Impact of STEM-Integrated Projects on Urban High School Science Engagement

Thesis Abstract

This study investigates how STEM-integrated projects influence science engagement among students in urban high schools, addressing persistent disparities in science participation and achievement linked to limited access to authentic, interdisciplinary learning experiences. The aim is to determine whether integrating STEM projects into standard science curricula enhances behavioral, affective, and cognitive engagement, and to identify contextual factors that mediate or moderate these effects. Specific objectives include (1) assessing changes in student behavioral engagement (participation, persistence, attendance) after a full academic year of project-based STEM instruction; (2) examining shifts in affective engagement (interest, value, self-efficacy) and intrinsic motivation; (3) evaluating cognitive engagement through evidence of persistence, elaboration, and self-regulated learning during project tasks; (4) exploring teachers’ instructional practices, time allocation, and perceived feasibility of STEM integration; and (5) testing whether classroom environment, community partnerships, and student sociodemographic variables moderate engagement outcomes. A mixed-methods design is employed, combining a quasi-experimental, pretest–posttest control group with concurrent qualitative inquiry. The population comprises 12 urban high schools within a metropolitan district, with a total target sample of 720 students in Grade 9 and Grade 10, drawn from six schools in the STEM-integrated cohort and six matched comparison schools. A stratified cluster sampling approach ensures representation across gender, ethnicity, and English language learner status. Data collection instruments include (a) the National Survey of Student Engagement adapted for science classrooms to quantify behavioral and affective engagement; (b) the Science Learning and Engagement Instrument to assess cognitive engagement; (c) classroom observation protocols using the STEM-Integrated Instructional Practices Framework; (d) teacher surveys on pedagogical practices, time-on-task, and collaboration with industry or higher education partners; and (e) semi-structured interviews with a purposive sample of 24 students and 12 teachers to capture nuanced experiences. Validity and reliability procedures involve pilot testing, factor analysis for survey scales, inter-rater reliability for observations (Cohen’s kappa ? .70), and triangulation across methods. Data analysis proceeds in two strands. Quantitative data will be analyzed using ANCOVA to compare post-intervention engagement scores between STEM-integrated and control groups while controlling for baseline differences and covariates such as prior achievement and socioeconomic status. Multigroup structural equation modeling will test a theoretical model linking STEM project features (authenticity, collaboration, interdisciplinarity) to engagement outcomes, with mediating roles for self-efficacy and perceived relevance. Hierarchical linear modeling will account for nesting of students within classes and schools. Qualitative data will undergo thematic analysis following an explicit coding framework aligned with the theoretical constructs of self-determination theory and expectancy-value theory, with member-checking to ensure credibility. A cross-method synthesis will integrate quantitative trends with qualitative narratives to interpret differential engagement effects and contextual contingencies. Expected findings anticipate that students in STEM-integrated classrooms will show statistically significant increases in behavioral and affective engagement, with moderate improvements in cognitive engagement relative to controls. The magnitude of effects is expected to be larger for students from underrepresented groups, historically marginalized in science, particularly when projects incorporate community partnerships and real-world problem-solving. The study also anticipates identifying critical instructional practices—such as structured collaboration norms, inquiry-based questioning, and explicit connections to career pathways—that amplify engagement. The contribution to knowledge lies in providing robust empirical evidence on the effectiveness of STEM-integrated projects in promoting science engagement in urban settings, elucidating mechanisms through which project features drive engagement, and offering scalable, context-sensitive guidelines for policymakers and practitioners. Based on findings, the study will advance theory by integrating constructs from self-determination theory and expectancy-value theory within an urban STEM education context and will inform practice by detailing scalable implementation strategies, assessment approaches, and professional development needs. Recommendations include adopting district-wide STEM project frameworks with built-in community partnerships, professional development focusing on collaborative pedagogy and data-informed instruction, and targeted supports for students with lower prior achievement or English language proficiency to sustain engagement gains. Limitations include potential variability in project quality across schools and the influence of external community factors; future research should explore long-term retention of engagement gains and translational effects on STEM course enrollment.