Procedimento estatístico para validação de escalas diagramáticas na quantificação de doenças

Abstract

Validation of disease diagrammatic scale is commonly done by using simple linear regression to test the identity between estimates and observed values of disease severity . In this case, when adjusting a linear regression, only evaluation of the hypotheses H01:β0 = 0 and H02: β1 = 1, are made without testing these hypotheses simultaneously, if β0 = 0 at same time that β1 = 1, or, if β1 = 1 at the same time that β0= 0. The objective of this work was to propose a novel statistical procedure to validate disease diagrammatic scales. The proposed procedure, namely L&O, results from the combination of the F test modified from Graybill (1976), t test for the medium error and analysis of the linear correlation coefficient. This procedure was applied to test the identity between estimates with diagrammatic scale and observed values of severity disease caused by Quambalaria eucalypti in eucalypt. The results showed that the proposed statistical procedure is efficient and not subjective and it could also be used to test the identity between any two vectors or groups of quantitative data.