This is, of course, a subtopic that appeared in a thread below, but it seems to me to be worth discussion. Suppose that your jurisdiction has a specification that says essentially that reporting the relative positions between two survey markers under 100 ft. apart with an error of more than 0.10 ft. (= 3cm Canadian) constitutes incompetent practice as is, for example, the case in Texas. What is the acceptable likelihood that some measurement process you employ will contain errors greater than the threshold of incompetent practice?
The nature of most random errors in surveys means that once you've decided whether you want to be 95% confident that you've met the spec or 99% confident, or whatever, you can then work backwards to methods that will give that result.
Naturally, in many jurisdictions if a surveyor certifies that a survey meets a more restrictive positional accuracy specification, such as those of the ALTA/ACSM minimum standards, it would be considered incompetent not to meet that more restrictive spec and it would control. But the general question is still in the same form. What likelihood of incompetence is acceptable?
I'll throw out the idea that the acceptable level of confidence to be used in designing measurement processes to meet the specification is somewhere between 95% and 99%, but not less than 95%.
Error Does Not Equal Incompetence
Assuming you are a careful measurer and use techniques such that there is a 99% probability of meeting the standards, there is still the possibility that a reported measurement is more than 0.10' off. One can prove one's techniques that it was an error and not incomptence.
It is my understanding having seen many of the simplistic methods endorsed here that 50% could not prove that it was an error.
Does that mean that that 50% is incompetent? That I cannot answer but your board can, and may in the future have to.
Paul in PA
13.The accuracy of the measurements for the survey shall be based upon the type of survey, and the current or expected use of the land. The accuracy of the measurements thus performed shall be substantiated by the computations of the traverse; the relative error of closure permissible shall be no greater than the following standards given below:
A.Where there is or will be zero lot line construction on small tracts in a high density urban area, the allowable closure error is 1:10,000 and the allowable positional error is plus or minus 0.10 feet.
B.In residential or commercial subdivisions where the length of lines does not exceed 300 ft, the area of tracts does not exceed 2 acres, and there is no plan for zero lot line construction, the allowable closure error is 1:10,000 and the allowable positional error is plus or minus 0.25 feet.
C.In suburban or rural residential or industrial tracts where the length of lines does not exceed 1000 feet and the area of tracts is between 2 and 40 acres, the allowable closure error is 1:7,500 and the allowable positional error is plus or minus 0.50 feet.
D.Rural tracts of 40 acres or more where the corners of the tract may be connected with traverse legs in excess of 1000 feet, the allowable closure error is 1:7,500 and the allowable positional error is plus or minus 1.5 feet.
E.Rural tracts of 40 acres or more in rough or tree covered terrain where the corners of the tract must be connected with short traverse lines because of poor visibility between the corners of the tract, the allowable closure error is 1:5,000 and the allowable positional error is plus or minus 3.0 feet.
F.Field work which has a closure error greater than the maximum allowed, or linear error of closure greater than the maximum positional error shown, shall be considered unacceptable and shall be corrected. Adjustment of a traverse must not shift the position of any point more than the maximum positional error listed above.
G.In lieu of maximum allowable positional error, the latest Accuracy Standards for ALTA/ACSM Land Title Surveys may be used for determining minimum accuracy requirements.
Error Does Not Equal Incompetence
> Assuming you are a careful measurer and use techniques such that there is a 99% probability of meeting the standards, there is still the possibility that a reported measurement is more than 0.10' off. One can prove one's techniques that it was an error and not incomptence.
Actually, that avoids the premise of the question. The premise is that it is given that an error above some threshold value that is present in the relative position of two survey markers you report is incompetent practice.
So, the question stands as one of the likelihood that the errors in your results as reported do not exceed that threshold value. What level of confidence would a prudent surveyor design the measurement process for? 95% confidence? 99% confidence? More? Much less?
> 13.The accuracy of the measurements for the survey shall be based upon the type of survey, and the current or expected use of the land. The accuracy of the measurements thus performed shall be substantiated by the computations of the traverse; the relative error of closure permissible shall be no greater than the following standards given below:
>[Okie standards omitted.]
That's nice that Oklahoma has any standards at all, but it doesn't answer the question does it?
Error Does Not Equal Incompetence
"The premise is that it is given that an error above some threshold value that is present in the relative position of two survey markers you report is incompetent practice."
NO! An error above the threshold is not incompetent practice.
Not being able to show and prove it was an error is incompetent practice.
Paul in PA
Error Does Not Equal Incompetence
> "The premise is that it is given that an error above some threshold value that is present in the relative position of two survey markers you report is incompetent practice."
>
> NO! An error above the threshold is not incompetent practice.
So, if your licensing board adopts a rule that says that you must meet a certain standard for minimum positional accuracy and you adopt methods that a reasonably prudent and competent surveyor would know could not be expected to meet the standard, your argument would be ... what? Are you arguing that it is merely negligent, not incompetent?
No state or forum member can answer your question but The Great State of Texas. Picking out other states for sarcasm is unseemly.
You can present your question for an answer here:
Please let us know how this turns out for you. It will be the only official answer to the question in your post. Until you have an answer there is no further basis for discussion here.
I'll throw out the idea that the acceptable level of confidence to be used in designing measurement processes to meet the specification is somewhere between 95% and 99%, but not less than 95%.
--
I think that's really good for GIS professionals. Most of us here are professional surveyors or surveyor interns, however, and we typically work at a much higher confidence level than the 95% swamp. Most of us would be out of business at that level. Inferring that some of us are incompetent or should be out of business will not be an acceptable response, as being totally unprofessional and against forum rules. Please don't go there.
The best text I know of for understanding surveying measurements is "Surveying Measurements and Their Analysis" by Dr. R. B. Buckner. Many of us here were trained by the book and/or the author. I assume from your interest in surveying measurements that you have a copy. If so we all have a basis for further discussion about surveying measurement theory.
Cheers!
> No state or forum member can answer your question but The Great State of Texas.
The Texas Board of Professional Land Surveying has adopted the standard in the form it was published. Until the rule is altered, the standard says what it says.
From the standpoint of professional land surveying, this isn't that difficult a problem. If the confidence level of relative positional certainty is to be 95%, then the standard errors of the N and E components of the positions of the two points in question must be less than 0.10/2.447 x 0.5 = 0.020 ft. If the confidence level is to be 99%, then the standard errors of the N and E components of the positions of the two points must be less than 0.10/3.035 x 0.5 = 0.016 ft.
In other words, you need to be positioning those points with uncertainties (standard errors) of less than +/-5mm or +/-6mm in N and E.
Anyone interested will find a relevant discussion on Page 33 of the 6th Ed. of "Surveying Theory and Practice", Davis, Foote, Anderson & Mikhail.
No, I Am Arguing That It Is An Error
It is mathematically impossible to eliminate all errors.
Unless you are saying, "Kent is infallible"?
Paul in PA
Kent,
The fallacy in your argument lies at its premise. Failure to meet the specification limits for measurement standards does not automatically equate to "incompetence."
If it were, then all surveyors would be incompetent 5% of the time when applying the 95% rule. Instead, that 5% is deemed as acceptably meeting the standard, not as incompetence.
Incompetence is when you are measuring between the wrong two points and can't tell the difference.
JBS
I like that!
Incompetence is when you are measuring between the wrong two points and can't tell the difference.
> The fallacy in your argument lies at its premise. Failure to meet the specification limits for measurement standards does not automatically equate to "incompetence."
In an isolated case, that might be argued. What is being discussed is a system of practice that is expected to deliver certain results. That is, if certain errors are quite likely with one method and quite unlikely with another, the issue of incompetent practice arises perfectly clearly.
> If it were, then all surveyors would be incompetent 5% of the time when applying the 95% rule. Instead, that 5% is deemed as acceptably meeting the standard, not as incompetence.
No, a practice that would be expected not to meet some standard is incompetent. If you're using a method with a 95% likelihood of meeting the standard, then you have chosen a method with a very high probability of success. On the other hand, if a surveyor chooses a method that has a very high probability of not meeting some standard, it is fair to say that his or her practices are incompetent. The only question is what level of confidence a professional surveyor would want to adopt.
I think the most defensible answer lies somewhere between 95% and 99% considering how widely those confidence levels are used for other important measurements.
No, I Am Arguing That It Is An Error
> It is mathematically impossible to eliminate all errors.
Hence the question: "What confidence level is appropriate?" My suggestion that a confidence level between 95% and 99% ought to be used has quite a bit of support in other allied fields as well as an established use in surveying.
A person competent with probability and statistics will understand that the slightest change in the requirement may lead to a change in the methodology required to achieve the requirement. So one must begin by finding the exact text of the applicable rule, and determine if it is a properly stated rule from a probability/statistical point of view. If not, file a complaint against the members of the state board of land surveyors for incompetence.
I’m going to regret this.
What Kent is pointing to is the fact that, in any jurisdiction that has a statutory positional tolerance standard, there must be an accepted level of practice that ensures that that level of positional accuracy is met. If the standard is not met, then, the practice can be considered incompetent.
In this case, I assume that Kent is defining incompetence as:
1. Not qualified in legal terms: a defendant who was incompetent to stand trial.
2. Inadequate for or unsuited to a particular purpose or application.
3. Devoid of those qualities requisite for effective conduct or action.
See incompetent at http://www.thefreedictionary.com/incompetent
The example Kent uses is the measurement of two points less than 100 feet apart must be reported with a positional tolerance of some relative accuracy plus an allowable distance, as in 1:10,000 + 0.10 feet (Texas Administrative Code, Title 22, Part 29, Chapter 663, Subchapter B, Rule §663.15(a)).
For example, two points that are actually 99.00 feet apart must be measured to an accuracy of 0.11 feet under this standard. (99/10,000 = 0.01, 0.10 + 0.01 = 0.11) Reporting the distance as anything between 99.11 feet and 98.89 feet meets the standards. Reporting a result of anything more or less than these values fails to meet the standard. Based on the definition above, this is incompetent in that the method used is not suited for the purpose.
The question that Kent raises is actually a legitimate one. My interpretation of his question is rather simple. Since we understand that no measurement is perfect, what standard of measurement tolerance do we apply to this on a practice level?
In other words, is it acceptable, in the example above, if Surveyor B reports 95 such measurements that meet the positional standard and 5 results that do not meet the standard, is Surveyor B guilty of failing to meet the standard or, at a confidence level of 95%, would this be considered and acceptable level of practice? Should the standard of practice require that only 1 measurement in 100 fail the positional tolerance standard (99% confidence level)?
Put another way, on a project where I make 100 measurements between points that are all 99.00 feet apart and report results which meet the standard 95 times but fail the standard 5 times, am deficient in meeting the standard? Would my peers accept that 5 of these measurements fail the standard or would they hold that I have failed to meet the standard? Would one "failure" in 99 be acceptable? Would my peers require that each reported measurement meet the tolerance standard?
It’s actually an interesting question.
Do we treat the group of measurements as a group or do we treat each measurement within the group as a discrete entity? Is the tolerance standard applied to the group of measurements or to the individual measurements?
I need to digest this for a bit.
Questions for Kent:
Have I properly understood your question in the first place?
Is your take on the standard such that it applies to the group of reported measurements or to each reported measurement?
> The question that Kent raises is actually a legitimate one. My interpretation of his question is rather simple. Since we understand that no measurement is perfect, what standard of measurement tolerance do we apply to this on a practice level?
> Questions for Kent:
>
> Have I properly understood your question in the first place?
Yes, I'd say that you've given a fair summary of the situation in this:
> The question that Kent raises is actually a legitimate one. My interpretation of his question is rather simple. Since we understand that no measurement is perfect, what standard of measurement tolerance do we apply to this on a practice level?
> Is your take on the standard such that it applies to the group of reported measurements or to each reported measurement?
I think that the most appropriate way to test compliance with the standard is on a methods and practices level. I wouldn't rule out some uncommon situation where damages result from a misreported distance between two incontestably stable monuments, though.
Surveyors who perform ALTA surveys have to meet a different standard, but one that is defined in such a way (using the length of the semi-major axis of a 95% confidence error ellipse as a measure) that methods and practices are implicitly the subject.
BTW since you have a good understanding of the question, you're in charge of answering the inevitable quibblers. I have to get into the field for a couple of days.
A related question is how could the standard be enforced. Looking at some guidance for control measurements, it would appear the typical pipe driven into Texas hardpan is fit only for monuments with a required horizontal accuracy of 1 meter. So unless the enforcement is done within a few days of the original measurement, the standard is unenforceable.
> A related question is how could the standard be enforced. Looking at some guidance for control measurements, it would appear the typical pipe driven into Texas hardpan is fit only for monuments with a required horizontal accuracy of 1 meter. So unless the enforcement is done within a few days of the original measurement, the standard is unenforceable.
In other words: how much can you get away with? You are supposed to be a professional. You should be enforcing it for your own measurements.
How is it enforced? Probably mainly if something goes to court and it comes out that you do not know your own precisions, and if it comes out that you did not meet the minimum standards.
But I get your point. A person can probably low-ball a lot of jobs, do them under minimum standards and never have to take responsibility for their actions. In fact I am positive that is done a lot.