Dear all,
I am using the point intercept transect method to estimate the proportion of
forest floor cover types. This method is rapid and easy and its use is widely
reported in the literature. I suppose that the accuracy of the estimate is
depending on the number and proportion of different cover types, their spatial
arrangement, and the number of sample points. I am not aware of any reports of
the accuracy of the estimates in published studies, but I suspect that their
accuracy is low, given that the sample size is often small (n << 50).
I was looking for an appropriate method to calcuate the accuracy of the
estimate, but didn't find anything. As a last resort I used a method applicable
to survey studies where the minimum sample size is equal to the product of the
proportions of responses (pi * pj) divided by the squared standard error
(SE^2). Obviously, this method does not really apply to the point intercept
transect method, one reason being that the points in the transect are not
independent. I suppose that as a consequence of this the accuracy of the
estimate is overestimated. Please note, that I am not really interested in the
spatial pattern at the transect level.
Can anyone suggest a more appropriate method to estimate the accuracy of the
point intercept transect method?
Thanks for your help.
Michael Drescher
---------------------------------
Yahoo! Messenger NEW - crystal clear PC to PC calling worldwide with voicemail
