Formative patterns of vegetation responses to cattle grazing can be difficult to detect because of innate heterogeneity of vegetation communities and grazing patterns, especially in semiarid environments. Nonetheless, some of this heterogeneity can be accounted for using appropriate experimental designs and statistical analyses. Bork and Werner (1999) suggested retention of negative vegetation consumption values for paired herbivory studies to concurrently analyze spatial heterogeneity and intensity of herbivory. Yet, inclusion of negative consumption metrics has not been widely utilized. Our objective is to determine if postulates proposed by Bork and Werner (interpretation of effects when negative consumption values are manipulated) are applicable to a real data set utilizing a paired experimental design in a heterogeneous, semiarid environment. We determined manipulation of negative consumptive values affects interpretation; it skews data distribution, over-estimates treatment effect, and results in a statistically weaker treatment effect. Therefore, we recommend employing a biologically paired experimental design and the retention of all negative vegetation consumption values so that: (1) resulting models represent a normally distributed population, (2) standing crop of vegetation resulting from the treatment is not over-estimated and (3) grazer treatment effect in heterogenous environments is recognized if existent. This will ensure conservative metrics guide management decisions.