Sixth-Grade Study Finds STEM Interest Does Not Guarantee Three-Dimensional Science Knowledge

A six-week classroom study of sixth-grade students in an integrated science and engineering program indicates that student interest and three-dimensional knowledge do not share a straightforward correlation. Researchers evaluating quantitative test data and qualitative student artifacts found that high self-reported interest does not guarantee strong conceptual understanding, nor does low interest prevent students from achieving top assessment scores. The mixed-methods investigation followed seventy-three sixth-graders completing the Iterative Science and Engineering curriculum, an instructional framework designed around local environmental problems. Rather than finding a direct relationship where elevated interest mirrors higher test scores, the authors identified six multi-faceted profiles showing significant divergence between student enthusiasm and academic mastery, challenging assumptions that boosting classroom interest automatically produces deeper learning gains.

The investigation addresses international concerns regarding student preparation in science, technology, engineering, and mathematics disciplines. Worldwide, educational agencies, corporations, and non-profits emphasize pre-university STEM instruction that prepares citizens to resolve ecological and health crises, including air pollution, extreme weather, and habitat destruction. Nevertheless, longitudinal data from the Programme for International Student Assessment demonstrate persistent academic stagnation. Science scores for United States eighth-graders showed no statistically significant change from 1995 to 2022 according to national and international education agencies. Furthermore, the score gap between high- and low-achieving students widened between 2018 and 2022 following pandemic school disruptions. Although national frameworks call for integrated three-dimensional science learning, conventional programs continue prioritizing passive memorization over active, iterative problem-solving and critical thinking.

To examine how conceptual understanding connects with personal interest, researchers Julia E. Calabrese, Nancy Butler Songer, Daniel Kalani Aina Jr., and Holly Cordner studied a public secondary school in the western United States. The research took place within a broader project comprising one hundred and thirty-three participants across six sixth-grade classrooms and one eighth-grade classroom. To maintain measurement consistency, the authors selected a purposeful sample of seventy-three sixth-grade students who completed all four assessment phases. Within the district, demographic records indicated that eighty-one percent of students were White, fourteen percent Hispanic, three percent of multiple races, and under one percent each Black, Asian, American Indian, or Pacific Islander. Additionally, five percent were English Language Learners, fifteen percent received disability services, and twenty-five percent qualified for meal assistance.

The classroom intervention utilized the Iterative Science and Engineering instructional model, implemented across three units spanning six weeks in regular fifty-minute science periods. The model systematically alternated between scientific investigation and engineering design to mirror authentic scientific work. Framing learning around an authentic task from the Department of Agriculture, students investigated local invasive insect populations. Participants collected outdoor neighborhood observations, analyzed biological patterns, and constructed their first formal scientific arguments. Student teams then researched a specific invasive insect, composed scientific arguments defending why it posed ecological harm, and designed physical insect traps using everyday materials. Students tested prototypes, incorporated peer critiques, and engineered a second build featuring enhanced trapping selectivity and efficiency before presenting their final solutions using infographics and posters.

Researchers measured learning and attitudes using identical tests administered before and after the curriculum. The three-dimensional knowledge examination contained ten questions totaling eighteen points, aligning with disciplinary core ideas, science and engineering practices, and crosscutting concepts. Items included multiple-choice questions assessing species richness, trophic levels, and carnivore classifications, alongside fill-in-the-blank questions on predator-prey dynamics. Constructed-response questions required students to construct scientific arguments with claims, evidence, and reasoning, incorporating one scaffolded item and two unscaffolded items. Students also modeled energy flow across five species and suggested design refinements for an insect trap. Concurrently, students completed a four-item survey adapted from the STEM Semantics Survey, evaluating their interest in and perceived meaningfulness of science and engineering on five-point Likert scales ranging from boring to interesting.

Using an explanatory mixed-methods design, researchers established quantitative benchmarks to group participants into descriptive profiles. For the knowledge assessment, researchers calculated means and standard deviations, defining high scores as at least one standard deviation above the mean, low scores as one standard deviation below the mean or less, and average scores between those cutoffs. On the pretest, high scores were nine or higher and low scores were two or lower; on the posttest, high scores were fourteen or higher and low scores were six or lower. Summing these cutoffs established composite knowledge thresholds of twenty-three for high and eight for low. For the interest survey, scores of sixteen or greater were classified as high and eight or less as low, yielding composite thresholds of thirty-two for high and sixteen for low.

Cross-referencing composite metrics produced six participant profiles: High Interest with High Knowledge, High Interest with Average Knowledge, High Interest with Low Knowledge, Average Interest with High Knowledge, Average Interest with Average Knowledge, and Average Interest with Low Knowledge. Notably, zero participants fell into a composite low-interest category, with all students reporting average or high interest overall. Knowledge scores were more divided, with over half the cohort scoring in the average range on both tests. Furthermore, only ten out of seventy-three students showed matching classifications across all four testing instruments, while nine students recorded scores spanning high, average, and low tiers across different measures. To unpack these variations, researchers selected four extreme cases from the least represented profiles for qualitative investigation through student science notebooks and teacher interviews.

The first extreme case examined was Melody, who represented the High Interest, High Knowledge profile. Melody was the only student in the sample to award the highest rating of five to every survey item both before and after the intervention. Her knowledge test scores ranked in the top ten percent on the pretest, and she increased her scores on both argument and non-argument questions on the posttest, missing only three total non-argument points across both administrations. Melody’s science notebook showed substantial effort, containing an elaborate trap drawing featuring five distinct viewing perspectives, detailed labels, and assembly instructions. Nevertheless, her teacher noted that despite being bright and active in extracurricular activities, Melody struggled to supply sufficient evidence and reasoning in unscaffolded notebook arguments, demonstrating that high interest does not remove all learning hurdles.

The second case focused on Beau, who belonged to the High Interest, Low Knowledge profile consisting of four participants. Beau maintained perfect ratings of five on engineering interest across both surveys, though his science ratings shifted slightly between curiosity and perceived meaningfulness. In stark contrast to his high personal enthusiasm, Beau scored only two points on the knowledge pretest and six on the posttest, earning merely one point in the argumentation section. His science notebook showed minimal depth, containing a brief claim without supporting evidence or reasoning. In interviews, his teacher described Beau as very smart but noted that he prioritized peer approval, was easily distracted by friends, and frequently missed instruction for traveling sports competitions. Beau’s profile demonstrates that high personal interest in engineering activities does not prevent students from struggling with rigorous academic expectations.

The third case study evaluated Camila, who exemplified the Average Interest, High Knowledge profile. Camila recorded the lowest composite interest score among all seventy-three students, dropping from a nine on the presurvey to an eight on the post-survey, with her engineering interest dropping to the lowest rating. Despite her stated disinterest, Camila excelled academically, posting above-average pretest scores and achieving the second-highest score on the posttest by missing points on only two questions. Her notebook demonstrated strong organizational skill, featuring custom hand-drawn scaffolding boxes separating claims, evidence, and reasoning, alongside detailed trap drawings showing smashed plums used as bait. Her teacher revealed that Camila possessed strong extrinsic motivation to outperform peers and attain top marks. Camila’s case illustrates that personal interest is not mandatory for academic success when disciplined work habits and extrinsic motivation drive learning.

The final case examined Simone, who represented the Average Interest, Low Knowledge profile. Simone was the only student in the sample classified as low on three of the four assessment administrations, scoring an eight on her initial interest survey and earning just two points on the knowledge pretest. Although she achieved modest gains on the posttest, her score remained below the group average. Simone’s science notebook was largely incomplete; while her invasive species prediction graph included four colored lines, her written explanation addressed only two organisms, and her scientific argument asserted that insects invade people without providing explanatory evidence. Simone’s teacher reported that she experienced low academic self-confidence and chose to submit minimal, rapid work to manage classroom stress and avoid taking school assignments home. Simone’s situation reveals that low confidence and task anxiety can constrain achievement independently of interest.

The researchers noted several study limitations, including the shortened four-item survey instrument adapted from the ten-item STEM Semantics Survey, the use of composite benchmark thresholds, and the restriction of analysis to students completing all four assessments. Additionally, the authors acknowledged that open-ended tasks capture only selected dimensions of three-dimensional competencies. Concluding their study, Calabrese and colleagues emphasized that student learning in science and engineering is far more complex than aligning lessons with personal likes and dislikes. Because high interest does not guarantee strong conceptual growth and low interest does not preclude academic excellence, instructional programs cannot rely solely on engagement strategies to drive achievement. Instead, the authors recommend that educators focus instructional time on core competencies, including iterative problem-solving, critical thinking, and evidence-based argumentation, providing students with structured opportunities to master complex three-dimensional science practices.

Subject of Research: Relationship between three-dimensional science and engineering knowledge and student interest

Article Title: Re-examining the relationship between three-dimensional science and engineering knowledge and student interest: a mixed methods study

Article References: Calabrese, J. E., Songer, N. B., Aina, D. K., & Cordner, H. (2025). Re-examining the relationship between three-dimensional science and engineering knowledge and student interest: a mixed methods study. Disciplinary and Interdisciplinary Science Education Research, 7(1), Article 17. https://doi.org/10.1186/s43031-025-00134-y

Image Credits: AI Generated

DOI: 10.1186/s43031-025-00134-y

Keywords: three-dimensional learning, science education, engineering design, student interest, STEM education, Iterative Science and Engineering, scientific argumentation, middle school science, Next Generation Science Standards, conceptual understanding, mixed methods research, invasive species curriculum, STEM Semantics Survey, classroom artifacts, PISA science assessments, disciplinary core ideas, science and engineering practices, crosscutting concepts