This study examines how fourth-grade elementary students infer the relative temporal order of events from the superposition of lunar craters. A two-tier diagnostic item adapted from a released PISA science unit was administered to 195 fourth graders in Busan, Korea: tier 1 required the full ordering of five superposed craters, and tier 2 elicited written justification under teach-back framing. An embedded mixed-methods analysis combined quantitative scoring of the orderings with language-network analysis and lexicon-based classification of justification. The sequencing performance of students varied widely: 38.5% reproduced the keyed order exactly, and errors were concentrated not on global judgment— identifying the large outermost crater as the oldest, which most students correctly identified—but on successive local judgment of which of the two overlapping inner craters covered the other. Incorrect ordering fell into three recurring patterns (full reversal, adjacent single error, and internal scrambling), and the written justification of students who were correct and incorrect differed systematically: students who were correct described the observable overlap relations using a differentiated set of superposition verbs, whereas students who were incorrect most often assumed that larger or deeper craters were older. Correct ordering and scientifically appropriate justification were strongly associated but did not coincide; some students identified the correct order without citing superposition evidence, whereas others cited such evidence yet misordered the craters. These results indicate that assessing relative-age reasoning at this age requires examining students’ evidence and justification alongside their answers and that instruction should explicitly address the local reading of overlap relations and the size–age misconception.

Abstract
요 약
1. Introduction
2. Theoretical Background
    2.1. Chronological Reasoning and TemporalSequencing
    2.2. Relative Dating from Superposition andCross-Cutting Relations
    2.3. The PISA Released Unit and the Positionof the Present Instrument
    2.4. Two-Tier Diagnosis and the Separationof Product from Process
3. Methods
    3.1. Research Design
    3.2. Participants
    3.3. Instrument: “Craters on the Moon”
    3.4. Answer Key and Scoring
    3.5. Data Analysis
4. Results
    4.1. Sequencing Accuracy and the PairwiseError Map
    4.2. Subtypes of Incorrect Sequences
    4.3. Language Networks of Written Justifications
    4.4. Integrating Correctness and ReasoningType
5. Discussion
    5.1. Identifying the Oldest Crater Was Easierthan Sequencing the Inner Craters
    5.2. Correctness and Reasoning Validity AreDissociable
    5.3. The Size–Age Conflation as the DominantMisconception
    5.4. Implications in the Context of PISA 2025
    5.5. Limitations
6. Conclusion
References

• Dong-Young Lee(Science Education Research Institute, Pusan National University, Busan 46241, Korea) Corresponding author