<![CDATA[Metacognition plays an essential role in solving statistical problems. However, many statistical tasks used in classroom practice emphasize procedural computation rather than encouraging students to regulate their thinking. Consequently, tasks specifically designed to elicit students’ metacognitive activity in statistical contexts remain limited. This study aimed to develop a Statistical Problem-Solving Task (SPST) capable of eliciting said activities and to examine the distribution and characteristics of those activities during problem solving. The study employed an educational design research approach consisting of two stages: (1) task development based on analytical, non-routine, and ill-defined problem characteristics, followed by expert validation of content and clarity; and (2) a pilot experiment with 29 twelfth-grade students. The data collection procedure was conducted through a modified Mathematics Metacognitive Inventory, students’ written responses to the SPST, and semi-structured interviews with selected subjects. Two students representing the dominant metacognitive level identified in the questionnaire results were selected for in-depth interviews to explore their metacognitive processes during problem-solving. Quantitative data were analyzed descriptively to determine the distribution of students’ metacognitive activity levels, while qualitative data from students’ written work and interview transcripts were analyzed thematically and cross-checked against one another to identify converging patterns of metacognition during problem-solving. The results gathered from the quantitative inquiry showed that many of the students (75.9%) demonstrated moderate activity. Meanwhile, high metacognitive activity was observed in 6.9% of the students, with the rest (17.2%) showing low metacognitive activity. Furthermore, concurrent interview analyses revealed that metacognitive regulation was frequently accompanied by conceptual errors and incomplete evaluation. Ultimately, the study suggests that the SPST can function as a diagnostic and instructional instrument for fostering metacognition in statistics learning, though additional instructional support remains necessary.]]>
