Good point Jerry, before you start calculating parametric descriptors
(mean) you should really ensure that you have normally distributed,
continuous data or you run the risk of getting erroneous results. If the
data is not normal then the non parametric descriptors of centre point
estimation, median or the mode, are likely to be more accurate. If we are
getting into that then you should probably be consider the error inherent in
the data and the degree of certainty with which you have described the
distribution; whether you're using the mean, median or the mode. If you don't
then you're pretty much guessing anyway..... If the variance is large you
may end up with an RPA average of 8m with a standard error of 4m. Should you
go with 4 or 12? It should be possible to conduct some surveys and do a bit
of analysis to determine what the best way to approach mulit stem
calculations actually is or that may be overkill.
________________________________
From: Jerry Ross <trees@xxxxxxxxxx.co.uk>
To: UK Tree Care <uktc@xxxxxx.tree-care.info>
Sent: Monday, 28 May 2012, 9:22
Subject: Re: BS5837-2012 - Combined Stem Diameter
Would the Median* x number of stems be better than using the average?
(*i.e. the diameter of the 'middling' stem -
http://en.wikipedia.org/wiki/Median )
That would tend to get rid of the distortion caused by outliers.
Quite how you find the median without measuring them all is, I admit, a bit
of a problem.
Personally I'll go on using the 4.6.1 a) for a measurable number of stems and
my judgement for the rest.
On 28/05/2012 09:03, Richard Nicholson (Leisure) wrote:
<Rupert wrote
I have continued to measure all the stems on a MS tree, added
x-sectional areas, and derived a figure from a notional stem of the same
cross-sectional area, in view of the gremlins in the new proposals.
<Tim added
that would work Rupert - in other words just use the first calculation.
Gets a bit tricky when there are 30 or 50 stems but in that situation they
are often of similar size and the second calculation works better then.
RN
You will recall that in the draft for public consultation there was only
one method, what is now the calculation in 4.6.1 a)
The general comment from arbs. in the form of DPC returns and comments at
the AA Reg. Con. meeting was that its not reasonable to expect an arb to
measure every stem. So we added an average calculation as in 4.6.1 b). It
will only give an average, that’s what averaging gives. It is possible in
any such averaging calculation to manipulate the input figures and produce
anomalies, but after all its about measuring trees not assessing a system
against the womping willow or any other tree which can grow a stem as you
look at it.
If Rupert and Tim want to continue to measure every stem, keep calm and
carry on.
-- The UK Tree Care mailing list
To unsubscribe send mailto:uktc-unsubscribe@xxxxxx.tree-care.info
The UKTC is supported by The Arbor Centre
http://www.arborcentre.co.uk/
--
The UK Tree Care mailing list
To unsubscribe send mailto:uktc-unsubscribe@xxxxxx.tree-care.info
The UKTC is supported by The Arbor Centre
http://www.arborcentre.co.uk/
