Abstract
We performed a computer assisted experiment to test the accuracy of different ratio scales in estimating vegetation cover. Sixteen subjects estimated the cover level of artificial vegetation patterns displayed on the screen for various levels of resolution (from presence/absence to 100 different states, each measured on the ratio scale). We found that estimation error is minimum when the range of cover is divided into ten equal parts. Finer resolution gives less precise estimation since subjects tend to divide cover level into ten or at most twenty intervals in their mind