A Cultural Mismatch Intervention to Increase Science Self-Efficacy Among STEM College Students

Abstract

Cultural Mismatch Theory (CMT) suggests mismatch between interdependent home norms and independent school norms can hinder academic success for historically marginalized (HM) students who are more likely to be first-generation college students and underrepresented in science, technology, engineering, and math (STEM). The effectiveness of a CMT intervention to increase science self-efficacy was tested among 213 (M_age = 22.99, SD = 5.74; 8.2% HM) STEM majors from a community college and baccalaureate-granting institution. CMT intervention students reported higher science self-efficacy relative to the control group. The findings support scalable CMT interventions to address STEM workforce disparities.

Increased diversification of the science, technology, engineering, and math (STEM) workforce has important implications for addressing the complex health needs of a rapidly diversifying U.S. population. Students from historically marginalized (HM) communities are often more familiar with, motivated to, and able to implement culturally-competent care in ways that can address ethnic and racial health disparities (Rabinowitz et al., 2000; U.S. Department of Health and Human Services, 2006; Valantine & Collins, 2015). However, while individuals from HM backgrounds currently make up approximately 32.6% of the U.S. population and are projected to grow to 44.2% in 2060 (U.S. Census, 2018), approximately 8.4% of doctorates in STEM fields were conferred to HM students in 2019 (NSF, 2019a). A report released by the National Center for Education Statistics documented that students of color, as well as firstgeneration college and lower-income students, who become interested in STEM-relevant majors are more likely to leave the field or drop out than white students or students in other majors (Chen & Soldner, 2013). Interventions that address sociocultural barriers by reconciling home and academic cultural values or highlighting how contextual differences in education play a role in adaption during college may support persistence in education (Stephens et al., 2012; Townsend et al., 2018).

Guiding Framework: Cultural Mismatch Theory (CMT)

Cultural mismatch theory (CMT) suggests that disparities in education can be explained by differences in cultural dispositions that are aligned or misaligned with the values of postsecondary institutions (Stephens, Fryberg et al., 2012; Stephens, Markus et al., 2012). Firstgeneration college students (students whose parents did not receive a college degree; Stephens, Fryberg et al., 2012) are more likely to be raised in working class homes that foster interdependent cultural values and academic motivations (e.g., “learn to work with others”). These values can mismatch with the independent values of postsecondary institutions (e.g., “learn to work independently”). However, these independent cultural values match with the socialization of continuing-generation college students (students whose parents received a college education and who often come from middle- and upper-class socioeconomic backgrounds). Emergent research on CMT suggests cultural value mismatch plays a critical role in college adjustment (e.g., lower grades, poorer performance on academic related tasks; Stephens, Fryberg et al., 2012; Vasquez-Salgado et al. 2015, Vasquez-Salgado et al., 2018; Vasquez-Salgado et al., 2021; Authors, in press). This cultural barrier has been documented as a major determinant of poor adjustment and educational disparities among social class groups at baccalaureate-granting institutions (BGI; Stephens, Fryberg et al., 2012; Vasquez-Salgado et al., 2021; Authors et al, in press).

CMT interventions may be particularly effective for HM students who are more likely to be first-generation college students and hold interdependent values (Stephens, Hamedani et al., 2014; Townsend et al., 2019; Vasquez-Salgado et al., 2021). For example, Stephens, Fryberg and colleagues (2012) showed that first-generation college students who were randomly assigned to read a welcome letter that incorporated interdependent messages (e.g., the university “has a tradition of learning through community”) performed significantly better on an academic task than those that read the traditional letter with independent messages (e.g., the university “has a tradition of independence”). First-generation college students who read the interdependent message also performed similarly to continuing-generation college students, demonstrating the power of interdependent messages to eliminate educational disparities. This galvanizing CMT intervention, along with other successful social psychological interventions aiming to foster belonging and academic adjustment, are implicit in nature (Stephens, Hamedani et al., 2014; Townsend et al., 2019). The current study aims to extend prior interventions by testing the effectiveness of an intervention that explicitly informs HM students about cultural values, CMT, as well as strategies for resolving cultural mismatch.

A final and important element of CMT work is identifying cultural mismatches in the lived experiences of HM students, our population of interest. In particular, Vasquez-Salgado and colleagues (2015; 2018) demonstrated that Latinx first-generation college students experience home-school cultural value mismatch – a mismatch between interdependent family obligations and independent academic obligations – during the transition to college. Evidence, across studies using qualitative, survey and experimental designs, suggests that this mismatch negatively disrupts academic performance (Vasquez-Salgado et al., 2015; 2018; 2021). In addition, Latinx first-generation college students may experience peer-peer cultural value mismatch – a mismatch between interdependent ideologies and practices of one peer and the independent ideologies and practices of another – during the transition to college that can also affect academic performance (e.g., problems with attention and learning, lower GPA; Burgos-Cienfuegos et al., 2015; Authors, in press). It is noteworthy to mention first-generation college status was the key variable that predicted experiences with both forms of mismatch (homeschool, peer-peer) in survey work with multi-ethnic samples above and beyond Latinx or HM background (Vasquez-Salgado et al., 2021; Authors, in press), suggesting that the results reflecting socioeconomic status differences, rather than ethnic differences. Taken together, these findings highlight that in addition to broader, cultural mismatch noted in CMT (Stephens, Fyberg et al., 2012), there are specific types of mismatches that first-generation college students experience in postsecondary institutional settings. However, there is no CMT intervention to that has incorporated these mismatches into an intervention among college students. In the current study, we extend prior work by focusing on these specific and salient cultural value mismatches between home, peer, and academic contexts.

Cultural Mismatch among STEM Fields of Study and Across Institution Types

Given the unsupportive and competitive nature of many science courses (Hurtado et al., 2009; Malcom & Feder, 2016), at odds with interdependent values (e.g., “learn to do collaborative research”), it is possible that cultural value mismatch is exacerbated in STEM fields. Indeed, research suggests that individualism is one main aspect that deters HM students and women from STEM undergraduate majors and careers (Diekman et al., 2010; 2011) as it mismatches with interdependent goals (Smith et al., 2014). In addition, Canning and colleagues (2020) found first-generation college students are negatively impacted by high competition within courses, which can lead to imposter syndrome. Together, this literature suggests that interdependent notions often mismatch with the emphasis on independence or individual competition within many STEM courses, creating difficulties for first-generation college students in STEM education. Programs that incorporate the importance of “community” within a science and larger educational context among students from HM backgrounds may cultivate a stronger sense of science identity and self-efficacy among student participants relative to those who do not participate (Saetermoe et al., 2017; Camacho et al., 2021; Vasquez-Salgado et al., 2023). However, those interventions can be 2 years in length. The intervention presented in the current study provides a scalable and cost-effective complement or alternative.

Additionally, the literature has focused on baccalaureate-granting universities (McGee et al., 2012), despite the fact that HM students are more likely than non-HM students to receive part of their training at a community college (NSF, 2019b; Solórzano et al., 2013). Approximately 40% of first-time college students begin their post-secondary education at CCs (NCES, 2016; Shapiro et al., 2016). Though CCs serve as a mechanism of upward mobility and advancement for HM students, unfortunately, few transfer to 4-year colleges and universities and they remain overrepresented at CCs and underrepresented at higher levels of educational attainment (McGee et al., 2012; Perez & Ceja, 2010; Solórzano et. 2005), generating a significant loss in talent in the STEM fields. Given the conflicting goals of CCs to provide vocational education and terminal associate’s degrees as well as transfer opportunities, limited resources mean that CCs have the lowest retention rates among STEM majors (Snyder & Cudney, 2017). Therefore, an early intervention at the CC level, even with a minimally-viable, 10-minute intervention, may increase science self-efficacy, and thus, potentially reduce the educational trajectory disruption experienced by HM, first-generation college and low-income students interested in STEM. Our research tests such a minimally-viable intervention at a 2-year CC and 4-year BGI. By minimally-viable, a term we adopt from the technology-business sector to describe a product developed to maximize benefit relative to risk and effort, we are proposing an intervention that can be easily and effectively deployed without unduly increasing participant burden or institutional resources. In this way, the minimally-viable intervention can be a stand-alone activity or flexibly embedded in a research curriculum or program to augment other intervention strategies. Therefore, in the current study, we test the hypotheses that STEM students from across institution types who experience a minimally-viable CMT intervention will report higher science self-efficacy relative to a control group. Additionally, we hypothesize that intervention effects will be stronger for first-generation college students.

Method

Participants

College students (N = 213, M_age = 22.99, SD = 5.74, 72% female) age 18 or older in social science, biomedical, and STEM courses were recruited via flyers, course announcements, and email. Participants were from African American (0.9%), Asian American (9.9%), Latinx (76.5%), White (4.2%), multiracial (6.6%), and other or unknown ethnic backgrounds (1.4%). The majority were from HM backgrounds (88.2%). The average parent education level was between high school diploma or equivalent and vocational or technical program after high school. First-generation college students (i.e., neither parents had some college experience) represented 68.3% of the sample and continuing-generation college students 31.7% (i.e., at least one parent had some college experience or above). Students were enrolled at a 2-year CC (81.2%) or 4-year BGI (18.8%) in Southern California. The majority reported intending to major in the social sciences (61.5%) or other STEM majors (e.g., Chemistry, Nursing, Nutrition, Computer Science; 34.7%).

Design

In this pre-test post-test control group design, participants first responded to questions in a 30 minute, Qualtrics survey about their background, experience in college courses, and preintervention science self-efficacy. After completing this portion of the survey, students were randomly assigned (“A/B” experiment) to a 10 minute lesson on Cultural Mismatch Theory (CMT intervention group; n = 104) or the history of higher education in California (control group; n = 109). Participants in the CMT intervention were provided information about cultural values (e.g., independent and interdependent beliefs and attitudes), CMT theory and research (e.g., home-school cultural value mismatch, peer-peer cultural value mismatch), and sample strategies for how to resolve cultural mismatch (e.g., thinking of their college as “their home”, connecting with classmates or other peers in a study group). Participants in the control group were provided information about higher education in California (e.g., Donahoe Higher Education Act, California Master Plan for Higher Education), functionality of the three-tier public higher education system in California, and sample strategies for how to navigate educational resources (e.g., programs and offices that offer student support, how to connect with the Academic Counseling or Learning Resource Center). The curriculum was developed by the authors to increase content knowledge (e.g., what is CMT?) and skills (e.g., how can cultural value mismatch be resolved?) in ways that would affect student self-perceptions (i.e., science self-efficacy) relative to standard orientations that focus on connecting students to resources. Learning check quizzes with two questions were provided after each presentation to ensure comparability in attention. After the two learning check questions, participants responded to post-intervention science self-efficacy items.

Science Self-Efficacy Measure.

Science self-efficacy (SSE) measures were modified based on Research Self-Efficacy Measures (Byars-Winston et al., 2016) developed and validated with historically underrepresented ethnic and racial groups in science. Participants reported their level of confidence in their ability to excel in their science major over the next two semesters (SSE1), persist with science courses even though they may be a minority in them (SSE2), complete a science degree (SSE3), pursue a graduate degree in science (SSE4), and complete a graduate degree in science (SSE5), and pursue a research science career (SSE6) on a scale from 1 (no confidence) to 5 (complete confidence). Items were averaged to form one score, with a higher score indicating higher science self-efficacy. Cronbach’s alphas = .94 and .95 for pre- and post-intervention, respectively, which indicates high scale reliability.

Power Analyses.

Power analysis in G*Power suggested a minimum sample size of 128 (64 in each of the conditions) to be sufficient to detect a medium effect size (i.e., η² = .063; Townsend et al., 2019) with 80% power, assuming an independent samples t-test with a significance level of α = .05.

Results

Manipulation Checks.

To test that random assignment to the two conditions was effective, we conducted preliminary analyses and confirmed that there were no significant differences between the CMT intervention and control conditions in pre-intervention science self-efficacy (t(195) = .19, p = .849). Additionally, as a manipulation check, participants in both the CMT (M = 2.00, SD = .00) and control conditions (M = 1.77, SD = .57) responded correctly to at least 50% of the two learning check questions on average.

Analytic Strategy.

Table 1 shows the correlations between age, parent education, and post-intervention science self-efficacy items. Given the correlational results and prior research that suggest that sociodemographic factors shape science self-efficacy (e.g., Byars-Winston et al., 2016; Townsend et al., 2019), age, parent education, gender, HM status, college generation status, and institution type were controlled for in final hypothesis-testing analyses. Regressions controlling for age, parent education, gender, HM status, college generation status, institution type, and pre-intervention science self-efficacy (Step 1) were modeled to test our prediction of the effectiveness of the CMT intervention on post-intervention composite science self-efficacy. To test the hypothesis that the CMT intervention would be more effective for first-generation college students, a college generation status x intervention interaction term was entered in Step 2.

Table 1.

Correlations of Covariates and Post-Intervention Science Self-Efficacy

	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Age	−.16*	.09	.15*	.10	.03	.09	.12†	−.01	.12†	.12	.12	.14†	.09	.09	.10
2. Parent Education	-	−.12†	.03	.16*	−.13†	−.13†	−.14†	−.12†	−.14†	−.04	−.07	−.18*	−.16*	−.14†	−.15*
3. Pre-Intervention SSE Composite		-	.71**	.86**	.90**	.92**	.91**	.92**	.85**	.65**	.78**	.82**	.78**	.79**	.79**
4. Pre-Intervention SSE1			-	.57**	.52**	.52**	.54**	.58**	.58**	.65**	.57**	.51**	.45**	.45**	.48**
5. Pre-Intervention SSE2				-	.74**	.73**	.71**	.75**	.72**	.55*	.75**	.67**	.64**	.63**	.66**
6. Pre-Intervention SSE3					-	.79**	.78**	.84**	.78**	.57**	.71**	.80**	.71**	.72**	.73**
7. Pre-Intervention SSE4						-	.92**	.82**	.80**	.53**	.68**	.79**	.79**	.79**	.75**
8. Pre-Intervention SSE5							-	.80**	.80**	.56**	.67**	.79**	.77**	.80**	.74**
9. Pre-Intervention SSE6								-	.77**	.58**	.68**	.73**	.72**	.73**	.74**
10. Post-Intervention – SSE Composite									-	.78**	.88**	.94**	.93**	.93**	.94**
11. Post-Intervention SSE1										-	.68**	.69**	.60**	.57**	.66**
12. Post-Intervention SSE2											-	.80**	.77**	.73**	.80**
13. Post-Intervention SSE3												-	.87**	.89**	.86**
14. Post-Intervention SSE4													-	.93**	.87**
15. Post-Intervention SSE5														-	.86**
16. Post-Intervention SSE6															-

Note. SSE = Science Self-Efficacy.

^†p < .10.

^*p < .05.

^**p< .01.

Hypothesis Testing.

As shown in Table 2, relative to the control group, students in the CMT condition reported significantly higher post-intervention science self-efficacy (descriptive means: M_CMT = 3.16, SD = 1.34; M_control = 2.90, SD = 1.35). The interaction between intervention and college generation status was not significant, suggesting that the intervention was effective for first- and continuing-generation college students (See Supplemental File for item by item analyses where first-generation college students report higher post-intervention science selfefficacy in completing a science degree relative to continuing-generation students). Additionally, there were no significant interactions by gender, HM status, or institution type.

Table 2.

Regression Model Predicting Post-Intervention Science Self-Efficacy

	SSE Composite
Step 1
Constant	.38 (.17)*
Age	.01 (.01)
Parent Education	−.08 (.10)
Gender	.00 (.06)
Underrepresented Minority Status	−.02 (.08)
College Generation Status	−.10 (.10)
Institution Type	−.06 (.08)
Time 1 Science Self-Efficacy Item	.89 (.04)**
CMT Intervention (vs. Control Condition)	.25 (.11)*
Step 2
College Generation Status x Intervention	.06 (.12)

Note. Age was mean-centered. Parent education was the sum of the z-score of mother’s education and father’s education centered at 0 (z-score). Gender was effect-coded with −1 = male and 1 = female. Underrepresented minority status was effect-coded by NIH definition (African Americans, Hispanics or Latinos, American Indians or Alaska Natives, Native Hawaiian) with −1 = non-URM and 1 = URM. College generation status was effect-coded where −1 = continuing-generation college student and 1 = first-generation college student. Institution type was effect-coded with −1 as 4-year institution and 1 as 2-year institution. CMT = Cultural Mismatch Intervention.

^†p < .10.

^*p < .05.

^**p< .01.

Discussion

The current study tested the effectiveness of a minimally-viable, 10-minute intervention aimed at introducing and providing strategies to resolve cultural value mismatch between home and academic cultures in increasing science self-efficacy. Results suggest that the Cultural Mismatch Theory (CMT) intervention can increase students’ science self-efficacy (e.g., confidence in exceling in a science major and pursuing a graduate degree in science) relative to a control group. These results remained even after controlling for other factors associated with science and academic career achievement (age, parent education, gender, HM status, college generation status, institution type, pre-intervention science self-efficacy), which suggests that the CMT intervention has the potential to be effective for all students. These results among STEM college students extend the growing body of literature by demonstrating the effectiveness of eliminating educational disparities by resolving cultural mismatch, particularly among first-generation students (e.g.,Stephens, Hamedani et al., 2014; Townsend et al., 2019). Our results also extend “lay theory” or “difference-education” interventions that are believed to empower students through normalizing challenge and helping them understand that difference can be context-specific and overcome (e.g., social differences can be explained by the ways people differentially adapt to various contexts and situations; Stephens, Hamedani, et al., 2019; Townsend et al., 2019; Yeager et al., 2016). By explicitly introducing the concept of cultural values, CMT, as well as strategies for resolution, our intervention is unique because it normalizes the phenomena of mismatch and provides students with tools to aid them in resolving the mismatch.

Although promising, there are several limitations of the current study. The participants represented here were mostly community college, female, Latinx, and first-generation college students. We believe this intersectional approach strengthens the study as these groups are a representative yet often understudied student population in the STEM literature (e.g., approximately 65% of Latinx students who graduate from high school move to a community college students; Solórzano et. 2005). While students in these demographic groups are most likely to be underrepresented in STEM and most likely to benefit from CMT interventions (e.g., Vasquez-Salgado et al., 2018; 2021), further research is needed to identify student populations who may be able to benefit most from related interventions or to establish the universality of these type of interventions. Additionally, though we focused on science self-efficacy, all measures were cross-sectional and self-reported. Future studies should employ longitudinal methods to examine the effect of CMT interventions on career outcomes as well as replicate these findings with other career development outcomes, such as STEM sense of belonging and identity (e.g., Byars-Winston et al., 2016). Additionally, given how the science self-efficacy composite and item-level analyses may vary across individual demographic groups (e.g., lower construct validity for Latinx females relative to African American males and females; Byars-Winston et al., 2016), the current study and future studies can contribute to advancing work on measurement tools. It is also interesting that participants in the CMT intervention group scored significantly higher in the learning check questions than participants in the control group. This may be indicative of the relevance of the CMT content. If this is the case, future studies should consider developing equally salient control conditions.

Altogether, the results suggest that interventions based on CMT, even a brief one, can be a powerful tool in addressing educational disparities for college students from various backgrounds and institutions, particularly first-generation college students in science majors. Online interventions such as the one presented in the current study that promote students’ understanding of cultural or contextual differences have been shown to be as effective in improving academic outcomes (e.g., GPA) among first-generation college students as in-person interventions (Stephens, Hamedani, et al., 2019; Townsend et al., 2019). However, to provide support for the development and refinement of scalable online CMT interventions, future studies should compare how this online CMT intervention performs when delivered in different formats given research that suggests college students may respond differently to content delivered online versus in-person and passively versus actively (e.g., reflective writing task) in terms of satisfaction, motivation, and engagement (e.g., Harackiewicz & Priniski, 2018; Means et al., 2020). Additionally, although psychological constructs like science interest are related to academic performance (e.g., Harackiewicz & Hulleman, 2010), examining distal academic outcomes (e.g., GPA, retention, maintenance of STEM major and career choice) will be critical next steps as academic metrics are important in engaging policy makers in eliminating achievement gaps (Harackiewicz & Priniski, 2018). As another means to scale intervention efforts, future interventions should identify the ways in which educating college professors, counselors, and other student support staff about CMT may facilitate student science engagement and persistence given research that suggests that parent and teacher expectations are particularly influential in student decisions to pursue a STEM major and career (Woong Lee, Min, & Mamerow, 2015). Together, CMT interventions can fuel and “fire the creative juices” of students at the individual level as well as create institutional level change that can improve HM persistence in STEM (Estrada et al., 2016) and, ultimately, shape the future of the national STEM workforce.