“…However, in these STA studies, the quantitative statistical method for STA was Spearman rank correlation, which was insufficient for inferring monotonicity due to the lack of consideration of sampling error ( Kalish et al, 2016 ). Several techniques for quantitative STA have recently been developed ( Prince et al, 2012 ; Kalish et al, 2016 ; Benjamin et al, 2019 ), one of which is based on conjoint monotonic regression (CMR; Kalish et al, 2016 ). CMR compares the goodness-of-fit between a partial order model, in which the x-coordinates have the same partial order as the y-coordinates for all dots, and a monotonic model, in which the two coordinates have the same order as well as monotonically related to each other.…”