김화경 Hwa Kyung Kim

DOI: JANT Vol.65(No.3) 403-420, 2026

---

This study analyzed Korean students' growth in mathematics at the content-domain and topic levels using the TIMSS 2023 Longitudinal Study, in which students who took the 2023 assessment were retested with the same items one year later (Grades 4-5: 4,507 students; Grades 8-9: 4,410 students). Because the longitudinal database provides no domain-level scale scores, abilities were estimated at both time points on a common scale using EAP estimation with the published IRT item parameters fixed, and growth was compared in standardized units. First, students in Grades 4-5 grew in all domains, but the data domain grew the most slowly and consequently ranked lowest among the three domains at the second time point, whereas growth in Grades 8-9 was stagnant except in algebra, indicating that the strengths and weaknesses reported in cross-sectional studies are being reproduced as differences in growth. Second, growth in Grades 8-9 was driven largely by female students, and low performers caught up in the number domain but fell further behind in functional thinking. Third, algebra - functions in particular - showed the largest contrast in growth between socioeconomically disadvantaged students who reached the high international benchmark and those who did not, and growth in data interpretation differed most by socioeconomic background. All group comparisons are descriptive and exploratory. Based on these findings, we propose restructuring statistics education in connection with other subjects such as social studies and science, and strengthening early algebra education.

Hoyun Cho , Jukyung Park

DOI: JANT Vol.65(No.3) 421-449, 2026

---

As artificial intelligence (AI) rapidly reshapes K-12 education, the distinctive pedagogical demands of mathematics instruction call for subject-specific policy attention. This study analyzes state-level AI policies across 29 U.S. states and Puerto Rico through March 2026, asking how jurisdictions address AI integration in mathematics education. Using qualitative document analysis, the study applies a graded classification of mathematical relevance, from Tier A (mathematics-substantive) through Tier C (mathematics-relevant but not subject-named) to Tier D (subjectneutral), and derives a typology of four policy approaches through inductive coding: Legislative and Task Force, Comprehensive Principle-Based Guidance, Toolkit and Resource-Based, and Targeted Use Case Application. A central finding is the structural scarcity of Tier A policies. Even after substantively relevant Tier C policies are credited, mathematics-substantive provisions are virtually absent, which leaves mathematics educators to adapt general, subject-neutral directives. Cross-cutting analysis identifies shared priorities in AI literacy and assessment integrity, as well as a persistent gap in mathematics teachers’ AI adoption and professional development.

황윤선 Yun Sun Hwang

DOI: JANT Vol.65(No.3) 451-472, 2026

---

This study examines how the Mathematical Logic-Based Instructional Model (MLIM) is operationalized in a multidisciplinary general education context and how learners perceive and experience mathematical logic learning. Quantitative and qualitative data were collected from the course “Introduction to AI Logic” over four semesters. The quantitative analysis investigated differences in learners’ perceptions of mathematical knowledge and logical thinking across academic disciplines, while the qualitative analysis examined cognitive difficulties and patterns of conceptual understanding based on open-ended responses. The findings indicate that mathematical logic learning involves structural cognitive barriers at the initial stage, where difficulties in understanding symbolic systems are combined with challenges in transitioning from natural language to formal representations. These difficulties appear across the overall learning process rather than within isolated concepts. In addition, learners’ thinking processes were gradually organized through repeated problem-solving and conceptual structuring. Despite initial differences across disciplines, learners reported relatively similar levels of logical thinking experiences after the course. Based on these findings, MLIM may be interpreted as an operational strategy in which the relative emphasis of questioning, modeling, and proof is flexibly adjusted according to learners’ cognitive states and instructional contexts. Rather than privileging a single phase, the model is characterized by the dynamic reconfiguration of all phases in response to learning conditions. In this process, the modeling phase often functions as a key component in supporting representational transition, while its relative emphasis, along with that of questioning and proof, varies depending on the learning context.

지민경 Minkyeong Ji , 황지현 Jihyun Hwang

DOI: JANT Vol.65(No.3) 473-496, 2026

---

This study aims to develop an instrument measuring secondary mathematics teachers’ perceptions of educational technology (EdTech) utilization based on the Expectancy-Value-Cost theory, designed to assess the motivational factors influencing their use of EdTech. Sub-factors were categorized within the dimensions of expectancy, value, and cost, and initial items were developed based on a comprehensive literature review and previous research. The results of a survey conducted with 213 secondary mathematics teachers confirmed that the final instrument, comprising 31 items across eight factors, is a valid and reliable measure of mathematics teachers' perceptions regarding EdTech utilization. Consequently, this developed instrument can serve as a foundational tool to systematically understand teachers' motivation, ultimately providing support for teachers in designing and implementing mathematics lessons utilizing EdTech.

이예진 Ye-jin Lee

DOI: JANT Vol.65(No.3) 497-514, 2026

---

This study aimed to develop an instrument for measuring teachers' beliefs about mathematical misconceptions and to examine the adequacy of the instrument and teachers’ belief tendencies using responses from in-service elementary teachers. Based on prior research on mathematical misconceptions and teacher beliefs, preliminary items were developed and revised through expert reviews in mathematics education and two pilot tests, yielding a final 28-item instrument, which was administered in the main survey. The survey was conducted with 108 inservice elementary teachers recruited through convenience sampling across multiple regions in South Korea. Descriptive statistics, reliability analysis, correlations among subdomains, and factor structure analyses were conducted to evaluate the instrument. The final instrument consisted of three subdomains: the nature and causes of misconceptions, instructional responses to misconceptions, and the instructional purposes and learning value of misconceptions. The overall Cronbach's alpha was 0.903, and subdomain reliability coefficients ranged from 0.719 to 0.899, indicating acceptable internal consistency. In addition, bifactor model analysis indicated the possible coexistence of subdomain-specific characteristics and a general belief factor common to all items. Teachers exhibited the highest constructivist-oriented beliefs regarding the instructional purposes and learning value of misconceptions, whereas their beliefs regarding instructional responses to misconceptions were relatively lower. These findings suggest that teachers recognize the educational value of misconceptions while also considering corrective and procedure-centered traditional beliefs when responding to misconceptions. This study provides an empirically grounded instrument and foundational data for research on mathematics instruction and teacher education related to misconceptions.

김소민 Somin Kim , 김희정 Hee-jeong Kim

DOI: JANT Vol.65(No.3) 515-537, 2026

---

This study compared the mathematics motivating efficacy between secondary in-service and pre-service mathematics teachers, with particular attention to whether group differences varied across types of motivating strategies. Data were collected from 143 in-service and 86 pre-service secondary mathematics teachers using the Motivating Efficacy Scale for Mathematics Teachers (MESMT). Confirmatory factor analysis confirmed a 21-item structure comprising four subfactors: providing successful experiences, eliciting attention and engagement, creating mathematics context-based relevance, and providing extrinsic rewards. Independent samples t-tests were then conducted at both the subfactor and individual item levels. The results showed that in-service teachers had higher mean scores than pre-service teachers across all four subfactors, but a statistically significant difference emerged only in the efficacy for providing extrinsic rewards, although the effect size was small (d=0.281). This difference, however, did not remain significant after correction for multiple comparisons and after controlling for background variables. At the item level, among the six items that differed significantly, in-service teachers scored higher on five items related to practice-based strategies, whereas pre-service teachers scored higher only on the item concerning helping students attribute their success to their own effort and ability. Only one item―maintaining consistency in the difficulty of tasks and assessments―remained significant after correction for multiple comparisons, and no item remained significant when background variables were additionally controlled. The mathematics context-based relevance subfactor showed the lowest mean efficacy in both groups. These findings suggest that mathematics motivating efficacy may not be uniformly higher or lower across all domains depending on whether teachers have classroom experience, but may instead form differently depending on the nature of the motivating strategy and the teaching task. This study suggests that mathematics teacher preparation and in-service professional development should be designed to strengthen strategy-specific competencies, particularly by providing opportunities to connect theoretical knowledge with practical classroom implementation.

장현석 Hyun Suk Chang , 김민아 Mina Kim , 이봉주 Bongju Lee

DOI: JANT Vol.65(No.3) 539-560, 2026

---

This study aimed to examine pre-service and in-service teachers’ perceptions of the use of three visual proof methods―two-column proof, proof map, and flow-chart method―in secondary proof instruction. For this purpose, a questionnaire was developed based on previous studies, and a survey was conducted with a total of 61 participants, including 31 pre-service teachers and 30 in-service teachers. The results of the study are as follows. First, the two-column proof method showed the highest mean responses in terms of perceived educational value and recommendation for its introduction into the Korean mathematics curriculum. However, for all three visual proof methods, the mean responses for curriculum introduction recommendations were relatively lower than those for educational value. This suggests that although teachers acknowledged the educational value of visual proof methods, they held a more cautious attitude toward their incorporation into the curriculum. Second, in terms of perceived necessity, the two-column proof method showed the highest mean responses across all subdimensions. In particular, teachers perceived the two-column proof method as the most helpful method for supporting students’ development of proof-related thinking compared with the conventional sentence-based proof writing method. Third, comparisons between in-service and pre-service teachers showed that, in terms of educational value and recommendations for curriculum inclusion, in-service teachers tended to perceive the proof mapping method more positively than pre-service teachers. In contrast, with respect to perceptions of necessity related to supporting students’ development of proof thinking, the two groups showed generally similar levels of positive responses to all three visual proof methods. These findings suggest that visual proof methods have the potential to be used in secondary proof instruction as teaching and learning methods that connect intuitive understanding with formal reasoning.

여승현 Sheunghyun Yeo

DOI: JANT Vol.65(No.3) 561-579, 2026

---

This study examines how AI chatbot-based simulation classes influence the development of mathematical noticing and questioning abilities among elementary pre-service teachers. A chatbot named Minji was designed to simulate a third-grade student who harbors misconceptions about fractions, specifically the belief that any shape divided into three parts represents one-third, regardless of whether the parts are equal in size. Sixty-two second-year students enrolled in the course Elementary Mathematics Textbook Research and Instructional Methods at D University of Education participated. The study classified mathematical noticing into three hierarchical levels: descriptive noticing, interpretive noticing, and responsive noticing. Data were analyzed through overall level distribution tables at the planned and actual stages, case analyses for each level, and a transitional (alluvial) analysis. Quantitative findings showed a meaningful shift: descriptive-level responses decreased from 35.5% (planned) to 22.6% (actual), while responsive-level responses increased from 25.8% to 35.5%. A Stuart-Maxwell test indicated that the change in marginal distributions approached significance (χ²=5.23, df=2, p=0.073). Transitional analysis revealed that 54.5% of the participants who had planned at the descriptive level successfully demonstrated higher-level responses during actual interaction. Qualitative case analyses illustrated how the chatbot’s student-role responses and an indialogue mentor teacher feature supported pre-service teachers’ progression toward interpretive and responsive noticing during interaction, while post-dialogue automated feedback supported their subsequent reflection. These findings suggest that AI chatbot-based simulations offer a promising pedagogical tool for strengthening pre-service teachers’ professional vision in mathematics education.

김주희 Juhee Kim

DOI: JANT Vol.65(No.3) 581-602, 2026

---

This study examined high school students’ interpretations of confidence intervals and confidence levels in a lesson on estimating a population mean, the public status of those interpretations, and explanations and justifications offered when different samples were considered. Drawing on Cobb and Yackel’s emergent perspective and Yackel and Cobb’s concept of sociomathematical norms, this study analyzed one lesson from a Probability and Statistics course using a qualitative case-study design. Data comprised video and audio recordings, transcripts, student worksheets, presentation materials, and researcher memos. The analysis focused on the statistical meanings of students’ utterances, the public status of explanations, the teacher’s discourse orchestration, and normative cues. Students interpreted a 95% confidence level as the probability that the population mean lies within a particular calculated interval. Through the teacher’s revoicing and peer agreement, this interpretation was presented as a candidate explanation for public consideration. Two non-overlapping confidence intervals then brought it into conflict with the fixed nature of the population mean. One student used independent-trial reasoning and a dice analogy to draw on repeated sampling, sample variability, and possible coverage failure as explanatory resources, linking the 95% confidence level to the possibility that each calculated interval contains the population mean and thereby rejustifying the problematized interpretation. This study interprets the utterance sequence as reparative justification incorporating productive statistical resources. The case traces how a misconception about confidence intervals was presented in classroom discourse as a candidate public explanation, problematized, and rejustified through new statistical resources and an analogy.

장민 Min Jang , 박영희 Younghee Park

DOI: JANT Vol.65(No.3) 603-619, 2026

---

This study designed and implemented operational strategies for generative AI-based mathematics performance assessment and integrated students’ perceptions of fairness with evidence from their actual AI-use processes. The participants were 59 twelfth-grade students: 53 in Calculus and six in AI Mathematics after duplicate participation was removed. A mixed-methods design was employed using a 12-item fairness perception survey, student prompt logs, teacher observation journals, and semi-structured interviews with 10 students. Based on prior research and policy documents, the survey was organized into four dimensions―fairness of opportunity, procedural transparency, content validity, and educational acceptability―and refined through review by two in-service mathematics teachers experienced in generative AI-supported instruction. Quantitative and qualitative data were integrated using the four dimensions as a common framework, with recurring meanings and behavioral characteristics identified across prompt logs, observation journals, and interviews. The survey showed high internal consistency (Cronbach’s α=0.906). Procedural transparency received the highest mean score (M=4.50), followed by educational acceptability (M=4.32), fairness of opportunity (M=4.21), and content validity (M=4.18); all four dimensions were rated positively. Qualitative analysis identified three patterns of AI use: logical inquiry and active verification, trial-and-error-based refinement, and error correction and conceptual reconstruction. Rather than accepting AI-generated outputs uncritically, students revised questions, checked calculations and conditions, corrected errors using mathematical concepts, and reconstructed explanations in their own words. The integrated findings indicate that students’ positive perceptions of fairness were associated not with the mere use of generative AI but with common devices and AI environments, prior instruction, explicit assessment criteria, prompt-log documentation, teacher observation, and assessment focused on mathematical verification, correction, and reconstruction. These findings suggest that students’ mathematical judgment and problem-solving processes, rather than prompt-writing skill or the completeness of final products, should serve as central assessment evidence. Because this exploratory study was conducted at a single school with markedly unequal course samples, the findings should be interpreted as students’ perceptions of fairness under the implemented conditions rather than as objective proof of assessment fairness.