In forest mensuration, forest inventory and forest modeling, it is common to compare the agreement between two sets of quantitative measurements. One set could be from a commonly-accepted...
Description
In forest mensuration, forest inventory and forest modeling, it is common to compare the agreement between two sets of quantitative measurements. One set could be from a commonly-accepted reference method, the other from a new or alternative method, device, or model. If the measurements are statistically the same or practically "close enough" to each other, the more complex, more destructive and/or costly method can be replaced by the simpler, less destructive and/or more efficient method. Traditional statistical approaches used for such comparisons based on bias, correlation coefficient, coefficient of determination and significance tests are inadequate and not really relevant for agreement analysis. Various agreement measures proposed for agreement analysis are also insufficient and not pertinent in many respects. An alternative approach based on the limits of agreement proposed by Bland and Altman (1986) is introduced and demonstrated in this study, along with the method...