCFI	TLI	RMSEA	ΔRMSEA	SRMR
All three institutions (N = 1258)	3194.381	1286			0.936		0.932	0.033		0.043
Institution A (N = 247)	2080.420	1286			0.856		0.845	0.048		0.064
Institution B (N = 141)	1868.064	1286			0.830		0.817	0.056		0.075
Institution C (N = 870)	2699.042	1286			0.936		0.931	0.033		0.047
Multi-Group Analysis: Institution
Configural model	6933.025	4014			0.904		0.901	0.040		0.056
All Path coefficient constrained-equal model	6992.116	4086	63.6.0 (72)	0.750	0.904	0.000	0.903	0.039	0.001	0.057

). We first specified a configural structural model in which the overall model was held constant across institutions while allowing structural paths to vary freely. This model demonstrated acceptable fit (χ²(4014, N = 1258) = 6933.025, p < 0.001, CFI = 0.904, TLI = 0.901, RMSEA = 0.040, SRMR = 0.056). Next, all structural paths were constrained to be equal across institutions, and the constrained model was compared to the configural model. The corrected chi-square difference test was not significant (p = 0.750), and changes in model fit indices were minimal (ΔCFI = 0.000; ΔRMSEA = 0.001), which supported structural invariance. This indicated that the strength and direction of relationships among constructs are equivalent across three institutions. Accordingly, we proceeded with the full multi-group SEM analysis accounting for the three institutions, with the understanding that variation across institutional contexts did not substantively alter the relationships specified in the model.

5.4. Results of Final Multi-Group SEM

The final model demonstrated acceptable model fit, with χ²(4086, N = 1258) = 6992.116, p < 0.001, CFI = 0.904, TLI = 0.903, RMSEA = 0.039, SRMR = 0.057. The final structural model’s estimated unstandardized and standardized structural path coefficients are presented in

Table 6. Coefficient estimates of the final multi-group structural model.

		Estimate		Std.Err	Std.All	95% CI	p (>|z|)
Direct Effect
Outcome	Predictor
Curriculum Momentum	Holistic Teaching Practices	0.763	***	0.199	0.768	[0.373, 1.152]	0.000
	Institutional Support	−0.029		0.144	−0.035	[−0.313, 0.254]	0.838
Cognitive Momentum	Holistic Teaching Practices	0.816	***	0.161	0.915	[0.501, 1.131]	0.000
	Institutional Support	−0.083		0.114	−0.111	[−0.306, 0.139]	0.464
Meta-Cognitive Momentum	Holistic Teaching Practices	0.705	***	0.149	0.685	[0.414, 0.997]	0.000
	Institutional Support	−0.161		0.102	−0.185	[−0.361, 0.038]	0.113
Career Readiness Self-Efficacy	Curriculum Momentum	0.160		0.105	0.117	[−0.046, 0.367]	0.128
	Cognitive Momentum	0.456	***	0.102	0.299	[0.256, 0.656]	0.000
	Meta-cognitive Momentum	0.078		0.066	0.059	[−0.051, 0.208]	0.235
	Holistic Teaching Practices	0.192		0.175	0.141	[−0.150, 0.535]	0.271
	Institutional Support	0.108		0.068	0.094	[−0.026, 0.242]	0.113
Professional and Interpersonal Self-Efficacy	Curriculum Momentum	0.083		0.076	0.079	[−0.066, 0.232]	0.273
	Cognitive Momentum	0.155	*	0.077	0.132	[0.003, 0.307]	0.045
	Meta-cognitive Momentum	0.318	***	0.053	0.312	[0.214, 0.421]	0.000
	Holistic Teaching Practices	0.214		0.129	0.204	[−0.039, 0.468]	0.098
	Institutional Support	0.147	**	0.050	0.166	[0.050, 0.244]	0.003
Predictor → Outcome	Mediator
Holistic Teaching Practices → Career Readiness Self-Efficacy	Curriculum Momentum	0.122		0.081	0.087	[−0.037, 0.282]	0.133
	Cognitive Momentum	0.372	***	0.096	0.264	[0.183, 0.561]	0.000
	Meta-cognitive Momentum	0.055		0.045	0.039	[−0.033, 0.144]	0.221
Holistic Teaching Practices → Professional and Interpersonal Self-Efficacy	Curriculum Momentum	0.064		0.056	0.063	[−0.046, 0.173]	0.257
	Cognitive Momentum	0.127	*	0.062	0.126	[0.005, 0.248]	0.041
	Meta-cognitive Momentum	0.224	***	0.051	0.223	[0.125, 0.323]	0.000
Institutional Support → Career Readiness Self-Efficacy	Curriculum Momentum	−0.005		0.023	−0.004	[−0.050, 0.040]	0.838
	Cognitive Momentum	−0.038		0.050	−0.033	[−0.137, 0.061]	0.451
	Meta-cognitive Momentum	−0.013		0.012	−0.011	[−0.036, 0.010]	0.283
Institutional Support → Professional and Interpersonal Self-Efficacy	Curriculum Momentum	−0.002		0.012	−0.003	[−0.026, 0.021]	0.836
	Cognitive Momentum	−0.013		0.017	−0.016	[−0.047, 0.021]	0.455
	Meta-cognitive Momentum	−0.051		0.031	−0.063	[−0.113, 0.010]	0.104
		Estimate		Std.Err	Std.All	95% CI	p(>|z|)
Outcome	Covariates
Career Readiness Self-Efficacy	Race/Ethnicity	0.072		0.049	0.044	[−0.024, 0.168]	0.141
	International Students	−0.006		0.073	−0.002	[−0.149, 0.137]	0.932
	DACA Recipient	−0.040		0.108	−0.007	[−0.252, 0.171]	0.707
	Gender	−0.010		0.038	−0.008	[−0.084, 0.064]	0.782
	Transgender	0.032		0.098	0.008	[−0.161, 0.224]	0.748
	Financially Dependent	−0.112	*	0.050	−0.069	[−0.210, −0.014]	0.025
	Low-Income	0.008		0.043	0.005	[−0.077, 0.093]	0.856
	First-Generation	0.008		0.043	0.005	[−0.077, 0.092]	0.857
	Adult Student	−0.047		0.053	−0.028	[−0.151, 0.057]	0.376
	Full-Time Status	−0.024		0.044	−0.014	[−0.111, 0.063]	0.591
	Employed	0.146	**	0.045	0.080	[0.057, 0.234]	0.001
	Veteran	−0.097		0.101	−0.019	[−0.295, 0.100]	0.334
	Foster	0.080		0.166	0.014	[−0.245, 0.406]	0.628
	Single Parents	−0.061		0.087	−0.018	[−0.231, 0.109]	0.484
	Disability	−0.027		0.058	−0.015	[−0.142, 0.087]	0.639
	English as First Language	0.052		0.050	0.026	[−0.046, 0.150]	0.300
	Been Incarcerated in Jail or Prison or a Juvenile Facility	0.031		0.121	0.007	[−0.205, 0.268]	0.794
	Have Access to Transportation	−0.048		0.078	−0.012	[−0.201, 0.105]	0.537
Professional and Interpersonal Self-Efficacy	Race/Ethnicity	0.028		0.034	0.022	[−0.040, 0.095]	0.421
	International Students	0.016		0.051	0.007	[−0.084, 0.116]	0.755
	DACA Recipient	−0.152		0.087	−0.034	[−0.324, 0.019]	0.082
	Gender	0.027		0.027	0.026	[−0.026, 0.081]	0.315
	Transgender	−0.039		0.070	−0.013	[−0.177, 0.098]	0.575
	Financially Dependent	−0.110	**	0.035	−0.088	[−0.178, −0.042]	0.001
	Low-Income	0.014		0.031	0.011	[−0.047, 0.074]	0.654
	First-Generation	−0.095	**	0.031	−0.074	[−0.155, −0.034]	0.002
	Adult Student	0.060		0.038	0.047	[−0.014, 0.133]	0.113
	Full-Time Status	0.000		0.031	0.000	[−0.061, 0.062]	0.994
	Employed	0.097	**	0.032	0.070	[0.035, 0.160]	0.002
	Veteran	0.033		0.081	0.008	[−0.125, 0.192]	0.680
	Foster	0.020		0.124	0.004	[−0.222, 0.262]	0.872
	Single Parents	−0.190	**	0.058	−0.073	[−0.304, −0.075]	0.001
	Disability	0.097	*	0.040	0.070	[0.019, 0.174]	0.015
	English as First Language	0.054		0.034	0.034	[−0.014, 0.122]	0.118
	Been Incarcerated in Jail or Prison or a Juvenile Facility	0.064		0.095	0.018	[−0.122, 0.250]	0.503
	Have Access to Transportation	−0.047		0.058	−0.015	[−0.160, 0.066]	0.417

Note. *** p < 0.001, ** p < 0.01, * p < 0.05.

. Below, we describe several key findings from the SEM analysis. Figure 3 presents the final model.

First, holistic teaching practices emerged as a robust and consistent predictor of curricular momentum as well as cognitive and meta-cognitive momentum in the teaching and learning domain. This suggests that when students experience integrated, supportive teaching practices, they are more likely to sustain academic progress, engage deeply with STEM learning, and regulate their own learning processes. Holistic teaching practices showed a moderate positive association with students’ professional and interpersonal self-efficacy (p = 0.098, β = 0.204), though this relationship did not reach conventional levels of statistical significance. The magnitude of the standardized coefficient suggests a potential practical contributory role of supportive instructional environments in fostering these competencies. However, this finding should be interpreted with caution and warrants further investigation.

Second, somewhat unexpectedly, institutional support did not significantly predict any form of momentum within the curricular as well as teaching and learning domains. This null finding should be interpreted with caution: after accounting for classroom-level teaching practices and other modeled relationships, institutional supports did not appear to exert an independent association with immediate, short-term momentum. This result takes on added nuance in light of the finding that institutional support was significantly and positively associated with students’ professional and interpersonal self-efficacy. This pattern suggests that, while institutional supports may not directly influence immediate momentum, they may contribute meaningfully to broader skill development and stronger self-efficacy for longer-term student outcomes.

Third, the momentum students built in the teaching and learning domain played a central role in shaping broader career-related self-efficacy. Both cognitive and meta-cognitive momentum were significant positive predictors of professional and interpersonal self-efficacy, while career readiness self-efficacy was significantly and positively predicted only by cognitive momentum, a finding that highlights the importance of sustained intellectual engagement and self-regulated learning for STEM skill development. In contrast, curricular momentum was not significantly related to either of the outcome variables, which suggests that its influence may be less direct or contingent on other forms of momentum.

Fourth, cognitive momentum significantly mediated the positive association of holistic teaching practices with both measures of self-efficacy in career-relevant competencies. Meta-cognitive momentum also transmitted the impact of holistic teaching practices to career readiness self-efficacy. Together, these findings indicate that holistic teaching practices contribute to students’ self-efficacy in career competencies both directly and indirectly through momentum, which reinforces the central role of classroom-level supports in fostering skill development. In contrast, institutional support showed a direct association only with professional and interpersonal self-efficacy, which suggests that institutional supports may influence confidence in broader skill development without necessarily operating through immediate momentum in the teaching and learning domain.

Among the covariates, several patterns are noteworthy for contextual understanding. Employment status was positively associated with both endogenous variables, while financial dependency was linked to lower levels of these outcomes. First-generation status and being a single parent were negatively associated with professional and interpersonal self-efficacy. Interestingly, students with disabilities reported higher levels of professional and interpersonal self-efficacy relative to their peers, net of other factors. These patterns provide important context for interpreting the study findings and highlight how personal, social, and economic circumstances may differentially shape confidence in career preparation and professional skills within community college STEM pathways.

6. Discussion, Implications, Limitations, and Conclusions

6.1. Discussion

Our study provides a nuanced examination of how holistic teaching practices and institutional support function within the community college STEM ecosystem and how they are associated with student momentum and self-efficacy in career-relevant competencies. The findings highlight the centrality of classroom instruction as a proximal mechanism through which engagement, skill development, and confidence in STEM competencies are cultivated. Holistic teaching practices consistently predicted cognitive and meta-cognitive momentum, indicating that when instructors provide integrated, supportive, and human-centered learning environments, students are better able to sustain engagement, develop self-regulated learning strategies, and internalize confidence in their professional and interpersonal abilities. In STEM classrooms, where rigorous disciplinary norms, high-stakes evaluation, and gatekeeping practices are prevalent (Gasiewski et al., 2012; Núñez, 2023; Seymour & Hewitt, 1997), these instructional approaches are particularly salient. By fostering repeated mastery experiences and validating diverse student identities, holistic pedagogy reinforces Bandura’s (1997) conceptualization of self-efficacy development while extending prior research demonstrating that student-centered and culturally responsive teaching is especially effective for historically underserved populations in STEM (Cohen & Kelly, 2018; Freeman et al., 2014; Ladson-Billings, 1995).

The differential effects of institutional support observed in this study illuminate the complexity of the STEM education ecosystem. While student-perceived institutional support did not independently predict immediate curricular or teaching and learning momentum, it was significantly associated with professional and interpersonal self-efficacy. Based on our findings, these results hold consistently across three institutional contexts. This pattern suggests that distal, structural supports, such as advising, tutoring, mentoring, and institutional policies, may operate more as enabling conditions than as direct drivers of short-term engagement. These supports likely function by creating scaffolds and opportunities that support students’ confidence in applying skills to academic and professional contexts, particularly in navigating complex STEM pathways.

Meanwhile, students’ lived experiences of support may not align neatly with two analytically distinct constructs (teaching practices and institutional support), even when those constructs are statistically separable in the measurement model. Institutional resources are often encountered both within and beyond the classroom, frequently mediated by faculty members. In community college contexts, students’ primary points of contact with the institution frequently occur through course-based experiences and interactions (Deil-Amen, 2011), which can blur perceived boundaries between teaching-related and institutional sources of support that influence momentum. Consequently, the effects of institutional support on momentum may be absorbed by, or contingent upon, classroom experiences. Nevertheless, institutional supports that foster belonging, provide clear guidance, and signal institutional commitment remain crucial for building professional confidence among students who are balancing work, family, and financial responsibilities.

This distinction is consistent with ecological perspectives (Bronfenbrenner, 1979), which posit that distal contextual factors influence outcomes indirectly by shaping proximal processes. In the STEM context, where sequential, high-stakes coursework and gatekeeping mechanisms dominate (Cohen & Kelly, 2018; Seymour & Hewitt, 1997), the efficacy of institutional support may manifest primarily through its capacity to foster students’ agency, confidence, and ability to translate learning into career-relevant competencies. Moreover, although many institutions are working to develop more holistic and student-centered support systems, awareness and accessibility of these services may remain uneven among community college students (Wang, 2024a). This challenge may be particularly pronounced for students who primarily enroll online, commute, or have limited social or familial exposure to college environments. As a result, institutional supports may not consistently translate into gains in students’ momentum or short-term academic outcomes, and may also constrain their potential influence on career readiness self-efficacy.

The mediating role of cognitive and meta-cognitive momentum further underscores the intertwined nature of instructional, institutional, and individual processes. Sustained intellectual engagement and self-regulated learning not only facilitate mastery of STEM content but also provide the foundation for developing confidence in professional and interpersonal competencies, such as communication, collaboration, and problem-solving. These dynamics reinforce the argument that proximal instructional experiences and the momentum they generate are critical levers for developing self-efficacy in competencies that extend beyond immediate coursework. Interestingly, curricular momentum, which is conceptually linked to course progression and clarity about academic pathways, did not emerge as a significant predictor. This lack of a significant relationship may reflect temporal dynamics not captured in cross-sectional data; that is, curricular momentum may require sustained progression over multiple terms before translating into confidence in career-relevant competencies. Alternatively, this null finding may also carry substantive implications in suggesting that linear progression alone is insufficient for cultivating the deeper skills and confidence necessary for STEM career readiness. More specifically, students might develop the forward motion through course experiences (Wang, 2017), yet the broader STEM program and curricular structures may not be fully aligned with current STEM workforce demands, such as emerging fields in data science, renewable energy, or biotechnology. This highlights the importance of instructional quality, engagement, and reflection in translating curricular progression into meaningful career-oriented outcomes, aligning with prior literature emphasizing the importance of active and contextualized learning experiences in STEM (Freeman et al., 2014; Seymour & Hewitt, 1997; Wang et al., 2017).

It is also important to bring to light, individual and social background characteristics that provide important context for interpreting students’ self-efficacy in career-relevant competencies. Employment status was positively associated with professional, interpersonal, and career readiness self-efficacy, suggesting that work experience may reinforce skill development and confidence (Bandura, 1997; Lent et al., 2000). In contrast, financial dependency, first-generation status, and single-parent responsibilities were linked to lower professional and interpersonal self-efficacy, which may highlight how structural and familial demands can constrain engagement with learning and mentorship opportunities (Núñez, 2022; Wang, 2020). Interestingly, students with disabilities reported higher professional and interpersonal self-efficacy, potentially reflecting resilience and adaptive strategies developed through navigating systemic barriers (da Silva Cardoso et al., 2013; Jenson et al., 2011). These patterns indicate that while holistic teaching practices and institutional support are key levers for fostering momentum and self-efficacy, personal and socio-economic circumstances shape how students experience and benefit from these supports. This underscores the need for flexible, equity-oriented interventions that ensure STEM pathways are accessible and enabling for all students.

6.2. Implications

Taken together, these findings suggest that the STEM ecosystem in community colleges operates as an interconnected network of proximal and distal supports, with complex, non-linear relationships among instruction, institutional structures, student engagement, and the cultivation of self-efficacy in skills development. Holistic teaching provides the immediate conditions for engagement and momentum, which then translate into confidence and competency development, while institutional supports may amplify or sustain these processes, particularly for students navigating structural barriers. This underscores that improving STEM outcomes requires attention not only to discrete interventions but also to the coherence and alignment of the broader ecosystem, as reflected in repeated calls in the literature for integrative, equity-centered approaches to postsecondary STEM education (Perez, 2023).

The implications of this research are multifaceted. At the classroom level, faculty development should prioritize culturally responsive, student-centered, and reflective pedagogies that explicitly attend to skill development and confidence building, rather than focusing solely on content coverage. Faculty could develop practices such as using inclusive language, designing transparent assignments, providing structured opportunities for real-world collaboration, and offering timely formative feedback. Community colleges should also recognize and leverage the assets of faculty with industry experience and allocate greater support and resources to recruit and retain instructors who could connect coursework to professional contexts. These practices have the potential to foster mastery experiences, increase engagement, and enhance students’ self-efficacy in career-relevant competencies.

At the institutional level, support services should be intentionally aligned with classroom experiences, offering coherent, visible, and actionable resources that complement instructional efforts. Examples include proactive advising models, clear transfer and degree maps, and career-integrated programming beyond classroom experience, all of which can help students envision attainable futures in STEM. The observed positive relationship between institutional support and professional and interpersonal self-efficacy suggests that institutions can leverage their broader structures to enhance students’ confidence in applying learned skills, even when these supports do not appear to immediately influence course engagement or cognitive momentum. This calls for a reimagining of institutional supports to be more integrated, responsive, and explicitly tied to students’ professional and STEM-related competencies. Their effectiveness also depends on the visibility and accessibility of supports, particularly when embedding advising, career exploration, and resource awareness into gateway STEM courses. Many students continue to rely on individual faculty members or limited networks to navigate college, which pinpoints the importance of proactively integrating essential information and resources into students’ experiences, rather than expecting them to seek out services on their own.

6.3. Limitations

This study contributes to a growing understanding of how STEM support systems function within community college contexts, where historically underserved and first-generation students disproportionately navigate complex academic, financial, and structural barriers. At the same time, we need to transparently acknowledge several limitations and caveats that situate our findings and highlight important directions for future inquiry. To begin, because the analyses draw on cross-sectional, correlational data, causal inferences cannot be made regarding the relationships among holistic teaching practices, institutional support, momentum, and self-efficacy in career-relevant competencies. Although the models were theoretically grounded and specified indirect pathways, the observed associations may reflect reciprocal or unmeasured processes that unfold over time. Future research using longitudinal, quasi-experimental, or experimental designs is needed to more fully capture how instructional and institutional contexts shape momentum, self-efficacy, and longer-term educational and career outcomes in STEM pathways.

Moreover, future research could benefit from including a broader range of geographical locations and larger sample sizes. While our study draws on data from three institutions across two states with a relatively robust sample, community colleges vary widely in regional context, institutional size, and student demographics. Incorporating more geographically and institutionally diverse samples could enhance the generalizability of findings across community college contexts.

In addition, the reliance on student self-reported measures may introduce measurement error and reflect perceptions rather than enacted practices or demonstrated competencies. While such perceptions are theoretically meaningful, particularly for constructs related to support, belonging, and self-efficacy, future studies would benefit from incorporating multiple data sources, including classroom observations, institutional records, and performance-based assessments of career-relevant competencies. Doing so would allow researchers to more precisely examine how instructional and institutional practices relate to measurable skill development.

The study’s focus on STEM students across three community colleges may also limit generalizability to other institutional or disciplinary contexts, given variation in organizational structures, resources, and labor market alignment. Nonetheless, this focus is a strength insofar as community colleges play a critical role in broadening participation in STEM and supporting students who face intersecting structural constraints. Future research could examine how these relationships vary across STEM disciplines, institutional types, or regional labor markets.

Finally, although multiple forms of momentum were modeled as mediating mechanisms, other developmental processes, including but not limited to STEM identity formation (J. Rodriguez et al., 2016), mentoring relationships (Crisp & Cruz, 2009; Packard, 2012), or workplace learning (Hora, 2019; S. L. Rodriguez et al., 2016), were not examined and may further shape students’ self-efficacy in career-relevant competencies. Building on these findings, future studies should attend not only to the direct effects of teaching and institutional practices but also to the mechanisms and interactions through which proximal and distal supports influence self-efficacy and longer-term career trajectories. Longitudinal and mixed methods approaches, in particular, hold promise for illuminating how momentum and competency development unfold over time and intersect with disciplinary norms, workforce expectations, and equity considerations within community college STEM pathways.

6.4. Conclusions

In conclusion, our research demonstrates that fostering career-relevant competencies and confidence among community college STEM students requires attending to both the immediate instructional environment and the broader institutional ecosystem. Holistic teaching practices serve as a strong predictor in the model of momentum and competency development and are associated with mastery experiences and engagement that translate into sustained self-efficacy. Institutional supports can complement these processes by scaffolding skill application and confidence, even when short-term academic engagement is not directly affected. Together, these empirical insights highlight the need for an integrated, equity-oriented approach to STEM education that aligns classroom practices, institutional supports, and student engagement to cultivate not only academic progress but also the confidence and competencies required for long-term career and educational success in STEM pathways. By framing these dynamics within the broader STEM ecosystem, our study advances understanding of how community colleges can strategically design coherent, responsive, and inclusive pathways that support diverse STEM learners in realizing their professional aspirations.