This article presents a methodology for teaching interval analysis in economic modeling, using the Leontief input-output model as a case study implemented in the GNU Octave environment. The limitations of traditional deterministic, probabilistic, and fuzzy approaches are discussed, particularly their reliance on precise data or subjective assumptions that hinder educational applications. Interval analysis, which formalizes uncertain parameters through guaranteed bounds, is shown to be methodologically preferable for economic problems characterized by incomplete information and statistical data scarcity. A complete computational workflow is developed using the freely available GNU Octave interval package (version 3.2.1), which fully implements the IEEE 1788-2015 standard for interval arithmetic. The paper provides a detailed algorithm for interval input-output balance problems and demonstrates its application to a three-sector economic system with 5% data variation. Through comparative analysis, it is shown that deterministic planning based on average values leads to underproduction and overproduction risks under worst-case scenarios, while the interval approach provides guaranteed planning ranges. The pedagogical value of this approach is emphasized: it develops critical thinking, fosters robust analysis skills, and prepares future specialists for decision-making under real-world uncertainty. All code is provided for full reproducibility, making the methodology directly applicable in educational settings.