Disease Severity Estimates – Effects of Rater Accuracy and Assessment Methods for Comparing Treatments
American Phytopathological Society, 2015
Online
academicJournal
Zugriff:
Assessment of disease severity is required for several purposes in plant pathology; most oftenthe estimates are made visually. It is established that visual estimates can be inaccurate andunreliable. The ramifications of biased or imprecise estimates by raters have not been fullyexplored using empirical data; partly because of the logistical difficulties involved in differentraters assessing the same leaves for which actual disease has been measured in a replicatedexperiment with multiple treatments. In this study nearest percent estimates (NPEs) ofSeptoria leaf blotch (SLB) on leaves of winter wheat from non-treated and fungicide treatedplots were assessed in both 2006 and 2007 by four raters and compared to assumed truevalues measured using image analysis. Lin’s concordance correlation (LCC, ρc) was used toassess agreement between the two approaches. NPEs were converted to Horsfall-Barratt (HB)mid-points and again compared for agreement with true values. The estimates of SLB severityfrom fungicide-treated and non-treated plots were analyzed using generalized linear mixedmodeling to ascertain effects of rater using both the NPE and HB values. Rater 1 showedgood agreement with image analysis (ρc = 0.986 to 0.999), while raters 3 and 4 had less goodagreement (ρc = 0.205 to 0.936). Conversion to the HB scale had little effect on bias oraccuracy, but reduced both precision and agreement for most raters on most assessment dates(precision, r = -0.001 to -0.132; and agreement, ρc = -0.003 to -0.468). Inter-rater reliabilitywas also reduced slightly by conversion of estimates to HB midpoint values. Estimates ofmean SLB severity were significantly different between image analysis and raters 2, 3 and 4,and there were frequently significant differences among raters (F=151 to 1260, P=0.001 to<0.0001). Conversion to the HB scale changed the means separation ranking of raterestimates on 26 June 2007. Nonetheless, image analysis and all raters were able todifferentiate control and treated plots treatments (F=116 to 1952, P=0.002 to <0.0001,depending on date and rater). Conversion of NPEs to the HB scale tended to reduce F-valuesslightly (2006: NPEs, F=116 to 276, P=0.002 to 0.0005, and for the HB converted valuesF=101 to 270, P=0.002 to 0.0005, and in 2007, NPEs, F=164 to 1952 P=0.001 to <0.0001,and for HB converted values F=126 to 1633 P=0.002 to <0.0001). The results demonstratedthe need for accurate and reliable disease assessment to minimize over or underestimatescompared to actual disease, and where multiple raters are deployed, they should be assignedin a manner to reduce any potential effect of rater differences on the analysis.
Titel: |
Disease Severity Estimates – Effects of Rater Accuracy and Assessment Methods for Comparing Treatments
|
---|---|
Autor/in / Beteiligte Person: | Bock, Clive ; El Jarroudi, Moussa ; Kouadio, Amani Louis ; Mackels, Christophe ; Chiang, Kuo-Szu ; Delfosse, Philippe |
Link: | |
Veröffentlichung: | American Phytopathological Society, 2015 |
Medientyp: | academicJournal |
DOI: | 10.1094/PDIS-09-14-0925-RE |
Schlagwort: |
|
Sonstiges: |
|