What Works

Participants

Four different models of specialized science high schools were included in the study. Residential schools serve high-ability students from all around the state, ensuring that every county has a chance at representation in the entering class. Comprehensive or self-contained schools are usually located in large metropolitan areas, serving an entire district or a city’s most talented students. Schools-within-schools are also typically located in urban areas and were often founded to bring cohorts of academically talented students and additional resources to schools with fewer resources. The schools-within-schools model allows students to participate in classes with their schoolmates outside of the specialized school curriculum and for some non-enrolled students in the school to take advanced courses in the program. The fourth category of specialized science schools is the half-day or part-time school model available in Virginia and Michigan, usually in poorly resourced or rural counties. These schools serve a geographical region, busing in high-achieving students from a number of secondary schools to participate in advanced science coursework in science-related disciplines for part of the day. The rest of the day, students are enrolled in their home high school.

Using data from 3,526 respondents, a response rate of 43.5%, from four different models of specialized science high schools (residential schools, comprehensive high schools, school within a school, and shared-day regional schools), our team of researchers was able to identify specific characteristics of teachers and classrooms associated with increasing the number of students in the STEM workforce or in STEM pathways (see Almarode et al., 2014; Subotnik, Edmiston, Lee, Almarode, & Tai, 2012; Subotnik, Tai, & Almarode, 2011; Subotnik, Tai, Almarode, & Crowe, 2013). Demographic information about the respondents is shown in Table 1.

OPEN IN VIEWER

Table 1.

Demographic Characteristics of Respondents

	Specialized HS graduates (N = 3,526)
Male	45.0%
Female	51.2%
Hispanic origin	4.3%
Caucasian	70.1%
African American	4.6%
Asian	21.2%
American Indian/Alaskan Native	1.1%
Native Hawaiian/Pacific Islander	0.3%
Other	3.7%
Immigrant status: Respondent	12.2%
Immigrant status: Male parent/guardian	28.6%
Immigrant status: Female parent/guardian	29.8%
English as primary language	78.8%
Parent/guardian education beyond high school	81.2%
Parent/guardian in STEM-related career	56.7%

Note. STEM = science, technology, engineering, and/or mathematics; HS = high school.

Data Collection

In addition to reporting on their demographics and undergraduate major or concentration, participants were asked about the specific characteristics of their high school, including identifying the frequency (e.g., not at all, once a month, once a week, several times per week, or every day) of several practices or strategies they experienced. We focus this article on the outcomes of our investigation of practices, strategies, and characteristics from the literature on effective science education listed in Table 2.

OPEN IN VIEWER

Table 2.

Classroom Practices and Strategies and Teacher Characteristics Identified in Respondents’ Specialized Science High School

Classroom practices and strategies	Teacher characteristics
Laboratory experiences (Labs)	Encouraged students to pursue own interests
Hands-on experiences	Show enthusiasm for STEM
Field trips	Show expertise in STEM
Guest lectures from scientists	Discuss interaction of STEM and society
Demonstrations	Encouraged participation in academic competitions
Internships/mentorships	Encouraged students to challenge and ask why
Asking questions with pre-determined answers	Enhance awareness of STEM careers
Asking questions with no pre-determined answers	Build self-confidence in STEM activities
Assigning collaborative projects	Accept student criticism of points made in class
Accelerating the pace of instruction
Focus on deep understanding of complex content
Make connections to other content areas

Note. STEM = science, technology, engineering, and/or mathematics.

Classroom practices and strategies were further categorized by teacher behaviors versus student experiences. For example, asking questions with pre-determined answers, asking questions with no-predetermined answers, assigning collaborative projects, accelerating the pace of instruction, focusing on deep understanding of complex content, and making connections to other content areas were categorized as teacher behaviors in the survey. We categorized hands-on experiences, field trips, guest lectures, demonstrations, and internships or mentorships as student experiences.

Data Analysis and Findings

Data were analyzed to determine classroom practices and strategies and teacher characteristics that related to an undergraduate degree in STEM fields.

Classroom practices and strategies

Logistic regression ¹ as applied to our study variables estimates the probability that a respondent will report earning an undergraduate degree in a STEM-related area (our primary outcome of interest) if they also report experiencing certain classroom practices and/or teacher characteristics. We can ask, for example, if a respondent indicates that he or she engaged in frequent hands-on experiences while attending his or her specialized science high school, will he or she then be more likely to report earning an undergraduate degree in a STEM-related area? Of the 3,526 respondents to the survey, almost 2,600 provided responses to this set of questions on teacher behaviors and student experiences, allowing us to draw inferences from the logistic regression model.

We found that certain classroom practices and strategies were associated with a greater likelihood of respondents reporting an undergraduate degree in STEM, while others were either not associated with this outcome or actually had a negative association with this outcome. Some of the outcomes were very surprising and counterintuitive to what we expected. For example, teachers asking questions with pre-determined answers as opposed to open-ended questions were positively associated with the reported earning of a STEM-related degree. Other outcomes associated with completing a STEM degree were less surprising, including teachers making connections between STEM subjects and other content areas, teachers focusing on deep conceptual understanding of complex content, and participation in an internship or mentorship. With logistic regression analysis, these positive associations are represented by an odds ratio that expresses the likelihood of earning a degree in a STEM-related area.

In each case of the analyses below, the predictor variables are classroom practices, strategies, and teacher characteristics, and the specific outcome of interest is reporting an undergraduate degree in a STEM-related area.

Odds ratios and their interpretations for classroom practices and strategies are shown in Table 3.

OPEN IN VIEWER

Table 3.

Classroom Practices and Strategies Positively Associated With the Reported Earning of a STEM-Related Degree

Classroom practices and strategies	Odds ratio	Outcome
Teachers asking questions with pre-determined answers	1.156	Respondents reporting that they were frequently asked questions with pre-determined answers were approximately 15% more likely to report earning a STEM-related degree rather than a degree in a non-STEM–related area.
Teachers making connections with other content areas	1.119	Respondents reporting that teachers made connections between STEM and other content areas were made in their classroom were approximately 11% more likely to report earning a STEM-related degree rather than a degree in a non-STEM–related area
Teachers focusing on deep understanding of complex content	1.250	Respondents reporting that their classes focused on deep understanding of complex content were approximately 25% more likely to report earning a STEM-related degree rather than a degree in a non-STEM–related area.
Participating in an internship or mentorship	1.173	Respondents reporting that they participated in an internship or mentorship were approximately 17% more likely to report earning a STEM-related degree rather than a degree in non-STEM–related area.

Note. STEM = science, technology, engineering, and/or mathematics.

Given what is often called for in the literature on science learning, we were surprised to find that collaborative assignments were negatively associated with the reported earning of a STEM-related degree. Specifically, respondents that reported frequent assignments that required working collaboratively with peers or in teams were approximately 17% less likely to report earning a STEM-related degree. In addition, although programs for high-achieving students are proud of the advanced pace of their curriculum and instruction, moving at a rate that was considered too fast for their students was associated with a 22% smaller likelihood of reporting a STEM-related undergraduate degree. For this particular question, respondents were asked how frequently their teachers moved at a pace that was too fast for their abilities and understanding of the course content.

None of the remaining classroom practices and strategies on the list above—laboratory experiences, hands-on experiences, field trips, guest lectures, demonstrations, and teachers asking questions with no pre-determined answers—was associated with the reported earning of a STEM-related degree.

Teacher characteristics

Just as with classroom practices and strategies, we found that certain teacher characteristics were associated with a greater likelihood of respondents reporting an undergraduate degree in STEM while others were either not associated with this outcome or actually had a negative association. Respondents reporting that their teachers accepted criticism of the points they were making and that their teachers reinforced their self-confidence in STEM activities were positively associated with earning a STEM-related degree. The corresponding odds ratios and interpretations are shown in Table 4.

OPEN IN VIEWER

Table 4.

Teacher Characteristics Positively Associated With the Reported Earning of a STEM-Related Degree

Teacher characteristics	Odds ratio	Interpretation
Accepting criticism	1.168	Respondents that reported that their teacher accepted criticism of his or her ideas were approximately 16% more likely to report earning a STEM-related degree in a non-STEM–related area.
Reinforcing self-confidence in STEM activities	1.569	Respondents that reported that their teacher reinforced their self-confidence in STEM activities were approximately 56% more likely to report earning a STEM-related degree rather than a degree in a non-STEM–related area.

Note. STEM = science, technology, engineering, and/or mathematics.

With regard to accepting criticism, the particular survey item was phrased as “To what extent did your high school STEM teachers accept student criticism of the points they were making.” Thus, this finding may suggest that when students are allowed to engage in discourse about scientific ideas, they are offered a greater opportunity to engage intellectually in the enterprise of science.

Surprisingly, respondents who reported that their teachers frequently discussed the interaction between STEM and society were 21% less likely to report earning a STEM-related degree. The remaining teacher characteristics (i.e., enthusiasm, expertise, encourages students to participate in academic competitions, encourages asking “why?” enhances students’ career awareness) were not associated with the reporting of a STEM-related degree.

Classroom Practices and Strategies

Teachers that make constructive connections with other subject areas within and outside of STEM, as well as focusing on deep understanding, promote a high level of cognitive engagement above and beyond listening to a guest lecture or watching a demonstration. Similarly, having the opportunity to “do science” with a “real scientist” through an internship or mentorship represents a high level of intellectual engagement beyond a preconceived hands-on activity or laboratory experiment. The pace of instruction should also match the students’ level of understanding, and understanding should not be sacrificed in the name of speed. As this particular finding suggests, moving at a pace that is too fast for the students’ abilities and understanding of the content, regardless of their talent, could be detrimental to their long-term engagement and persistence in STEM. This, of course, is different from the well-supported idea of acceleration through content that the student is able to master at a more rapid pace (see Colangelo, Assouline, & Gross, 2004; Colangelo & Davis, 2003; C.-L. Kulik & Kulik, 1982; J. A. Kulik & Kulik, 1984). The message here may be that in specialized science high schools, the content will challenge these individuals and thus requires additional time to progress toward proficiency and mastery in that content.

There were two additional surprising findings: (a) the relative ineffectiveness of working collaboratively with peers and, (b) the relative effectiveness of receiving teacher-directed instruction. Although difficult to provide an evidence-based rationale for these outcomes, the first finding suggests that students are more successful at building teacher-directed background and foundation knowledge individually. This finding is in line with the work of Klahr and Nigam (2004), as well as Kirschner, Sweller, and Clark (2006) and Hattie (2009) reporting on the effectiveness of direct instruction over other approaches (e.g., problem-based learning). What this finding does not imply is that we should abandon open-ended questioning or inquiry and require students to work by themselves. For example, teachers may want to consider the use of more convergent than divergent questions in light of these findings in science classrooms. Less formal contexts may wish to encourage both convergent and divergent question use with gifted learners. Similarly, administrators and teachers should keep in mind that students may prefer to devote some of their learning time to individually acquiring, processing, and reflecting on content as they build background and foundational knowledge in STEM classrooms.

Teacher Characteristics

In terms of teacher characteristics, these findings suggest that building the self-efficacy of talented students and permitting them to engage in critical conversations about content further enhance the cognitive engagement inside the classrooms of these specialized science high schools. As suggested by the body of work around classroom discussions and scientific discourse, these conversations are best facilitated and guided by the teacher (see Hattie, 2012).

Other findings related to the teacher characteristics variables do not align with mainstream beliefs in STEM education. For example, respondents who reported that their teachers frequently discussed the interaction between STEM and society were less likely to report earning a STEM-related degree. Also enthusiasm, expertise, encouraging students to participate in academic competitions, and encouraging the asking of “why?” enhancing students’ career awareness) were also not associated with the reporting of a STEM-related degree. As stated previously, these counterintuitive findings should not prompt us to eliminate efforts in areas like academic competition, career counseling, encouraging students to inquire about their environment, or, certainly not being enthusiastic about the content in each of the STEM disciplines. Instead, these findings should prompt us to explore the differential effects of these characteristics on different subgroups.