Persistent excellence gaps document disparities in advanced achievement, particularly in mathematics. The study examined student- and school-level correlates of mathematics achievement among adolescents scoring at or above the dataset-specific 90th percentile in mathematics by employing a comparative analysis of two large-scale datasets: the international Programme for International Student Assessment (PISA) 2022 and the nationally representative High School Longitudinal Study of 2009 (HSLS:09). Using multilevel modeling (PISA) and hierarchical regression (HSLS:09), the study estimated the associations of student demographics (sex, language background, socioeconomic status) and context-specific school administrative structures. Results revealed socioeconomic status as a robust and cross-contextual correlate of achievement. However, the salience of other factors was context-dependent: demographic predictors varied between datasets and school-level administrative factors (e.g. ability grouping, extracurricular offerings) showed no significant association in the PISA 2022 model. In stark contrast, within the U.S.-specific HSLS:09 context, the institutional availability of a robust advanced curriculum such as offering multiple Advanced Placement or International Baccalaureate mathematics courses was significantly and positively associated with achievement. The findings demonstrate that while socioeconomic inequality universally shapes elite performance, the measurable impact of school structures is contingent on assessment context and granularity. Implications, limitations, and future directions of the present study are discussed.
This study investigates the impact of artificial intelligence (AI)-supported learning environments enriched with environmentally friendly scenarios on the cognitive and problem-solving skills of gifted students. As part of the study, students developed an "Environmentally Friendly Assistant" using Machine Learning for Kids (MLK), incorporating text classification, sentiment analysis, and chatbot applications focused on recycling, energy saving, and nature protection. The research was conducted with 41 gifted middle school students through a single-group pretest - posttest experimental design over four weeks. Data were collected through an achievement test and a rubric measuring cognitive skills, problem-solving, and environmental awareness. The findings indicated significant improvements in scores from pretest to posttest within the study sample, particularly in algorithmic reasoning, data selection, and error analysis. Rubric evaluations similarly showed notable increases in higher-order cognitive and problem-solving skills. Furthermore, higher levels of awareness regarding environmentally friendly behaviors were observed following the intervention, along with their ability to adapt AI applications to social and environmental issues. The AI-based instructional process provided a holistic learning experience by enhancing gifted students' academic skills and environmental awareness. These findings highlight the potential of AI integration to support sustainable education and foster 21st-century skills.
Our study examined how mentoring success in promoting girls in STEM depends on mentees' environmental and motivational prerequisites. Following a systemic approach, we analyzed how participation in a one-year online mentoring program in STEM was associated with growth in endogenous STEM learning resources (learning capital) among 330 talented girls (M age = 13.48), and what role (a) initial resources for STEM learning in their personal environment (educational capital), and (b) their motivation to engage with STEM played in this growth. On average, participants' learning capital slightly improved during mentoring. We found a compensatory effect of mentoring: girls who began mentoring with low educational capital demonstrated larger increases in learning capital. This effect was particularly pronounced among participants with a comparatively low motivation to engage with STEM, suggesting that mentoring can compensate for both resource constraints and low motivation. Girls who started mentoring with high educational capital and high motivation also tended to improve more, indicating a positive Matthew effect. However, among well-resourced participants with lower motivation, online mentoring was detrimental to development. These findings underscore the importance of considering mentees' prerequisites in explaining mentoring success and offer practical implications for talent mentoring for girls in STEM and beyond.
The present study examines how elementary school students' home learning workspaces relate to academic performance. We investigate these workspaces using the sociotope approach, which views environments as action spaces that differ in objective affordances that make actions (e.g. learning) feasible and in normative affordances that signal which actions (e.g. learning) are expected and valued. To test whether the home learning workspaces of 84 German third- and fourth-grade students were related to their achievements, 282 teacher-education students rated anonymized photographs of these workspaces using the sociotope approach. To complement this, the elementary students also provided self-reports on their home learning workspaces. Grade point average (GPA) served as the achievement indicator, and socioeconomic status (SES) was included as a control variable. Higher-achieving students' workspaces received more favorable third-party ratings of the feasibility of learning actions and normative support for learning, although not all workspace characteristics differed between achievement groups. Third-party ratings of the workspaces predicted achievement-group membership and provided information beyond SES. Self-reported workspace characteristics also predicted GPA and explained additional variance beyond SES. The findings are consistent with viewing home learning workspaces as structured constellations of affordances and highlight the value of combining third-party and learner perspectives when examining academic achievement.
Previous research has tracked how students' reading abilities grow over time, but little is known about how much variability exists in reading abilities for rural students in low-socioeconomic schools nor how the ecology of the classroom environment facilitates student growth. Thus, the instructional reading levels of three cohorts of first-grade students (n = 468) attending a rural, low-socioeconomic primary school were assessed in the fall, winter, and spring. Results indicated substantial variability in student instructional reading level within classrooms (n = 27). In the average classroom, student instructional reading levels ranged from readiness to third grade at the beginning of first grade and then readiness to fourth grade at the end of first grade. We also investigated whether the alignment between student instructional reading levels of students and classrooms at the beginning of the year was predictive of changes in instructional reading levels. Results indicated alignment - not misalignment - was predictive of changes in instructional reading levels. Specifically, students whose initial instructional reading level matched their classroom's average experienced differential change - those students with lower initial instructional reading levels demonstrated smaller increases in their instructional reading levels than students with higher instructional reading levels. Implications, limitations, and future directions are discussed.
