Why do workshops and seminars on subjects such as accuracy and precision or errors get less response from the profession, as measured by attendance, than subjects such as legal descriptions or boundaries? I'll admit that the premise of my question is based on personal observation and not hard, scientific evidence. But based on the apparent response by surveyors to measurements and analysis topics, it seems that my anecdotal measurements of this phenomenon are not too far off and not likely to bother a majority of the profession. If someone out there has hard, scientific evidence to prove (or disprove) my premise, I'd like to know. I am willing to be surprised and am equally willing to change my mind.
Rather than this column being a reflection on surveyors' interest (or non-interest) in workshops on surveying measurement errors, I'd like to focus on my observations about surveyors' measurement techniques-that is, the techniques I've heard about in my seminars when I've asked various questions about accuracy, precision and errors. Many of my workshops pertain to instrumentation with the hope that by understanding one's instrumentation, a surveyor can better use the measurement tools, and even more importantly, understand and analyze the measurements made with that tool.
During the course of some of these seminars, I ask my audience members about their understanding of survey instrument accuracy and errors. My observations and recommendations to these inquiries follow.
Precision and accuracy - About half of the surveyors in an average group will be able to distinguish between these two terms. I do get some verbal or visual clue that they seem to. Do you wonder about the remaining half? I do.
Errors - Seldom will I get an indication that the audience generally understands that there are two types of errors, though most appear to readily relate to the terms "random" and "systematic." The responses indicating knowledge of the sources of errors-environmental, technological and human-are usually quite a bit higher.
Standard deviation - Almost everyone in a given audience will have heard this term, but very few are able to explain what it is as a statistical concept or how it applies to surveying. Additionally, most will not know how to calculate it or how to use it if it is used in conjunction with a measurement.
Experiment to obtain empirical data - I don't use this expression directly, but the reaction of my audiences will often come down to a lack of understanding on how to design a process that isolates a part of the measurement portion, and then collect and analyze data to evaluate the accuracy or precision of that process.
PPM - Audience members have difficulty verbalizing their understanding of this acronym. Some will tell me that it stands for "parts per million" and stop there. Others will tell me that it represents the error (or the correction used) in EDM measurements contributed by the temperature and pressure of the atmosphere. I rarely find that they understand PPM as a dimensionless expression that can be equivalently stated as a precision ratio or some other type of fraction, which can be applied to many types of measurements to indicate accuracy or precision.
I realize that expert measurement is only part of what a surveyor does, but regardless of the sub-specialty within surveying-boundary, topographic, construction, engineering, hydrographic and so on-it is a significant part of delivering the professional service. So my recommendations in regard to these areas are written in the form of wishes.
Wish #1 - Surveyors read, understand and integrate into their thinking (and work) the published accuracy of their instrumentation, including accessories such as optical plummets, and prism and range poles. This includes the people who operate and use them. Wish #2 - Surveyors always strive for accuracy, which means that they strive to understand which systematic errors are applicable to the quality of results they wish to obtain, that they apply corrections to their observations and that they devise methods to monitor random error, actively encouraging the normal occurrence of random error by having truly independent observations. This, by the way, means redundant measurements (precision) so that by being precise and removing the pertinent systematic errors the desired accuracy will be achieved. Wish #3 - Surveyors estimate errors before doing a survey so as to estimate the resulting compounded error. This way, if the estimated resultant is unsatisfactory, they can revise their methods to achieve desired results. Wish #4 - Surveyors observe the results they get from their surveys, attempt to quantify the performance of their surveying systems (i.e., accuracy delivered by the particular combinations of instruments, accessories and operators), and confirm their suspicions periodically by collecting data for data's sake to have an empirical conclusion to their system accuracies. Wish #5 - Surveyors periodically (no more than two years apart and more frequently under severe conditions) have a competent service shop inspect, clean and adjust their instrumentation (including accessories). At much more frequent intervals, I wish that surveyors check the adjustment of their instrumentation (including accessories) and perform calibration processes to evaluate instrument accuracies.
The holiday season is past us, so if you'd like to adopt my wishes, perhaps you'll convert some of them into resolutions for the coming year and the rest of your career. See you at the next seminar.