Epistemological Silence in Excel’s MODE Functions: Tensions Between Statistical Definition and Computational Mediation

Abstract

Objective: To critically examine the behavior of the Excel 365 functions MODE, MODE.SNGL, and MODE.MULT when applied to amodal, unimodal, and polymodal datasets, analyzing the mismatch between the statistical definition of the mode and its computational implementation in a tool widely used in educational contexts.

Methodology: A qualitative study of an analytical-conceptual nature was conducted, based on three complementary procedures: documentary analysis of Excel’s statistical functions, computational experimentation using ten simulated datasets, and comparative analysis with responses generated by Copilot. As a technical counterpoint, a custom VBA function was also developed to explicitly identify different types of mode.

Results: The tests showed that Excel does not explicitly recognize amodality, returning the error #N/A, and presents limitations in representing ties of maximum frequency, particularly when it relies on automatic cell expansion in the MODE.MULT function. In contrast, Copilot correctly identifies polymodal situations and explicitly explains the statistical reasoning involved. The VBA function developed in this study demonstrated that it is technically feasible to align computational implementation with the formal statistical definition.

Conclusions: The results indicate that implementation decisions embedded in digital tools may generate conceptual tensions in statistics education when their outputs are interpreted as equivalent to formal mathematical definitions. The analysis reinforces the need for critical use of computational technologies in educational contexts and highlights the importance of making explicit the conceptual assumptions that guide the technological mediation of statistical knowledge.