The rapid advancement of digital technologies has significantly altered learning styles, transitioning from traditional VAK (visual, auditory, kinesthetic) methods to E-VAK, incorporating computers and interactive applications. This shift necessitates continuous updates to educational materials. Augmented reality (AR), a technology merging real and digital interactions, presents an opportunity to enhance learning environments, boosting both practical skills and conceptual understanding. The integration of AR with other technologies like VR and IoT aligns with learner-centered strategies. AR uniquely links physical reality with virtual data, providing enhanced visual and auditory representations of knowledge. AR's potential to create a constructivist learning environment, emphasizing active learning and realistic scenarios, is significant. Further, AR principles align with modern communication theory, focusing on self-learning and interactive knowledge acquisition. AR shows promise in STEAM (Science, Technology, Engineering, Arts, and Mathematics) education, which aims to provide creative education utilizing cutting-edge technology. STEAM competencies are crucial for future economic growth, and AR can address the decreasing engagement of youth in STEM fields. Critical thinking (CT), a multifaceted cognitive construct, involves inductive and deductive reasoning, and creative problem-solving. Developing CT skills is a key objective in education, requiring learners to engage with real-life experiences and practice data collection, interpretation, teamwork, and decision-making. This study addresses the challenge of low STEAM test scores among Saudi intermediate students and their limited computer technology skills, aiming to improve science learning outcomes using AR applications.

Augmented reality (AR) combines virtual and real environments using specialized software on smart devices to display digital content such as images, videos, and 3D models. AR's educational applications are increasing, with studies exploring its benefits and drawbacks. AR systems require a mix of real and virtual elements, real-time interaction, and 3D registration. AR is categorized into location-based AR, using GPS data, and vision-based AR, using image recognition techniques. Vision-based AR can be marker-based (using QR codes etc.) or markerless. The integration of AR in STEAM education enhances interest and engagement in science, fortifying learner-centered activities and promoting knowledge building. Key benefits include increased motivation, collaboration, and improved spatial awareness. Interactive simulations and virtual laboratories enhance the learning experience and aid in memorizing procedural knowledge. AR can also foster higher-order thinking skills. However, implementing AR faces challenges, such as the usability of AR devices, technical issues with applications (limited scaffolding, low simulation accuracy), and institutional constraints. Despite these challenges, AR can significantly improve comprehension, memory, and knowledge transfer by engaging students actively in their learning. Critical thinking (CT) involves analyzing arguments, inductive/deductive reasoning, evaluation, and problem-solving. STEAM education aims to enhance CT skills, fostering problem-solving, collaboration, self-learning, and critical thinking. Studies show that STEAM improves CT and 21st-century skills. CT involves high-order reasoning, evaluating information, and finding solutions. Factors influencing CT skills include learning preferences, conceptual understanding, and problem-solving abilities. Mental capacity, determined by factors like work efficiency and adaptability, is also relevant. Augmented reality can support various learning styles and cognitive structures, enhancing mental capacity and motivation. Learning retention (LR) refers to retaining information for later use. Variables affecting LR include attention, self-efficacy, relevance, and testing. AR has shown potential in improving test scores and retention rates, although the impact on long-term memory needs further investigation.

This quantitative study employed a quasi-experimental pretest/posttest control group design. 120 8th-grade students (63 female, 57 male) from six private schools in Ha'il, KSA, participated. Students were divided into a control group (traditional teaching) and eight experimental groups based on gender (male/female) and mental capacity (high/low, determined using the Juan Pascual Leone cross-shape scale). The HP Reveal AR application was used in the experimental groups, with lessons on light waves, refraction, reflection, and the use of mirrors. Data collection involved achievement tests for science knowledge (validity 0.78, reliability 0.88) and the Cornell Arabized Saudi Environment Scale for CT skills (reliability 0.92). The intervention period lasted three weeks, including 120 min of AR application use and 240 min of face-to-face instruction.

MANOVA analysis revealed a significant interaction effect of mental capacity, AR design (image/marker), and gender on learning retention (LR). Post-hoc LSD tests showed that the highest LR was observed in the male, high mental capacity, marker-based AR group (G1), followed by other groups with high mental capacity, while the lowest LR was in the female, low mental capacity, image-based AR group (G8). The results suggest that incorporating diverse multimedia elements like images, audio, and video in AR applications improves student engagement and attention, leading to enhanced LR. Regarding critical thinking (CT) skills, MANOVA showed a significant effect of mental capacity but not gender or AR design. LSD tests indicated significantly better overall CT skills in high mental capacity students compared to low mental capacity students. Finally, analysis of learning outcomes showed a significant effect of mental capacity, but no significant effects of gender, AR design, or interactions. Post-hoc tests revealed better learning outcomes for male students than female students. These findings support the notion that AR applications enhance learning retention and that mental capacity plays a crucial role in both learning retention and CT skill development. While the study did not find a significant effect of AR type or gender on CT skills, it is important to note the positive effect AR had on male students' learning outcomes.

The findings support the hypothesis that AR enhances learning retention and CT skills, particularly for students with high mental capacity. The superior performance of male students in science learning with AR may be attributed to differences in learning styles or engagement with the technology. The study's results have implications for educators, highlighting the potential of AR to improve learning outcomes in STEAM education. The use of diverse multimedia elements within AR applications enhances student engagement and attention, leading to better learning retention. The significant role of mental capacity underscores the importance of catering to individual learning needs. The lack of significant differences in CT skills based on gender or AR type suggests that AR can benefit students regardless of these factors. However, the finding of better learning outcomes for male students requires further exploration. Future research should investigate the specific factors that contribute to gender differences in AR-enhanced learning.

This study demonstrates the effectiveness of AR technology in enhancing learning retention and critical thinking skills in a STEAM educational setting. The findings highlight the importance of mental capacity and suggest further research on gender differences in AR-enhanced learning. Future research should explore the long-term effects of AR, investigate its use in diverse populations, and address practical issues like cost and implementation strategies. The integration of AR-based active learning in future curricula could significantly improve STEAM skill acquisition.

The study's limitations include a relatively small sample size, focusing on a specific science topic and private school students. The short intervention period may also limit the generalizability of the findings. The study also assumes that all learners can benefit from AR, without considering the needs of students from underprivileged backgrounds or those with special educational needs.