The Effects of a Science-Focused STEM Intervention on Gifted Elementary Students’ Science Knowledge and Skills

Abstract

To develop Science, Technology, Engineering, and Mathematics (STEM) talents, both researchers and policy developers recommend that educators begin early. In this randomized study, we document the efficacy of teacher professional development and a rich problem-based inquiry curriculum to develop the science talent of elementary students. The intervention, STEM Starters, a federally funded Jacob K. Javits project, provided sustained and embedded professional development to classroom teachers and to pull-out gifted program teachers to support the implementation of a problem-based curriculum in their classrooms. During the intervention, randomly assigned teachers participated in 120 hr of professional development that focused on science content, inquiry-based instruction, technological applications, and differentiated instruction within problem-based curriculum units. Statistically significant gains in science process skills, science concepts, and science content knowledge were found among gifted students in the treatment group when compared with gifted students in the comparison group.

Keywords

gifted inquiry instruction problem-based curriculum professional development science STEM

The National Science Board (NSB; 2010) stressed the importance of Science, Technology, Engineering, and Mathematics (STEM) innovators for the long-term prosperity of our nation. Future STEM innovators are dependent upon an educational system to identify and nurture their STEM talents and to guide their career decisions. The NSB recommended STEM opportunities for all students begin in the elementary grades and include inquiry-based learning, peer collaboration, and open-ended, real-world problem solving. To foster STEM talent, the NSB also recommended policy actions to encourage research-based STEM professional development for elementary teachers that will cultivate investigative classrooms and identify promising STEM learners.

Talent development in science should begin in the early grades with investigative opportunities to encourage curiosity and engagement so that interest in science will be maintained as students progress across grade levels (Brandwein, 1995; Maltese & Tai, 2010; S. P. Marshall, McGee, McLaren, & Veal, 2011; Metz, 2008; National Association for Gifted Children, Task Force on Math and Science, 2008; Roberts, 2010). In addition, Keeley (2009) recommended early intervention in science to maximize the cumulative learning processes critical to science talent development. In support of early opportunities, researchers argued that science interest is usually ignited before middle school and is instrumental in motivating students to develop their talents and to pursue science as a career option (Lindahl, 2007; Maltese & Tai, 2010).

Perspectives and Theoretical Framework

The theoretical framework of this study was guided by the principle that implementation of a problem-based science curriculum enacted by teachers supported with sustained professional development can positively influence STEM accomplishments in young students. Researchers maintained that the combination of teacher professional development and the implementation of a problem-based/inquiry curriculum impacts teacher instruction and, ultimately, student achievement (Buczynski & Hansen, 2010; Johnson, Kahle, & Fargo, 2007; Santau, Maerten-Rivera, & Huggins, 2011). For example, Santau et al. (2011) examined the effects of a 3-year professional development intervention focused on inquiry on the science achievement of Grade 4 English language learners. In an assessment created from the National Assessment of Educational Progress (NAEP) and the Trends in International Mathematics and Science Study (TIMSS) released test items, researchers found students significantly improved their science achievement scores from pre- to posttest. When comparing student scores with national and international norm groups, students in the intervention group outperformed the norm group across similar items on the posttest despite scoring lower than the norm group on the pretest.

In a similar study, Cotabish, Dailey, Robinson, and Hughes (2013) reported increased achievement for general education students on measures of science process skills, content, and concept knowledge after their teachers participated in 2 years of professional development focused on improving science instruction within the context of a problem-based learning (PrBL) curriculum. In another intervention, Project Clarion, teachers participated in professional development focused on implementing inquiry-based science. Students, including those from low socio-economic backgrounds, who received services for 3 years, demonstrated increased science achievement when compared with students who had not participated in the program (Kim et al., 2012). Taken together, these recent studies documented the value of sustained professional development and the implementation of a problem-based/inquiry curriculum on student achievement in science. Part of a larger study reported elsewhere (Cotabish et al., 2013), the current study extends this research literature by disaggregating the effects and focusing specifically on elementary students in gifted and talented programs.

Developing the STEM Talent of Students

To encourage the development of STEM talent in the elementary grades, a demonstration project, STEM Starters, was developed and subsequently funded through the Jacob K. Javits Gifted and Talented Students Education Program. STEM Starters was primarily focused on the discipline of science with the integration of mathematics and engineering concepts. The project provided sustained and embedded professional development to classroom and gifted program pull-out teachers to support the implementation of problem-based curriculum units for students in Grades 2 through 5. During the intervention, teachers participated in 120 hr of professional development that focused on STEM content, inquiry-based instruction, specific problem-based curriculum units, technological applications, and differentiation of instruction. During the school year, embedded professional development, defined in STEM Starters as peer coaching, supported teachers in the implementation of the curriculum and in building science content knowledge.

Professional development

Roberts (2010) recommended that teachers view themselves as talent developers and understand that gifted students, regardless of background, are entitled to opportunities for educational excellence in STEM disciplines. To provide these opportunities, teachers need ongoing professional development and support to engage gifted students in quality learning endeavors such as inquiry-based learning and PrBL (C. A. Little & Paul, 2011; J. C. Marshall & Horton, 2011). In particular, researchers found that key factors in professional development that lead to change in the classroom included extended contact time (Gerard, Varma, Corliss, & Linn, 2011; Johnson & Fargo, 2010; Yoon, Duncan, Lee, & Shapley, 2008), follow-up support (Appleton, 2008; Lumpe, Czerniak, Haney, & Beltyukova, 2012), and explicit instruction on teaching practices using classroom-specific curriculum (Penuel, Gallagher, & Moorthy, 2011). Each of these features was incorporated into the STEM Starters intervention.

Extended contact time

In a review of professional development studies, Darling-Hammond, Wei, Andree, Richardson, and Orphanos (2009) found evidence supporting extended contact hours for increased teacher and student learning. In particular, their review indicated the greater the intensity and duration of professional development, the greater the impact on teachers and students. In other studies supporting extended contact time, Corcoran, McVay, and Riordan (2003) and Supovitz and Turner (2000) found changes in inquiry-based instruction did not occur until after 80+ hr of professional development. In a similar study, Roehrig, Dubosarsky, Mason, Carlson, and Murphy (2011) reported that it took a minimum of 180 hr of professional development for teachers to fully integrate inquiry-based strategies.

The duration of professional development can also impact student achievement. Johnson et al. (2007) found a statistically significant difference in science achievement among students in the experimental group when compared with the control group after teachers had participated in a 3-year longitudinal study involving 100+ hr of science professional development. In particular, Johnson et al. reported that statistically significant differences in student achievement were not found in the first year (M_e = 8.93, SD = 3.13; M_C = 8.28, SD = 3.96) of the study but were found in Year 2 (M_e = 12.28, SD = 3.16; M_C = 8.17, SD = 3.02) and Year 3 (M_e = 13.28, SD = 3.36; M_C = 8.16, SD = 3.19) indicating that the duration of professional development ultimately impacted student achievement results. In a similar study, Shymansky, Wang, Annetta, Yore, and Everett (2010) also found a positive relationship between student achievement scores and the number of hours teachers participated in a professional development program, F(1, 32) = 9.43, p = .004, R² = .233.

Follow-up support

Simon, Campbell, Johnson, and Stylianidou (2011) argued that teacher learning is complex and that external professional development experiences with no follow-up support are unlikely to create change in teaching practices. The researchers found that support systems were essential features of effective professional development. In particular, teachers expressed the value of peer observations, constructive feedback, and one-to-one meetings with mentors in changing their own instructional practices.

Teachers benefit from having classroom-embedded follow-up support to assist them in implementing new instructional strategies. In particular, Schleigh, Bosse, and Lee (2011) recommended that professional development be situated in teachers’ own classrooms. Peer coaching is an example of this type of professional development that allows teachers to transfer newly acquired learning and skills from the workshop to the classroom by providing teachers a natural support system (P. F. B. Little, 2005; Showers & Joyce, 1996). For example, Showers (1982) found that coaching was a statistically significant predictor of teachers’ ability to transfer training, F(1, 15) = 8.20, p < .01, from the professional development workshop to the classroom. With respect to student achievement, Showers (1984) found coaching also contributed to improved student scores on a science concept measure, F(2, 138) = 4.34, p = .01.

Explicit instruction on teaching practices

To improve elementary science instruction, the National Research Council (NRC) recommended that professional development provide teachers with explicit models of how to teach science (Duschl, Schweingruber, & Shouse, 2007). The NRC also suggested professional development mirror the instruction that will occur in the classroom by using the classroom-based curriculum selected for implementation. According to Desimone, Porter, Garet, Yoon, and Birman (2002), to improve inquiry-based instruction, professional development must be focused on the specific practice used by the teachers. For example, Penuel et al. (2011) compared three professional development programs to analyze differences in student learning in science. All three programs were similar except for their use of explicit instruction on pedagogical strategies provided to the teachers. The two programs that used explicit teaching instructions resulted in statistically significant increases in student science achievement with effect sizes of 0.34 and 0.29, whereas the program without explicit instructions failed to demonstrate significant increases in student science achievement.

In summary, professional development that contributes to teacher change and ultimately to student achievement requires adequate contact hours, follow-up support, and explicit instruction on teaching practices. Unfortunately, professional development utilizing practices deemed most effective are not common across school-wide communities (Darling-Hammond et al., 2009). For example, an examination of the professional development prevalent in the National Science Foundation’s Math and Science Partnership program found the majority of activities were courses, workshops, and institutes (63%), and less than 2.5% of them offered mentoring/coaching, internships, learning communities, or teacher collaborations (Moyer-Packenham, Bolyard, Oh, & Cerar, 2011).

Science curriculum for high-ability learners

In addition to providing elementary teachers with sufficient science preparation, high-ability students need early experiences in science that serve to increase their interest and engagement in the discipline (S. P. Marshall et al., 2011; Robbins, 2011; VanTassel-Baska, 1998). In a review of the literature, Robinson, Shore, and Enersen (2007) suggested the traditional use of basal texts with the superficial study of a wide-range of topics does little to promote high-ability students’ interest or engagement in science. Instead, to encourage student interest and enthusiasm in science, science curriculum should emphasize overarching concepts, higher level thinking, inquiry, technology, and scientific processes (VanTassel-Baska, 1998). More recently, Robbins suggested curriculum provide opportunities for students to apply scientific reasoning, encourage reflection and collaboration, engage students in quantitative problem solving, and expose students to the real work of scientists. These strategies emphasized by VanTassel-Baska and Robbins are typically found in a PrBL curriculum, which has been found to be effective with gifted students (Gallagher & Stepien, 1996; Gallagher, Stepien, & Rosenthal, 1992).

Science curriculum units, developed by researchers at the College of William and Mary, incorporate a problem-based unit design. These units emphasize student problem solving, scientific research, and experimental design, and focus on overarching concepts that provide a systematic link across the content. In a national study on one of the units, Acid, Acid, Everywhere, VanTassel-Baska, Bass, Ries, Poland, and Avery (1998) indicated that classrooms using the problem-based units demonstrated greater student interest, enthusiasm, and engagement when compared with classrooms using the traditional curriculum.

In another study using the William and Mary science units, Feng, VanTassel-Baska, Quek, Bai, and Oneill (2005) reported increased scores in science research skills among gifted students in Grades 3 to 5 (Grade 3: d = 1.37; Grade 4: d = 1.09; Grade 5: d = 1.00). In a similar study on the units, Kim and colleagues (2012) found increased science achievement (η²_p = 0.132) among all students, including those from low socio-economic backgrounds. As the evidence revealed, curriculum based on understanding concepts, real-world problem solving, and inquiry-based learning benefited high-ability learners leading to greater interest, engagement, and ultimately achievement (Feng et al., 2005; Kim et al., 2012; VanTassel-Baska et al., 1998).

Akinoglu and Tandogan (2007) stated that a PrBL curriculum increases student achievement and engagement attributable to the focus on active learning, real-world problem solving, and collaboration among stakeholders. They reported that classrooms using PrBL curriculum resulted in students scoring statistically significantly greater on measures of science achievement and student engagement when compared with students in a traditional science class—science achievement: t(48) = −2.27, p < .05; student engagement: t(48) = −2.34, p < .05 (Akinoglu & Tandogan, 2007). In a similar study, Drake and Long (2009) compared the effects of PrBL versus direct instruction on two groups of students. Both groups were exposed to the same science curriculum unit, but the interventions differed in the mode of instruction. Drake and Long reported positive results among students who participated in PrBL. For example, on a measure of content knowledge, the PrBL students demonstrated a greater gain in content knowledge than students in the direct instruction condition, t(27) = −1.85, p = .038. In addition, PrBL students exhibited greater time-on-task behavior than students involved in direct instruction and displayed greater skills in identifying problem-solving strategies and appropriate resources.

In addition to utilizing PrBL curriculum in elementary science classrooms, researchers have also recommended integrating science into literacy instruction to increase opportunities for science lessons (Cervetti, Barber, Dorph, Pearson, & Goldschmidt, 2012; Romance & Vitale, 2011). Through the Science IDEAS model, teachers used an interdisciplinary approach to teaching literacy through a science discipline (Romance & Vitale, 2011). The curriculum involved students in hands-on experiments, reading and writing about science, reflection through science journaling, and concept mapping. Romance and Vitale reported significant increases in both science and reading comprehension achievement scores for students who participated in Science IDEAS in Grades 3 to 5 when compared with students who received traditional instruction. In an experimental study, Cervetti and colleagues found that implementing a science-literacy integrated curriculum with an emphasis on inquiry-based instruction also resulted in improved science learning and literacy outcomes among Grade 4 students. In particular, researchers reported moderate effect sizes (ES = 0.65) for the treatment students on a measure of science understanding and on a measure of science writing (ES = 0.40).

In summary, teacher professional development and a rich problem-based/inquiry curriculum are two key features necessary to develop the science talent of students. Engaging teachers in professional development opportunities that involve extended contact time and follow-up support, and providing teachers with explicit instruction on teaching practices are key factors in enacting change in the classroom. In conjunction with an inquiry science curriculum that provides students with opportunities to solve real-world problems and that can be integrated with literacy components, interventions such as STEM Starters support the development of science talent in young students.

Purpose of the Study

The purpose of the study was to measure the impact of a STEM intervention on gifted students’ science learning, including science process skills, content knowledge, and concept knowledge. The mediating effect of the STEM intervention on the science learning of teachers and general education students has been reported elsewhere (Cotabish, Dailey, Hughes, & Robinson, 2011; Cotabish et al., 2013). The STEM Starters intervention included sustained, embedded professional development in science for teachers and the implementation of a problem-based science curriculum in both general education grade-level classrooms and gifted and talented program pull-out classrooms. The research questions were as follows:

Research Question 1: To what degree are elementary gifted students’ understandings of science process skills impacted by participation in a STEM intervention?

Research Question 2: To what degree is elementary gifted students’ science content knowledge impacted by participation in a STEM intervention?

Research Question 3: To what degree is elementary gifted students’ science concept knowledge impacted by participation in a STEM intervention?

Method

Design

The current study was part of a larger randomized field study on the effects of teacher professional development and the implementation of an inquiry-based curriculum on teaching and learning in science. Only results for identified gifted students are reported here. Randomly selected from five low-income schools in a southern state, 70 teachers from Grades 2 through 5 were assigned to the experimental and control conditions. Students assigned to experimental teachers were designated as students in treatment classrooms, and students assigned to control teachers were designated as students in comparison classrooms. Randomization occurred at the teacher level. The number of gifted students per grade-level treatment and comparison classrooms varied with most classes containing only one, two, or three gifted students. Students in the treatment classrooms received 1 to 2 years of intervention depending on their grade level. For example, students who were in Grade 5 during Year 1 participated in only 1 year of intervention because the project concluded at Grade 5. Students who were in Grade 2 during Year 2 received 1 year of intervention.

Participants

Students were identified as gifted according to state guidelines that included multiple criteria, such as standardized achievement test scores, cognitive ability tests (verbal and nonverbal), creativity measures, and teacher and parent recommendations. Table 1 summarizes the number of gifted students in the treatment and comparison groups in this study.

Table 1.

Number of Treatment and Comparison Students in the Gifted Program by Grade Level and Year.

	Treatment students		Comparison students
	Year 1	Year 2	Year 1	Year 2
Grade level	Gifted	Gifted	Gifted	Gifted
2	20	3^a	17	2^a
3	25	18	20	17
4	24	25	20	20
5	18	21	13	21
Total	87	67	70	60

Due to a school district policy change, fewer students were identified as gifted in Year 2 of the study.

Intervention

The main components of the intervention were as follows: (a) inquiry- or problem-based science curriculum units and (b) teacher professional development, totaling 120 hr per teacher over 2 years.

Curriculum

The William and Mary science curriculum units utilized in this study situated science learning in the context of a real-world problem (VanTassel-Baska & Stambaugh, 2008). Each unit introduced students to advanced content, engaged students in problem solving and critical thinking, and was focused on specific overarching concepts that were integrated throughout the unit, including change (Grades 2 and 3) and systems (Grades 4 and 5). In addition to the William and Mary curriculum units, students were engaged in the study of scientists and inventors through the Blueprints for Biography®-STEM Series curriculum units. Blueprints for Biography® are a series of teacher curriculum guides with high-level discussion questions, creative and critical thinking activities, a persuasive writing component, and rich primary resources (Robinson, 2006). STEM Blueprints focused on eminent individuals for whom exemplary children’s biographies existed in trade book form. Each guide concluded with a classic science experiment for students to carry out.

During each year of the intervention, students in the treatment classrooms received instruction in two William and Mary units and one STEM Blueprint; whereas, the gifted students in the comparison classrooms received science instruction using the school-adopted science curriculum, and their teachers conducted science as usual. Both experimental and control teachers followed the Arkansas State Science Frameworks to guide their science instruction; therefore, students were being taught the same content and overarching concepts, but the experimental teachers were using the curriculum implemented through STEM Starters. Meanwhile, the control teachers utilized Harcourt Brace Jovanovich (HBJ) readers to read about the science topics in the early grades and Scott Foresman texts in Grades 4 and 5. The curriculum units, their discipline foci, and the biographies used in the treatment classrooms are summarized in Table 2.

Table 2.

Curriculum Units by Grade Level for Treatment Students in the Gifted Program.

Grade level	Problem-based curriculum units and interdisciplinary STEM connections	Blueprints for Biography®	Arkansas State Science Standards^a
2	Weather Reporter (Earth and Physical Science; Mathematics—Measurement & Data) Budding Botanist (Life Science; Mathematics—Measurement & Data)	G. W. Carver Louis Pasteur	NS. 1.2.1-1.2.8 PS. 5.2.2; 7.2.2 ESS. 8.2.1-8.2.3; ESS. 8.2.5-8.2.9 LS. 2.2.3-2.2.4; 2.2.6
3	What’s the Matter (Physical Science; Mathematics—Measurement & Data) Dig It (Earth Science; Engineering Design; Mathematics—Measurement & Data)	Galileo Galilei Albert Einstein	NS. 1.3.1-1.3.9 PS. 5.3.1-5.3.4; 7.3.2 ES. 8.3.1; 9.3.1
4	Electricity City (Physical Science; Engineering Design; Mathematics—Measurement & Data) Invitation to Invent (Physical Science; Engineering Design; Mathematics—Measurement & Data)	Thomas Edison Michael Faraday	NS. 1.4.1-1.4.2; 1.4.4-1.4.14 PS. 6.4.1-6.4.2; 7.4.2-7.4.3
5	Acid, Acid Everywhere (Earth and Physical Science; Engineering Design; Mathematics—Measurement & Data) Nuclear Energy (Earth and Physical Science; Engineering Design; Mathematics—Measurement & Data, Operations & Algebraic Thinking)	Marie Curie A. Graham Bell	NS. 1.5.1-1.5.9 LS. 4.5.1; 4.5.3-4.5.4; 4.5.12; 4.5.16; 4.5.18 PS. 5.5.1-5.5.5; 5.5.7

Note. STEM = Science, Technology, Engineering, and Mathematics.

Description of specific frameworks found in the appendix.

Teacher professional development

Across two summers, STEM Starters teachers, including gifted and talented teachers and facilitators, participated in week-long summer institutes focused on science content, inquiry-based instruction, specific curriculum units, technological applications, differentiation of instruction, and biography study. As depicted in Table 3, the summer institutes provided 60 hr of professional development for the implementation of the curriculum units. The institutes were structured so that teachers took the role of students while expert science instructors led them through the problem-solving units modeling effective science instruction. As recommended by VanTassel-Baska (1998), instructional emphasis was placed on overarching concepts, higher order thinking skills, inquiry-based learning, experimental design, and the use of technology.

Table 3.

Teacher Professional Development Across 2 Years.

	Summer institute	Peer coaching
Year 1	30 hr (out-of-school) Curriculum units Inquiry-based strategies Technology applications	30 hr (in-school) Implementation of curriculum units Demonstration lessons in teachers’ classrooms Instructional facilitator Materials facilitator Science content expert
Year 2	30 hr (out-of-school) Science content from the specific curriculum units Inquiry-based strategies Classroom management Technology applications	30 hr (in-school) Instructional facilitator Materials facilitator Science content expert

In addition to the summer institutes, STEM Starters provided peer coaching on a weekly basis to the participant teachers. The peer coach was a former secondary science and elementary gifted and talented teacher. Once the school year began, the peer coach extended the institute learning by providing embedded, classroom support to each teacher. In the classroom, the peer coach modeled the lesson, assisted the teacher with instruction, and monitored and encouraged student involvement. Outside the classroom, the peer coach made certain that all necessary science activity materials were in the schools and maintained contact with all teachers by phone or email to ensure their needs were being met. As summarized in Table 3, peer coaching support occurred across 2 years, providing each teacher with 60 additional hr of professional development.

Instrumentation

Raters

The research team recruited undergraduate education majors and in-service teachers to assist in scoring student assessments that included problem-based and open-ended items. Each rater participated in 6 hr of training led by two STEM Starters science experts. Raters were provided with an overview of the curriculum, focusing specifically on the assessments. The pool of potential raters practiced grading each question and then participated in a discussion facilitated by the science experts. Potential raters assessed a content, a process, and a concept test from all grade levels, and only those who achieved an inter-rater reliability, in this case percentage of agreement (.90) with the ratings of the STEM Starters science experts, were retained for the study. A total of seven raters from the pool met the criterion and were used to score student responses for the study. During supervised scoring, the science experts randomly selected assessments to monitor consistency and accuracy and to ensure that inter-rater reliability remained above .90.

Student science process skills

To address Research Question 1, the Scoring Rubric for Scientific Processes–Diet Cola Test (Fowler, 1990) was used to assess students’ understanding of the design of science experiments. For example, participants were asked to design an experiment to answer the question, “Are earthworms attracted to light?” Participant responses to the questions were scored across five criteria: (a) generates a prediction, (b) lists materials needed, (c) lists experiment steps/arranges steps in sequential order, (d) plans the data collection, and (e) states plan for interpreting data for making predictions. Ratings range from no evidence (0) to strong evidence (3) with two additional points possible on one criterion resulting in 17 points possible.

Callahan, Hunsaker, Adams, Moore, and Bland (1995) described the Diet Cola Test as having promise for evaluating science process skills. The researchers stated the content validity of the instrument was adequate due to the match between the instrument’s task and the literature-supported criteria for science aptitude. Adams and Callahan (1995) reported alternate-form reliability to be .76, and inter-rater reliabilities ranged from .90 to .95 when calculated among four raters. According to Adams and Callahan and Callahan et al., the examination of convergent and discriminant validity revealed significant but weak patterns of correlation with other science process skills tests (e.g., the Group Embedded Figures Test and the Test of Basic Process Skills), thereby limiting the use of scores for making decisions for individuals; however, the results did support its use for decisions about groups of participants.

Student knowledge of science content

To address Research Question 2, students’ science content knowledge was assessed using pre–post embedded curriculum-based assessments for each unit to capture student learning gains. These assessments were specifically tied to the content in each curriculum unit. From the results of field testing these units and their embedded assessments, VanTassel-Baska and Stambaugh (2008) reported internal consistency reliability for scores on the content assessment to be .69 and the inter-rater reliability to be .89. The assessment format varied across grade levels. For example, in Grades 2 to 3, the curriculum-based content assessments utilized open-ended concept mapping to measure students’ understanding of selected content topics in science, whereas the students in Grades 4 to 5 were assessed with open-ended short-answer questions. Students from both the treatment and comparison groups were familiar with concept mapping, a tool commonly used to organize ideas during prewriting exercises. Both the Arkansas Department of Education (ADE; 2003) English Language Arts Curriculum Frameworks and the Common Core State Standards (CCSS; 2012) emphasized the use of graphic organizers such as concept mapping. In addition, the Arkansas Science Frameworks encouraged the use of graphic organizers to communicate observations (ADE, 2003). Prior to the pretests and posttests, students in both the treatment and comparison groups were given instructions on how to generate a concept map.

Student knowledge of science concepts

To address Research Question 3, students’ science concept knowledge was assessed using pre–post embedded curriculum-based assessments from each unit to capture student learning gains. These assessments were specifically tied to the content and overarching concepts in each curriculum unit. From the results of field testing these units, VanTassel-Baska and Stambaugh (2008) reported internal consistency reliability for scores on the concept assessment to be .68 and the inter-rater reliability to be .85. Using open-ended short-answer questions, the curriculum-based assessments measured students’ understanding of overarching concepts that unify STEM disciplines (e.g., systems and change) and their connection to the curriculum-based science content and process skills.

Data Collection

Data from the three instruments were collected from students in the treatment and comparison classrooms at the beginning of Year 1 (September 2009), at the end of Year 1 (May 2010), at the beginning of Year 2 (September 2010), and at the end of Year 2 (December 2010). Data were collected by teachers and provided to project staff and researchers for scoring. The STEM Starters science expert and peer coach was frequently in the buildings and responded to teacher questions and monitored the data collection fidelity and schedule.

Results

Analytic Strategy

All data were entered into SPSS Version 20 for analysis. As a conservative analytic strategy, researchers used listwise deletion of missing data; all cases remained in the sample, and participants were excluded from analysis only if they had missing data on the variables needed for that analysis; therefore, sample sizes for each statistical test differ. The point values for the Grade 5 versions of the Concept and Content tests differed from the Grades 2 to 4 tests so only Grades 2 to 4 were included in the Concept and Content analyses. Although multilevel modeling is often the most appropriate technique for analyzing data collected from students nested within classrooms (O’Connell & McCoach, 2008), in this study, gifted students were spread across classrooms and do not meet the suggested minimum cluster size of 10 (Bickel, 2007). For each measure (Process, Concept, and Content), researchers planned to conduct 2 one-way ANCOVA tests (one for the Year 1 data and one for the Year 2 data) to compare students’ posttest scores between the treatment and comparison groups using the pretest scores as a covariate to control for initial differences between the two groups on the measure. The researchers also planned to use eta squared to report the effect sizes. Eta squared is the proportion of the total variance that is attributed to the independent variable (Becker, 1999). Each analysis began by determining the reasonableness of the assumptions before interpretation. Researchers evaluated the homogeneity-of-slopes assumption by examining the interaction term between the group variable and covariate and used Levene’s test for the homogeneity-of-variance assumption. For the Year 1 data, there was a difference between the size of the treatment and comparison groups and this difference was considered as the assumptions were evaluated. Results of the assumption checks are presented in Table 4.

Table 4.

Test Results for the Homogeneity-of-Slopes and Homogeneity-of-Variance Assumption Checks.

	Homogeneity-of-slopes	Homogeneity-of-variance
Year 1 Process	F(1, 75) = 0.03, p = .875	F(1, 77) = 0.96, p = .331
Year 2 Process	F(1, 84) = 9.54, p = .003	F(1, 112) = 0.36, p = .549
Year 1 Concept	F(1, 74) = 4.36, p = .040	F(1, 76) = 5.41, p = .023
Year 2 Concept	F(1, 78) = 1.22, p = .272	F(1, 80) = 1.15, p = .287
Year 1 Content	F(1, 71) = 0.26, p = .615	F(1, 73) = 0.48, p = .489
Year 2 Content	F(1, 78) = 0.09, p = .767	F(1, 80) = 10.46, p = .002

Levene’s test on Year 1 Concept and Year 2 Content revealed concerns with the homogeneity-of-variance assumption. Further inspection revealed the variances were within the 10:1 tolerable aspect ratio of largest to smallest variances suggested by Tabachnick and Fidell (2007). The homogeneity-of-regression slopes assumption did not appear reasonable for the Year 2 Process and the Year 1 Concept; therefore, researchers chose to conduct one-way ANOVAs on the difference scores between the pre- and posttest as suggested by Tabachnick and Fidell. Due to the change in statistical analysis, omega squared was used to determine the effect size, which is described as an estimate of the total variance attributed to the independent variable (Becker, 1999).

Science Process Skills

For the Year 1 data, researchers conducted an ANCOVA comparing the Process posttest scores between the two groups using the pretest scores as a covariate. For Year 2 data, researchers conducted a one-way ANOVA on the difference scores between the pre- and posttests. The means and standard deviations for each group are displayed in Table 5. The results for Year 1, F(1, 76) = 10.40, p = .002, η² = 0.12, and Year 2, F(1, 112) = 11.80, p = .001, ω² = 0.08, both indicated a statistically significant difference between the adjusted posttest scores for the two groups with students in the treatment group scoring higher than students in the comparison group.

Table 5.

Means and Standard Deviations for Gifted Students’ Process Pre- and Posttest Scores.

	Pretest	Posttest	Adjusted posttest M	n	Effect size
	M (SD)	M (SD)	Adjusted posttest M	n	Effect size
Year 1 Treatment	8.53 (3.27)	11.51 (2.21)	11.59	57	0.12^a
Year 1 Comparison	10.32 (2.40)	9.82 (2.91)	9.60	22^b
Year 2 Treatment	7.85 (3.89)	10.24 (2.43)		59	0.08^c
Year 2 Comparison	7.49 (4.36)	8.11 (3.56)		55

Note. Total points possible = 17.

η².

Parents of students in the comparison group were initially reluctant to sign consent form.

ω².

Concept

For the Year 1 data, researchers conducted one-way ANOVAs on the difference scores between the pre- and posttests. For the Year 2 data, an ANCOVA was used to compare the Concept posttest scores between the two groups using the pretest scores as a covariate. The means and standard deviations for each group are displayed in Table 6. The results for Year 1, F(1, 76) = 50.02, p < .001, ω² = 0.39, and Year 2, F(1, 82) = 52.79, p < .001, η² = 0.40, both indicated a statistically significant difference between the adjusted posttest scores for the two groups with students in the treatment group scoring higher than students in the comparison group.

Table 6.

Means and Standard Deviations for Gifted Students’ Concept Pre- and Posttest Scores.

	Pretest	Posttest	Adjusted posttest M	n	Effect size
	M (SD)	M (SD)	Adjusted posttest M	n	Effect size
Year 1 Treatment	10.93 (4.14)	15.09 (3.38)		55	0.39^a
Year 1 Comparison	12.13 (3.28)	8.04 (4.78)		23^b
Year 2 Treatment	11.70 (3.86)	16.14 (3.33)	16.09	44	0.40^c
Year 2 Comparison	9.38 (3.54)	10.61 (3.12)	10.62	38

ω².

Parents of students in the comparison group were initially reluctant to sign consent form. Twenty points possible for Grades 2 to 3, and 35 points possible for Grades 4 to 5.

η².

Content

Researchers conducted two ANCOVA tests comparing the Content posttest scores between the two groups for each year using the pretest scores as covariates. The means and standard deviations for each group are displayed in Table 7. The results for Year 1, F(1, 72) = 31.03, p < .001, η² = 0.30, and Year 2, F(1, 79) = 203.01, p < .001, η² = 0.72, both indicated a statistically significant difference between the adjusted posttest scores for the two groups with students in the treatment group scoring higher than students in the comparison group.

Table 7.

Means and Standard Deviations for Students’ Content Pre- and Posttest Scores.

	Pretest	Posttest	Adjusted posttest M	n	Effect size
	M (SD)	M (SD)	Adjusted posttest M	n	Effect size
Year 1 Treatment	0.80 (1.47)	6.78 (3.14)	6.80	55	0.30^a
Year 1 Comparison	1.45 (1.85)	2.40 (2.56)	2.35	20
Year 2 Treatment	0.77 (1.03)	9.11 (2.94)	9.12	44	0.72^a
Year 2 Comparison	0.79 (1.07)	1.53 (1.61)	1.52	38

Note. Twenty points possible for Grades 2 to 3, and 35 points possible for Grades 4 to 5.

η².

Examples of a low-score response (pretest) and a high-score response (posttest) are displayed in Figures 1 and 2, respectively. These responses are from the same student who was asked to complete a concept map about matter. On the pretest, the student identified three phases of matter and provided examples of each. On the posttest, the student provided greater detail and demonstrated an understanding of more complex concepts. For example, the student indicated that heat causes the phases of matter to change states and also noted that condensation occurs when there is less heat.

Figure 1.

Low-score response: Grade 3 treatment student pretest response to the prompt “tell me everything you know about matter.”

Figure 2.

High-score response: Grade 3 treatment student posttest response to the prompt “tell me everything you know about matter.”

Discussion

The results from the Process test indicated gifted students in the treatment classrooms scored significantly higher on posttest scores than their gifted counterparts in both Year 1 and Year 2. As indicated by the effect sizes (η² = 0.12, ω² = 0.08), student participation in the intervention accounted for approximately 12% of the variability in student posttest scores on the Process assessment in Year 1 and 8% in Year 2. According to the results, gifted students in the treatment classroom were better able to design an experiment when presented with a testable question than gifted students in the comparison classrooms. These results are notable because the development of science process skills allows students to better model and utilize authentic scientific practices (Duschl et al., 2007) as encouraged by the Next Generation Science Standards (NGSS Lead States, 2013). This finding is consistent with results found from an earlier study on the effects of the William and Mary science curriculum units on students’ science process skills. VanTassel-Baska et al. (1998) also found small but significant gains in students’ science process skills when compared with similar students not exposed to the units.

The results from the Concept test revealed treatment students scored significantly higher on posttest scores than the comparison students in both Year 1 and Year 2.The effect sizes (ω² = 0.39, η² = 0.40) indicated that participation in the intervention accounted for approximately 39% of the variability in student posttest scores on the Concept assessment in Year 1 and 40% in Year 2. These results were expected due to the increased emphasis on overarching concepts both in the curriculum units and in the teacher professional development. Even though the overarching concepts (systems and change) are commonly highlighted in science curricula, many comparison teachers may not have used the concepts to connect the content as encouraged in the treatment classrooms. These results indicate that student understanding of concepts is substantially aided when teachers explicitly articulate overarching concepts and link content to them.

The results from the Content test revealed treatment students scored significantly higher on posttest scores than the comparison students in both Year 1 and Year 2. The effect sizes (η² = 0.30, η² = 0.72) indicated the student participation in the intervention accounted for approximately 30% of the variability in student posttest scores on the Content assessment in Year 1 and 72% in Year 2. The curriculum units used by the treatment classrooms provided multiple opportunities for students to explore the particular science topics in greater depth than those used by the comparison classroom (VanTassel-Baska et al., 1998). Students in the treatment classrooms were involved in a 9-week study on their particular science unit. For example, students in Grade 3 spent a full 9 weeks investigating matter, and it was subsequently referenced throughout the year. In contrast, the comparison classrooms typically read about matter in their textbooks and did not engage students in deep exploration of the content.

Limitations

Limitations of this study included the small sample size of gifted students across the five schools. Student attrition from the schools limited the number of 2-year participants. Attrition did not appear to be related to participation in the intervention but rather to the student mobility common in low-income districts. Another limitation of the study was the use of curriculum unit tests rather than an external achievement test to compare content and concept achievement between the experimental and comparison groups. Although an external achievement measure, the state science content test, was considered, the state test was not yet piloted nor revised during the first year of the study and was designed for Grade 5 only. Finally, both groups had familiarity with concept mapping as an instructional tool and were instructed in its use as part of the pre- and posttest data collection, but the experimental group might have used this type of graphic organizer more frequently in their science classrooms.

Conclusion and Scholarly Significance

The results from this study supported the implementation of a rigorous differentiated science curriculum focused on improving science concept and content knowledge, and process skills. Gifted education students in the treatment classrooms demonstrated statistically significant and, in several cases, practically significant gains in science process skills, science concepts, and science content knowledge when compared with gifted students in the comparison classrooms. Results indicated that gifted students in the treatment classrooms were better able to design science experiments when presented with a real-world problem, make scientific connections using overarching concepts such as change and systems, and benefit from being allowed to fully explore the age appropriate content in an investigatory manner as recommended by the NRC (Duschl et al., 2007) and the NSB (2010).

The results of this real-world field study are noteworthy, given the introduction and implementation of the Next Generation Science Standards (NGSS). The three dimensions that comprise the NGSS (Science and Engineering Practices, Crosscutting Concepts, and Disciplinary Core Ideas) were accentuated in the STEM Starters intervention and provided an important foundation for developing the STEM talent of young gifted learners. Through the emerging NGSS dimensions, students are encouraged to actively engage in inquiry-based, problem-centered experiences (NGSS Lead States, 2013) emphasized as a key feature of STEM Starters. In summary, to provide students with these experiences and to develop the STEM talent of young gifted learners, students need inquiry-based, problem-centered experiences (S. P. Marshall et al., 2011) and their teachers need professional development to identify and cultivate students’ talents (NSB, 2010).

Footnotes

Appendix

Grade 2: Arkansas State Science Frameworks (Arkansas Department of Education [ADE], 2003)
NS.1.2.1 Communicate observations orally, in writing, and in graphic organizers NS.1.2.2 Develop questions that guide scientific inquiry NS.1.2.3 Conduct scientific investigations individually and in teams NS.1.2.4 Estimate and measure length and temperature using International System of Units (SI) NS.1.2.5 Collect measurable empirical evidence in teams and as individuals NS.1.2.6 Make predictions in teams and as individuals based upon empirical evidence NS.1.2.7 Use age appropriate equipment and tools in scientific investigations (e.g., balances, hand lenses, rulers, and thermometers) NS.1.2.8 Apply lab safety rules as they relate to specific science lab activities LS.2.2.3 Identify basic needs of most plants LS.2.2.4 Compare different types of flowering plants and conifers LS.2.2.6 Describe the function of the following plant parts: leaves, stems, flowers, roots PS.5.2.2 Investigate the effect of physical phenomena on various materials (e.g., heat absorption by different colored materials) PS.7.2.2 Compare temperatures using the Celsius scale ESS.8.2.1 Conduct investigations to distinguish among the following components of soil: clay, sand, silt, and humus ESS.8.2.2 Recognize and discuss the different properties of soil: color, texture, ability to retain water, and ability to support plant growth ESS.8.2.3 Conduct investigations to determine which soil best supports bean plant growth ESS.8.2.5 Chart weather conditions every day ESS.8.2.6 Demonstrate safety procedures related to severe weather ESS.8.2.7 Describe characteristics of cumulus, stratus, and cirrus clouds ESS.8.2.8 Predict weather based on cloud type ESS.8.2.9 Read a Celsius thermometer
Grade 3: Arkansas State Science Frameworks (Arkansas Department of Education [ADE], 2003)
NS.1.3.1 Communicate observations orally, in writing, and in graphic organizers NS.1.3.2 Develop questions that guide scientific inquiry NS.1.3.3 Conduct scientific investigations individually and in teams NS.1.3.4 Communicate the results of scientific investigations NS.1.3.5 Estimate and measure length and temperature using International System of Units (SI) NS.1.3.6 Collect and analyze measurable empirical evidence in teams and as individuals NS.1.3.7 Make and explain predictions in teams and as individuals based upon empirical evidence NS.1.3.8 Use simple equipment and age appropriate tools, technology, and mathematics in scientific investigations (e.g., balances, hand lenses, rulers, and thermometers) NS.1.3.9 Apply rules as they relate to specific science lab activities PS.5.3.1 Compare and contrast objects based on two or more properties PS.5.3.2 Demonstrate physical changes in matter PS.5.3.3 Demonstrate the mass of solids PS.5.3.4 Compare and contrast solids and liquids ESS.8.3.1 Distinguish among Earth’s materials. ESS.9.3.1 Analyze the effect of wind and water on Earth’s surface
Grade 4: Arkansas State Science Frameworks (Arkansas Department of Education [ADE], 2003)
NS.1.4.1 Communicate observations orally, in writing, and in graphic organizers NS.1.4.2 Refine questions that guide scientific inquiry NS.1.4.4 Conduct scientific investigations individually and in teams NS.1.4.5 Communicate the results of scientific investigations NS.1.4.6 Estimate and measure length, mass, temperature, capacity/volume, and elapsed time using International System of Units (SI) NS.1.4.7 Collect and interpret measurable empirical evidence in teams and as individuals NS.1.4.8 Develop a hypothesis based on prior knowledge and observations NS.1.4.9 Identify variables that affect investigations NS.1.4.10 Identify patterns and trends in data NS.1.4.11 Generate conclusions based on evidence NS.1.4.12 Evaluate the quality and feasibility of an idea or project NS.1.4.13 Use simple equipment, age appropriate tools, technology, and mathematics in scientific investigations NS.1.4.14 Apply lab safety rules as they relate to specific science lab activities PS.6.4.1 Investigate the relationship between force and direction PS.6.4.2 Investigate the relationship between force and mass PS.7.4.2 Classify electrical conductors and insulators PS.7.4.3 Construct simple circuits from circuit diagrams
Grade 5: Arkansas State Science Frameworks (Arkansas Department of Education [ADE], 2003)
NS.1.5.1 Make accurate observations NS.1.5.2 Identify and define components of experimental design used to produce empirical evidence NS.1.5.3 Calculate mean, median, mode, and range from scientific data using International System of Units (SI) units NS.1.5.4 Interpret scientific data using data tables/charts, bar graphs, circle graphs, line graph’s, stem and leaf plots, Venn diagrams NS.1.5.5 Communicate results and conclusions from scientific inquiry NS.1.5.6 Develop and implement strategies for long-term, accurate data collection NS.1.5.7 Summarize the characteristics of science NS.1.5.8 Explain the role of observation in the development of a theory NS.1.5.9 Define and give examples of hypotheses LS.4.5.1 Distinguish among and model: organisms, populations, communities, ecosystems, and biosphere LS.4.5.3 Design food webs in specific habitats to show the flow of energy within communities LS.4.5.4 Evaluate food webs under conditions of stress LS.4.5.12 Conduct investigations in which plants are encouraged to thrive LS.4.5.16 Evaluate positive and negative human effects on ecosystems LS.4.5.18 Investigate careers, scientists, and historical breakthroughs related to populations and ecosystems PS.5.5.1 Identify the relationship of atoms to all matter PS.5.5.2 Conduct scientific investigations on physical properties of objects PS.5.5.3 Identify common examples of physical properties PS.5.5.4 State characteristics of physical changes PS.5.5.5 Identify characteristics and common examples of physical changes PS.5.5.7 Demonstrate the effect of changes in the physical properties of matter

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research and/or authorship of this article: This work was funded through the U.S. Department of Education, Jacob K. Javits Gifted and Talented and Students Program under Award Number S206A080026.

References

Adams

Callahan

C. M.

(1995). The reliability and validity of a performance task for evaluating science process skills. Gifted Child Quarterly, 39, 14-20. doi:10.1177/001698629503900103

Akinoglu

Tandogan

R. O.

(2007). The effects of problem-based active learning in science education on students’ academic achievement, attitude, and concept learning. Eurasia Journal of Mathematics, Science & Technology Education, 3, 71-81.

Appleton

(2008). Developing science pedagogical content knowledge through mentoring elementary teachers. Journal of Science Teacher Education, 19, 523-545. doi:10.1007/s10972-008-9109-4

Arkansas Department of Education. (2003). English language arts curriculum framework. Retrieved from http://www.arkansased.org/public/userfiles/Learning_Services/Curriculum%20and%20Instruction/Frameworks/Eng%20Lang%20Arts/ela_K_8_2003.pdf

Becker

L. A.

(1999). Measures of effect size (strength of association). Retrieved from http://www.uccs.edu/lbecker/glm_effectsize.html

Bickel

(2007). Multilevel analysis for applied research: It’s just regression. New York, NY: Guilford Press.

Brandwein

P. F.

(1995). Science talent in the young expressed within ecologies of achievement (RBDM 9510). Storrs, CT: The National Research Center on the Gifted and Talented, University of Connecticut.

Buczynski

Hansen

C. B.

(2010). Impact of professional development on teacher practice: Uncovering connections. Teacher and Teacher Education, 26, 599-607. doi:10.1016/j.tate.2009.09.006

Callahan

Hunsaker

Adams

Moore

Bland

(1995). Instruments used in the identification of gifted and talented students [Research Monograph 95130]. Storrs, CT: The National Research Center on the Gifted and Talented.

10.

Cervetti

G. N.

Barber

Dorph

Pearson

P. D.

Goldschmidt

P. G.

(2012). The impact of an integrated approach to science and literacy in elementary school classrooms. Journal of Research in Science Teaching, 49, 631-658. doi:10.1002/tea.21015

11.

Common Core State Standards. (2012). English language arts: Literacy-writing. Retrieved from http://www.corestandards.org/ELA-Literacy/W/3

12.

Corcoran

McVay

Riordan

(2003). Getting it right: The MISE approach to professional development. Philadelphia, PA: Consortium for Policy Research in Education.

13.

Cotabish

Dailey

Hughes

Robinson

(2011). The effects of a STEM professional development intervention on elementary teachers’ science process skills. Research in the Schools, 18(2), 16-25.

14.

Cotabish

Dailey

Robinson

Hughes

(2013). The effects of a STEM intervention on elementary students’ science knowledge and skills. School Science and Mathematics, 113, 215-226.

15.

Darling-Hammond

Wei

R. C.

Andree

Richardson

Orphanos

(2009). Professional learning in the learning profession: A status report on teacher development in the United States and abroad. Dallas, TX: National Staff Development Council.

16.

Desimone

L. M.

Porter

A. C.

Garet

M. S.

Yoon

K. S.

Birman

B. F.

(2002). Effects of professional development on teachers’ instruction: Results from a three-year longitudinal study. Educational Evaluation and Policy Analysis, 24(2), 81-112.

17.

Drake

K. N.

Long

(2009). Rebecca’s in the dark: A comparative study of problem-based learning and direct instruction/experimental learning in two 4th-grade classrooms. Journal of Elementary Science Education, 21, 1-16.

18.

Duschl

Schweingruber

H. A.

Shouse

(2007). Taking science to school: Learning and teaching science in grades K-8. Washington, DC: The National Academies Press.

19.

Feng

A. Z.

VanTassel-Baska

Quek

Bai

Oneill

(2005). A longitudinal assessment of gifted students’ learning using the integrated curriculum model (ICM): Impacts and perceptions of the William and Mary language arts and science curriculum. Roeper Review, 27(2), 78-83.

20.

Fowler

(1990). The diet cola test. Science Scope, 13(4), 32-34.

21.

Gallagher

S. A.

Stepien

(1996). Content acquisition in problem-based learning: Depth versus breadth in American studies. Journal for the Education of the Gifted, 19, 257-275.

22.

Gallagher

S. A.

Stepien

W. J.

Rosenthal

(1992). The effects of problem-based learning on problem solving. Gifted Child Quarterly, 36, 195-200.

23.

Gerard

L. F.

Varma

Corliss

S. B.

Linn

M. C.

(2011). Professional development for technology-enhanced inquiry science. Review of Educational Research, 81, 408-448. doi:10.3102/0034654311415121

24.

Johnson

C. C.

Fargo

J. D.

(2010). Urban school reform enabled by transformative professional development: Impact on teacher change and student learning of science. Urban Education, 45, 4-29. doi:10.1177/0042085909352073

25.

Johnson

C. C.

Kahle

J. B.

Fargo

J. D.

(2007). A study of the effect of sustained, whole-school professional development on student achievement in science. Journal of Research in Science Teaching, 44, 775-786. doi:10.1002/tea.20149

26.

Keeley

(2009, April 22). Elementary science education in the K-12 system. NSTA WebNews Digest. Retrieved from http://www.nsta.org/publications/news/story.aspx?id=55954

27.

Kim

K. H

VanTassel-Baska

Bracken

B. A.

Feng

Stambaugh

Bland

(2012). Project Clarion: Three years of science instruction in Title 1 schools among K-Third grade students. Research in Science Education, 42, 813-829. doi:10.1007/s11165-011-9218-5

28.

Lindahl

(2007, April). A longitudinal study of students’ attitudes towards science and choice of career. Paper presented at the National Association for Research in Science Teaching, New Orleans, LA.

29.

Little

C. A.

Paul

K. A.

(2011). Professional development to support implementation of content-based curriculum. In VanTassel-Baska

Little

C. A.

(Eds.), Content-based curriculum for high-ability learners (2nd ed., pp. 413-436). Waco, TX: Prufrock Press.

30.

Little

P. F. B.

(2005). Peer coaching as a support to collaborative teaching. Mentoring & Tutoring, 13, 83-94.

31.

Lumpe

Czerniak

Haney

Beltyukova

(2012). Beliefs about teaching science: The relationship between elementary teachers’ participation in professional development and student achievement. International Journal of Science Education, 34, 153-166. doi:10.1080/09500693.2010.551222

32.

Maltese

A. V.

Tai

R. H.

(2010). Eyeballs in the fridge: Sources of early interest in science. International Journal of Science Education, 32, 669-685. doi:10.1080/09500690902792385

33.

Marshall

J. C.

Horton

R. M.

(2011). The relationship of teacher-facilitated, inquiry-based instruction to student higher-order thinking. School Science and Mathematics, 111(3), 93-101. doi:10.1111/j.1949-8594.2010.00066.x

34.

Marshall

S. P.

McGee

G. W.

McLaren

Veal

C. C.

(2011). Discovering and developing diverse STEM talent: Enabling academically talented urban youth to flourish. Gifted Child Today, 34(10), 16-23.

35.

Metz

K. E.

(2008). Narrowing the gulf between the practices of science and the elementary school science classroom. The Elementary School Journal, 109, 138-161.

36.

Moyer-Packenham

P. S.

Bolyard

J. J.

Cerar

N. I.

(2011). Common features of professional development activities for mathematics and science teachers. Professional Development in Education, 37, 571-589. doi:10.1080/19415257.2010.531597

37.

National Association for Gifted Children, Task Force on Math and Science. (2008). The STEM promise: Recognizing and developing talent and expanding opportunities for promising students of science, technology, engineering and mathematics. Retrieved from http://www.nagc.org/uploadedFiles/STEM%20white%20paper%281%29.pdf

38.

National Science Board. (2010). Preparing the next generation of STEM innovators: Identifying and developing our nation’s human capital (NSB-10-33). Retrieved from http://www.nsf.gov/nsb/publications/2010/nsb1033

39.

NGSS Lead States. (2013). Next generation science standards: For states, by states. Washington, DC: Achieve. Retrieved from http://www.nextgenscience.org/next-generation-science-standards

40.

O’Connell

A. A.

McCoach

D. B.

(2008). Multilevel modeling of educational data. Charlotte, NC: Information Age Publishing.

41.

Penuel

W. R.

Gallagher

L. P.

Moorthy

(2011). Preparing teachers to design sequences of instruction in earth systems science: A comparison of three professional development programs. American Educational Research Journal, 48, 996-1025. doi:10.3102/0002831211410864

42.

Robbins

J. I.

(2011). Adapting science curricula for high-ability learners. In VanTassel-Baska

Little

C. A.

(Eds.), Content-based curriculum for high-ability learners (2nd ed., pp. 437-465). Waco, TX: Prufrock Press.

43.

Roberts

J. L.

(2010). Talent development in STEM disciplines: Diversity—Cast a wide net. National Consortium for Specialized Secondary Schools of Mathematics, Science, and Technology, 16(1), 7-9. Retrieved from http://www.ncsssmst.org/publications/journal/63-spring-2010-journal

44.

Robinson

(2006). Blueprints for biography: Differentiating the curriculum for talented readers. Teaching for High Potential, 7-8.

45.

Robinson

Shore

B. M.

Enersen

D. L.

(2007). Best practices in gifted education: An evidence-based guide. Waco, TX: Prufrock Press.

46.

Roehrig

G. H.

Dubosarsky

Mason

Carlson

Murphy

(2011). We look more, listen more, notice more: Impact of sustained professional development on Head Start teachers’ inquiry-based and culturally-relevant science teaching practices. Journal of Science Education and Technology, 20, 566-578.

47.

Romance

N. R.

Vitale

M. R.

(2011). A research-based instructional model for integrating meaningful learning in elementary science and reading comprehension: Implications for policy and practice. In Stein

N. L.

Raudenbush

S. W.

(Eds.), Developmental cognitive science goes to school (pp. 127-142). New York, NY: Routledge.

48.

Santau

A. O.

Maerten-Rivera

J. L.

Huggins

A. C.

(2011). Science achievement of English language learners in urban elementary schools: Fourth-grade student achievement results from a professional development intervention. Science Education, 95, 771-793. doi:10.1002/sce.20443

49.

Schleigh

S. P.

Bosse

M. J.

Lee

(2011). Redefining curriculum integration and professional development: In-service teachers as agents of change. Current Issues in Education, 14(3), 1-12.

50.

Showers

(1982). Transfer of training: The contribution of coaching (Report No. ED 231 035). Eugene: Center for Educational Policy and Management, College of Education, University of Oregon.

51.

Showers

(1984). Peer coaching: A strategy for facilitating transfer of training (Report No. ED 271 849). Eugene: Center for Educational Policy and Management, College of Education, University of Oregon.

52.

Showers

Joyce

(1996). The evolution of peer coaching. Educational Leadership, 53(6), 12-16.

53.

Shymansky

J. A.

Wang

T. L.

Annetta

L. A.

Yore

L. D.

Everett

S. A.

(2010). How much professional development is needed to effect positive gains in K-6 student achievement on high stakes science tests? International Journal of Science and Mathematics Education, 10(1), 1-19.

54.

Simon

Campbell

Johnson

Stylianidou

(2011). Characteristics of effective professional development for early career science teachers. Research in Science & Technological Education, 29, 5-23. doi:10.1080/02635143.2011.543798

55.

Supovitz

J. A.

Turner

H. M.

(2000). The effects of professional development on science teaching practices and classroom culture. Journal of Research in Science Teaching, 37, 963-980. doi:10.1002/1098-2736(200011)37:9<963::AID-TEA6>3.0.CO;2-0

56.

Tabachnick

B. G.

Fidell

L. S.

(2007). Using multivariate statistics (5th ed.). New York, NY: HarperCollins.

57.

VanTassel-Baska

(1998). Planning science programs for high ability learners. ERIC Clearinghouse on Disabilities and Gifted Education. Retrieved from http://www.ericdigests.org/1999-3/science.htm

58.

VanTassel-Baska

Bass

Ries

Poland

Avery

(1998). A national study of science curriculum effectiveness with high ability students. Gifted Child Quarterly, 42, 200-211.

59.

VanTassel-Baska

Stambaugh

(Eds.). (2008). What works: Twenty years of curriculum development and research for advanced learners. Retrieved from http://education.wm.edu/centers/cfge/curriculum/documents/WhatWorks.pdf

60.

Yoon

K. S.

Duncan

Lee

Shapley

(2008, March). The effects of teachers’ professional development on student achievement: Findings from a systemic review of evidence. Paper presented at the annual meeting of the American Educational Research Association, New York, NY.