Development and evaluation of a standard area diagram set for the severity of phomopsis leaf blight on eggplant

Abstract

A standard area diagram set (SADs) with eight severity values ranging from 0.5 to 32% was evaluated as a tool to improve the accuracy of the estimates of Phomopsis leaf blight severity on eggplant by ten inexperienced raters. A first assessment in a 50-leaf dataset of digital images was performed unaided (UN, no SAD). Two further aided (A) assessments, conducted two (A1) and four weeks after the first (A2) unaided one were performed. The precision of the estimates, as indicated by the Pearson’s correlation coefficient (r), improved significantly when using the aid (A1: r = 0.96 and A2: r = 0.96) compared to unaided (UN: r = 0.79). However, the generalized bias (C b) was not significantly affected because the unaided estimates were already quite accurate (C b = 0.93). The overall concordance (ρ c) was significantly improved due to the large gains in precision when using the SADs. The raters’ estimates were more uniform (ρ c > 0.9) when using the aid, with progressive gains in the reliability of the estimates among them, indicated by the concordance correlation coefficient statistics. The SADs proposed in this study will be useful in severity estimation during field work involving multiple raters, especially as use of the SADs results in less variable estimates, improving accuracy and reliability among raters.