A Non-Parametric Analysis of Interrelationships Among Mathematics Learning Interest Indicators: Kendall's Tau-b and Kendall's W Approaches Among Junior High School Students

Abstract

Most studies treat mathematics learning interest as a single composite construct linked to learning outcomes, without first testing whether its constituent indicators, namely enjoyment, attraction, attention, and engagement, are themselves statistically interrelated. This study addresses that gap by examining the strength of pairwise associations among these indicators and their overall concordance. A quantitative, cross-sectional correlational design was employed using data from 53 seventh-grade students at a junior high school in Yogyakarta, selected through simple random sampling. Because the Kolmogorov-Smirnov normality test indicated that three of the four indicators departed from normality, non-parametric tests were applied: Kendall's tau-b for pairwise correlation and Kendall's W for overall concordance. The results show that only three of six indicator pairs were significantly and fairly associated, enjoyment-attraction (τb = 0.463), attraction-attention (τb = 0.281), and attraction-engagement (τb = 0.258), while the remaining three pairs showed very weak, non-significant associations. Kendall's W of 0.264 (statistically significant) indicates that the four indicators exhibit real but modest overall concordance. These findings offer both methodological and practical contributions: aggregated interest scores may obscure heterogeneous student profiles, so diagnosis and instructional intervention should target specific indicators, with attraction serving as the most central connective dimension, rather than treating learning interest as a homogeneous construct.