QTRA - I'm sorry i don't agree with it!

[RobArb]

I just can't be drawn into a system that relies on statistics as its main factor. Statistics in any walk of life can be read (or interpreted) to twist how that person wants to put those statistics across, for example (not related and quite simplistic, but bear with) if i was to say that 8 in 10 people when asked "who was the best arborist" said "robin" than that would be pretty conclusive, but....

 

what you have to ask yourself is-

 

Who did the study?

What are the statistics measuring?

Who was asked?

How were they asked?

Compared with what?

 

So in my case, i actually asked 10 people with 8 being members of my family..... still a statistical study though...

 

 

 

Every statistical study uses numbers to represent the view or opinion that person is trying to put across, to support their own ideas and to make you arrive at a conclusion. Just be aware of how those statistics were brought together.

 

In the case of QTRA, the statistics are seem to be measuring risk. A complicated subject on its own, it has been deemed by QTRA that an acceptible risk is 1/10000. This from what i can tell has been derived from historical tree failures, but IMO history isn't always correct, the laws of averages isn't always right. No two people are the same, so why should two trees be?

 

This statement "Rodney Helliwell (Helliwell 1990)

suggests that 1/10,000 might be a suitable figure to start with as a limit of acceptable risk." seems to be the starting basis for QTRA but the words "suggest" and "might" make it sound like its been plucked from the air!!

 

Its still all based on interpretation and subjectivity (of which that Australian judge in the paper attached agreed with me) regardless of how we look at the figures, QTRA is a time wasting money making machine, there i said it

 

Anyway i'm wittering on now, i'll leave you with this...

 

"There are three ways to not tell the truth: lies, damned lies, and statistics."

Australian court case goode v burnside.pdf

[Amelanchier]

As far as I can see QTRA doesn't pretend to be anything but subjective.

[Albedo]

As far as I can see QTRA doesn't pretend to be anything but subjective.

 

I could be wrong Tony but something that is, or claims to be, Subjective... by definition would not rely on a quantitative measure like statistics.

[RobArb]

Not what the website suggest, but i suppose that is based upon entirely how you read it...

 

"The Quantified Tree Risk Assessment (QTRA) system applies established and accepted risk management principles to tree safety management. Firstly, the use of land upon which trees could fail is assessed and quantified, thus enabling tree managers to determine whether or not and to what degree of rigour a survey or inspection of the trees is required. Where necessary, the tree or branch is then considered in terms of both impact potential (size) and probability of failure. Values derived from the assessment of these three components are used to calculate the probability of harm.

 

The system moves the management of tree safety away from labelling trees as either ‘safe’ or ‘unsafe’ and thereby away from requiring definitive judgements from either tree assessors or tree managers. Instead, QTRA quantifies the risk of significant harm from tree failure in a way that enables tree managers to balance safety with tree values and operate to predetermined limits of tolerable or acceptable risk."

 

 

Ellison himself notes in his paper, with reference to the criterion “Probability of Failure’ that

 

Accurately assessing the probability that a tree or branch will fail is highly dependent upon the skill and experience of the assessor.

[Amelanchier]

Sorry that was a bit abrupt - your post is worth a longer response!

 

Statistics is not a dark art, it's a method of evaluating and presenting data. Sure it can be manipulated but what can't be? Subjectivity is... well... subjective!?!?

 

The application of stats in QTRA is to define a suitable threshold for action - 1/10,000 is presented because that is typically what is used in other areas of risk decision making. The calculation of the final QTRA value (the Risk of Harm or RoH) depends on a fundamentally subjective assessment of the tree - the likelihood that a certain part is going to fail and the remainder just quantifies the impact of that risk in a rigourous and systematic way. Having a threshold to benchmark those hazards against that can be justified places you in a stronger position than if you made one up yourself...

 

As for that Australian ruling - its odd how much that crops up. I was once reminded of it in the feedback sheet from a tutor in relation to some coursework that dealt with UK tree management as if it somehow rendered the whole QTRA enterprise redundant. AFAIK UK courts aren't obliged to heed passing comments from their peers down under anymore than they abide by ours.

[RobArb]

What do you make then of the response of David Lonsdale in the Bowen V National trust case 2011 in regards to QTRA?

 

 

There is in my opinion no need to use formal risk assessment during general tree

inspections. Inspectors should, however, understand the underlying principles

sufficiently to be able to identify any trees that clearly warrant further attention. Also,

tree risk assessment can be used retrospectively in order to help determine whether

a previous general inspection has been appropriately conducted. On the basis of the

visual evidence available on 2 January 2007, a retrospective assessment of the

incident branch is presented in the following paragraphs, by multiplying numerical

values for the three risk factors (A), (B) and © listed below5. General information

about the estimation of each of the three values, both in QTRA© and in a modified

version that can be alternatively applied5, is given in Appendix 4.

 

5 In the opinion of the author of the present report, there are instances where a re-definition of the three risk

factors in QTRA© would be appropriate. QTRA© Factor (A), Impact-potential, is based on the size of the tree

or part thereof, but it does not always represent the likely severity of impact. Factor (B), Probability of failure, is

straightforward, albeit inherently subjective in its assessment. Factor ©, Target value, incorporates a formula

for including property as well as human life in the estimation of the probability of harm, but it might be more

logical to re-configure factors (A) and © so that A is the potential outcome of impact (i.e. degree of harm) and

© is simply the probability of the target being present. In the case under consideration, use of this modified

approach would, however, not affect the calculation of the risk index, since the outcome of impact is simply

related to the weight of the incident branch and the probability of a person being present.

6 This factor is called “target value” in QTRA (rather than “occupancy”), because it includes an estimate of the

risk of damage to property, which is valued in proportion to the notional pecuniary cost of a human life.

 

Do you feel it has a place in tree assessments when each tree is so very different?

[Amelanchier]

I could be wrong Tony but something that is, or claims to be, Subjective... by definition would not rely on a quantitative measure like statistics.

 

I would argue that given our individual experiences and physiology, we are not capable, even for a minute, of an objective view. We can certainly strive for it and some of us are even obliged to (the judiciary more so than arb consultants!), but we can never truly achieve it.

 

Therefore we can only ever make a subjective interpretation of a quantative measure! We colour the data with our bias and history.

 

Also statistics can be qualitative. How would you rate the clarity of this reponse on a scale of 1-10, with 1 being the lowest and 10 being the highest...

[Amelanchier]

Not what the website suggest, but i suppose that is based upon entirely how you read it...

 

"The Quantified Tree Risk Assessment (QTRA) system applies established and accepted risk management principles to tree safety management. Firstly, the use of land upon which trees could fail is assessed and quantified, thus enabling tree managers to determine whether or not and to what degree of rigour a survey or inspection of the trees is required. Where necessary, the tree or branch is then considered in terms of both impact potential (size) and probability of failure. Values derived from the assessment of these three components are used to calculate the probability of harm.

 

The system moves the management of tree safety away from labelling trees as either ‘safe’ or ‘unsafe’ and thereby away from requiring definitive judgements from either tree assessors or tree managers. Instead, QTRA quantifies the risk of significant harm from tree failure in a way that enables tree managers to balance safety with tree values and operate to predetermined limits of tolerable or acceptable risk."

 

 

Ellison himself notes in his paper, with reference to the criterion “Probability of Failure’ that

 

Accurately assessing the probability that a tree or branch will fail is highly dependent upon the skill and experience of the assessor.

 

Look at it in context. What are the options for assessing a tree? You can go to it, poke around a bit, find a problem, follow VTA and if neccessary come up with a spec that will reduce the risk associated with the problem.

 

Great but is that tree worth doing more than its neighbour? Is it worth doing at all? What do you compare it with? How can you show that you assessed both it and its neighbour on as equal a footing as possible? How can you show that the work you specified reduces the risk? How do you ensure there is a paper trail of your decisions?

[Albedo]

I would argue that given our individual experiences and physiology, we are not capable, even for a minute, of an objective view. We can certainly strive for it and some of us are even obliged to (the judiciary more so than arb consultants!), but we can never truly achieve it.

 

Therefore we can only ever make a subjective interpretation of a quantative measure! We colour the data with our bias and history.

 

Also statistics can be qualitative. How would you rate the clarity of this reponse on a scale of 1-10, with 1 being the lowest and 10 being the highest...

 

My own subjective score for your post Tony is 10/10 as always:001_smile:

 

My Science / quantitative head gives you a -2/10

 

Unless of course you wan't to get into paradigm busting then the second score goes up a bit:001_smile:

 

I hate it when the doctor says 'how much does it hurt...on a scale of 1-10'. I always want to argue with him on the parameters or basis of his scale:001_smile:

 

Edit: Peer review and replicability takes some of the subjectivity out of data interpretation.

Edited November 27, 2012 by Albedo

[Amelanchier]

What do you make then of the response of David Lonsdale in the Bowen V National trust case 2011 in regards to QTRA?

 

There is in my opinion no need to use formal risk assessment during general tree

inspections....

 

Do you feel it has a place in tree assessments when each tree is so very different?

 

Clearly Mr Lonsdale is entitled to his opinions (well that is to say, they are certainly his to have).

 

I don't know that I would agree that every tree different in respect to hazard assessment. There are certainly a great number of trees that are much the same (i.e., plantations / shelter belts) and even more trees that are just too young, small or remote to be a hazard. Of those that remain gravity ensures that there are only so many ways that a tree can fail and those failure modes affect a predictable area around the tree. Very few explode; even fewer implode.

 

It is within that small set of possibilities that we are expected to differentiate probability with a liberal smattering of precaution and I am personally of the opinion that it is reasonable to set out a system for doing so. Quite often, I use QTRA for doing so - especially when there are a lot of trees to compare and prioritise for a client.