Multivariate data analysis has been applied to study the effect of process variability on ring formation. The literature suggests that temperature fluctuation in a kiln may allow for mechanisms by which rings become hard and subsequently grow. Since accurate temperature data from inside a kiln is unavailable, this study examines the possible relationship between temperature fluctuation in a kiln and the phenomenon of ringing. In particular, partial least squares discriminant analysis (PLS-DA) was applied to historical kiln data from three different kilns, from three different mills.The results of this study suggest that PLS-DA is able to distinguish, characterize, and monitor clean and dirty kiln operation involving ringing. The results also suggest that ringing may be the result of a combination of variables, that it may be related to the amount of fluctuation in the variables, and that the experiences of ring formation differ from one mill to another.