There is recognition on the part of governmental leadership that a healthy, globally competitive U.S. economy is dependent on continuing to attract qualified secondary students into academic tracks that lead to STEM or STEM-allied careers—often called the STEM pipeline (NAS, 2007; NRC, 2008). Of particular concern is that many populations have been historically underrepresented in STEM fields, representing a great loss of talent (Lowell & Salzman, 2007). These underrepresented populations include women and people with African-American, Hispanic/Latino, and Native American backgrounds, among others. In recent years, the proportion of students from underrepresented groups who are pursuing STEM degrees/careers has been flat or falling (Horwedel, 2006; Taningco, et al., 2008). Unfortunately, many of the reported outcomes of successful programs specifically aimed at underrepresented groups (e.g., Nave, Frizell, Obiomon, Cui, & Perkins, 2006) do not use robust enough analytic evaluation techniques to enable others to understand the source(s) of innovation and, therefore, to replicate and scale the program.

 

Strategies taken to address STEM pipeline issues usually center on approaches that address perceived shortcomings in the K-12 educational system. One such area is the reported shortage of highly qualified STEM teachers (AACTE, 2007; Forum, 2006; NCTE, 2009). The lack of qualified teachers is, in part, the result of the vicious cycle of not enough students with STEM concentrations in college leading to not enough K-12 teachers with appropriate STEM content exposure (Roundtable, 2008). One approach to this issue has been to focus on improving pre-service (e.g., Hora & Millar, 2009) or in-service (e.g., Klein, 2009)teacher professional development, while others focus on developing improved curricular materials for teachers to use (e.g., Winn, Lewis, & Curtis, 2009). Research has shown that sustained professional development for teachers is related to student achievement, however most teachers do not have access to these kinds of activities (Darling-Hammond, et al., 2009; Duschl, et al., 2008).

Student access to and enrollment in challenging secondary school courses are central to preparing them for college-level math and science courses (ACT, 2006). High quality STEM courses in K-12 schools lead to better college and workforce preparation, even for students who don’t pursue a STEM career (EdSource, 2008). Researchers note by the time students reach high school, science and math interest becomes a strong predictor of STEM degree choice in college (Maltese & Tai, 2010; Lynch, 2009). This points to the importance of aggressively promoting interest in science and math among students before the time they reach high school. One strategy for influencing these attitudes has been with the inclusion of mentors in STEM classrooms. This can be in the form of post-secondary STEM majors coming into classrooms through programs such as NSF’s GK-12 (e.g., Laursen, Thiry, & Liston, 2006; Thompson & Lyons, 2009) or through adult professional mentoring programs (e.g., Carline, Patterson, Davis, & Irby, 1998; Lowery, Kurpius, & Kerr, 2005). While researchers have noted the value of student exposure to role models and mentors who work in STEM fields (Brody, 2006; Gasbarra, Johnson, & Agenda, 2008), there are challenges with scaling intensive one-to-one programs (Markowitz, 2004).

 

Student experiences outside of formal secondary schooling, including venues such as informal science centers or conferences, are also strategies for impacting the STEM pipeline (Watkins, et al., 2009). Informal science centers can play a role as a gathering place for affinity groups and interactions with mentors (Fadigan & Hammrich, 2004). Research has pointed to the importance of early exposure to STEM career options (ACT, 2006; Gasbarra, et al., 2008), and alternate venues outside of school can provide for such opportunities. Collectively, reported research points to a diversity of approaches and venues to impact persistence in the STEM pipeline. Lent and colleagues have used social cognitive career theory models to demonstrate that numerous environmental and structural contextual factors influence student major/career choices at every stage of the STEM pipeline (Lent, et al., 2003; Lent, et al., 2005).

 

A strategic area of focus that has seen increased interest is looking at the points of transitions, or junctures, between the established educational systems along the STEM pipeline continuum. Perhaps none is more important than the critical juncture between secondary and post-secondary education, where students are moving from the K-12 educational system into either a two-year or four-year college. Successful transition across this juncture means proper preparation—both cognitive and affective—during secondary school, proper course selection, advising, and support in college (especially early on), and in some cases, special transition programs for students. College admissions policies have been shown to have impact on the enrollment of underrepresented groups (Harris & Tienda, 2010). Similarly, researchers have pointed to a need for college admissions to be informed by a deeper understanding of the characteristics of students who persist in STEM degree (Scott & Tolson, 2009).

Studies looking into Hispanic and female persistence in STEM college degrees point to the importance of robust campus infrastructure to support these students (Crisp, Nora, & Taggart, 2009; Lowery, et al., 2005; Merisotis & Kee, 2006). One small, but important, factor is the importance of college advisors being knowledgeable of high school curriculum for correct placement into college math courses (Norman, 2008). Many universities have set up special pre-college transition programs (e.g., Horwedel, 2006), specifically targeting underrepresented populations (e.g., Nave, et al., 2006). However, important meta-analyses into the secondary to post-secondary transition have pointed out that the lack of sufficiently rigorous research and evaluation in this area limits the ability to understand what strategies work and why (Goldrick-Rab, 2007; McLendon, Heller, & Lee, 2009). Even where there are rigorous studies, often they only include data sources from one side of the transition (e.g., just data from the freshman year in college (Hurtado, et al., 2007)). Alternately, they may look at both secondary and post- secondary data, but are exploratory or only include a narrow range of majors (e.g.,Mendez, et al., 2008).

Support for the critical secondary to post-secondary transition needs to come in the form of individual initiatives, but also in the establishment of larger, ongoing partnerships. Testimony before Congress on the America COMPETES Act, noted that deep commitment to coordination among key stakeholders (including K-12 and higher education) was crucial to improving the quality of K-12 STEM education (Snider, 2010). These partnerships should go beyond a single university campus to reach out to the K-12 school systems and other entities actively involved in supporting the STEM pipeline. For example, collaboration might take the form of formal coordination between high schools, community colleges, and universities(Hoffman, Vargas, & Santos, 2009) or with cooperating government labs, NGOs, and informal science centers (Simkin & Futch, 2006). Findings from a state-wide STEM pipeline initiative in Ohio pointed to the importance of state-wide coordination and communication in addressing the program challenges (AYPF, 2009). A similar program in Michigan brought together K-12, higher education, and business and showed positive impact on student career goals (Amato-Henderson, et al., 2009).

Gauging the success of various initiatives supporting the STEM pipeline through K-12 education and across the critical juncture to post-secondary education requires purposefully developing and supporting systems that facilitate the collection and analysis of data to inform the improvement of educational initiatives (Tolley & Shulruf, 2009). In some cases, evaluating the impact of innovations means the collection of new data, in other cases the merging data from existing databases has shown promise (Topping & Sanders, 2000). There has also been a call to collect and analyze data longitudinally in order to more fully understand the impact of educational innovations such as outreach programs (Subotnik, Rayhack, & Edmiston, 2006). A 21-year-old biomedicine outreach program has demonstrated the value of using longitudinal data to track students through the program and inform the continued improvement of the program (Winkleby, 2007; Winkleby, et al., 2009). On a larger scale, the Tennessee Value-Added Assessment System (TVAAS) used longitudinal data to understand the impact of teacher effectiveness on student learning (Sanders & Horn, 1998) and how student characteristics (and experiences) also interact with teaching to affect outcomes (e.g., Ballou, Sanders, & Wright, 2004). Recent editorials have pointed to the TVAAS data system as one of the key reasons as to why Tennessee was able to compete successfully for the U.S. Department of Education’s Race to the Top funds (Editors, 2010).

 

Educational researchers have demonstrated the value of advanced statistical techniques using longitudinal data in gauging the impact of, and informing the revision of educational outreach programs. Domina (2009) showed that longitudinal data analysis could be used to assess the impact of on students and drive improvements in these type of outreach programs. Not surprisingly, researchers have found that the relationship between individual characteristics and interest and persistence towards a STEM degree/career is complex (Valian, 2007). Looking at gender as a factor, programs that specifically target girls around STEM topics have shown promise, but often lack hard data concerning relationship between outreach intervention, course selection, and college degree choice (e.g., Koenig & Hanson, 2008). Available rigorous research demonstrates the importance of studying these factors (cf., Su, Rounds & Armstrong, 2009) and has pointed to fundamental effect of gender on activity preferences, with gender being a primary source of difference in self-efficacy beliefs related to STEM career choices (Zeldin, Britner & Pajares, 2008). These differences carry over to academic achievement where research shows women are less likely to pursue STEM degrees or enroll in highly competitive colleges when compared to men with same GPA (Viadero, 2009). Girls may enroll in high-achievement STEM-track courses in middle school but still have lower self-efficacy towards STEM degree/career tracks (Catsambis, 1995). Longitudinal research with a long-running mathematics outreach program for young women points to a complex relationship between high school math course selection and envisioned majors/careers (Berenson, Michael, & Vouk, 2006). Similarly, Jayaratne and colleagues (2003) point to important differences based on race and ethnicity as critical in how outreach programs impact female student attitudes (i.e., science self-concept and interest, persistence and aspirations in science).

 

Part of the difficulty of assessing the impact of K-12 STEM outreach programs is the challenge of meta-evaluations of secondary education STEM enrichment programs when evaluation has not been coordinated across programs (Carline, et al., 1998). This lack of coordination is hampered by the wide range of variables found to predict enrollment in STEM majors (Nicholls, Wolfe, Besterfield-Sacre, Shuman & Larpkiattaworn, 2007). In a meta-study of NSF MSP projects (Hora, Millar, Arrigoni & Kretchmar, 2009), the researchers pointed to the obstacles in collecting data on STEM learning outcomes and large variation in evaluation approaches across NSF projects. In the public discourse over the STEM pipeline issue, the lack of hard data has often meant that advocacy has won out over research. “Advocacy has often taken the place of research in education. Given the attractive and intuitively persuasive arguments set forth by education advocates in the past, coupled with the challenges faced in collecting the rigorous data necessary to carry out a thorough analysis, it is no small wonder that the difference between advocacy and research has been obscured.”(Subotnik, et al., 2010) Advocacy for a single project, however, rarely provides a strong foundation for an institution to collectively and creatively move forward, in an informed way, towards higher impact in their STEM outreach work.

In summary, the relationship between student abilities, attitudes, and choices, the contexts in which these are formed, and persistence along a STEM career trajectory are complex and map across many factors that are not well understood. However, a number of large-scale projects have pointed to the efficacy of taking systemic approaches to addressing support of students along the trajectory, especially those from underrepresented populations. This work has also pointed to the importance of paying close attention to the critical juncture between secondary and post-secondary education. Part of the way that this juncture can be best supported is through an integrated, systemic effort to collect, share, and analyze data on students, teachers, and programs that factor into STEM pipeline support. This data can be used directly to help make good decisions for individual students but, perhaps most importantly, in an analytical and strategic way, to learn about the efficacy of different STEM outreach and support approaches. The same networks that have been used to collect data and communicate findings should receive feedback from participating programs as part of a process of continuous improvement.

 

 

 

The material on this website is based upon work supported by the National Science Foundation under Grant No. 1038154.   Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the view of the National Science Foundation.