Geosynthetics April/May 2020 - 13

ACCURACY, PRECISION AND BIAS
Q: I am going through a few items for the
accreditation and there has always been one
thing that I have never been too sure of in the
standard when it talks about precision and bias.
I am never sure how to use this information
and have always simply calculated the
coefficient of variation and if that falls below
the 10%, then results are acceptable. I am
beginning to doubt myself and wanted to
check whether you see it the same way or
if there is a different way.
A: First, a few definitions are in order. Accuracy
refers to the closeness of a measured value to
a standard or known value. Precision refers to
the closeness of two or more measurements
to each other. Using the following example, if
you weigh a given substance five times, and
get 7.1 pounds (3.2 kg) each time, then your
measure is very precise. However, if one was
actually weighing a 6.6-pound (3.0-kg) mass,
you are not very accurate. Both accuracy and
precision are determined in a single lab and
generally are characterized numerically by the
term "repeatability." Repeatability equals the
standard deviation divided by the average
written as a percent for a single data set in
one given lab.
Bias on the other hand is a systematic
distortion of the relationship between
a treatment, risk factor, or exposure and
laboratory outcomes. Several types of bias can
be distinguished: information, equipment,
personnel, selection, confounding, etc. These
types of bias and their potential solutions are
discussed at great length in ASTM E9. Bias is
determined in multiple labs and generally
are characterized numerically by the term
"reproducibility." Reproducibility equals the
standard deviation divided by the average
written as a percent for multiple data sets
by several laboratories.
One needs both precision and bias
(repeatability and reproducibility) to calculate
the ultimate goal: Uncertainty. Unfortunately,
there is not a right way (i.e., one) to calculate
uncertainty. GAI-LAP has its favorite way, as
shown in Equation 1.

U = √ r s2 + R s2
U = uncertainty
rs = repeatability or precision
Rs = reproducibility or bias
Equation 1

To further complicate matters, one can calculate
uncertainty based on equipment, personnel,
time, location and material variability. GAI-LAP
annually calculates uncertainty based on its
internal numbers (Sr=repeatability) and the
external number from the GAI-LAP proficiency
tests (SR=reproducibility). A summary paper
on this subject is attached.
Now for my public service announcement (i.e.,
why are we requiring all of this?). Knowledge
of testing result uncertainty is fundamentally
important for laboratories; their clients and
stakeholders use these results for comparative
purposes. Uncertainty of measurement is
a requirement of ISO 17025 and a key to
mitigating risks, improving quality and reducing
costs. Despite the established requirements
of a test standard or norm, no measurement
is exact. No matter how carefully the result is
obtained, every measurement result contains
an independent amount of uncertainty.
Therefore, if measurement is important, then
measurement uncertainty is equally important.
According to both ASTM and ISO, no
measurement is complete without an
accompanied statement of the associated
amount of uncertainty. Creating awareness for
the importance of measurement uncertainty
is the key to ensuring that the geosynthetic
industry has a focus on measurement quality.

WELDING ROD

THE GEOSYNTHETIC
INSTITUTE ANSWERS
YOUR QUESTIONS
gmatechline@ifai.com
View past questions and answers at
https://tinyurl.com/GMATechline. Email
gmatechline@ifai.com anytime for fast,
free, direct technical answers.

Q: This question is in regard to the welding
rod intended for welding of geomembranes
(both high-density polyethylene [HDPE] and
linear low-density polyethylene [LLDPE]).
Are there any GRI GM13- or GM17-based
requirements for welding rods or are there any
white papers or specifications? I am mainly
interested in the parameters "Carbon Black
Content" and the "Carbon Black Dispersion."
Shall the welding rod also comply with the
GM13 and GM17 specification for these two
parameters (2%-3% content and the 1, 2
or 3 categories for the dispersion)?
A: You are correct, "adequate extrudate
welding rods or pellets, of the same
composition as the geomembrane, must be
used" when welding polyethylene. This is
clearly written in the EPA Technical Guidance
Document 30SW-91-051. See attached
for details. It is also written into the BAM
(Germany) regulations and most construction
quality assurance plans. We think that it is
an important consideration and is often
unchecked or overlooked on projects.
www.GeosyntheticsMagazine.com

0420GS_cv2-15.indd 13

13

3/23/20 6:20 PM
https://www.tinyurl.com/GMATechline http://www.GeosyntheticsMagazine.com
Geosynthetics April/May 2020

Table of Contents for the Digital Edition of Geosynthetics April/May 2020
