Glossary Terms and Equations
|
| Accuracy
The degree of agreement between an observed value and an
accepted reference value. The concept of accuracy includes
both bias (systematic error) and precision (random error). |
| Bias A
persistent positive or negative deviation of the measured
value from the true value. In practice, it is expressed as
the difference between the value obtained from analysis of a
homogeneous sample and the accepted true value.
Bias = True value - Measured mean value |
| Box Plot A
graphical summary representation of the distribution of a
set of data, the top and bottom of the box representing the
25th and 75th percentile. The horizontal line represents the
median concentration, and the lower and upper Ts
extend to the 10th and 90th percentile concentrations. |
| Control Chart A graphical plot of test
results with respect to time or sequence of measurement,
together with limits within which they are expected to lie
when the system is in a state of statistical control
(Taylor, 1987). |
| Critical Concentration A
calculated concentration used to determine whether the
measured bias is statistically significant (Anderson, 1987).
critical concentration =
t * ssp *
where: 
ssp = pooled standard deviation
s1 = standard deviation of reference solution measurements
s2 = standard deviation of daily QCS measurements
n = number of values
t = t
statistic at the 95% confidence level and
(n1 +
n2)
!2 degrees of freedom |
| External Blind Sample A Quality
Assurance sample of known analyte concentrations submitted
to the laboratory by an external agency. These samples
arrive at the CAL as normal weekly rain samples and undergo
routine processing and analysis. The identity of the sample
is unknown to the CAL until all analyses are complete. Data
are used to assess contamination potential from handling and
shipping. |
| Internal Blind Sample A Quality
Assurance sample of known analyte concentrations submitted
to the laboratory by the QA Specialist. The identity of the
sample is known to the processing staff only. The analyte
concentrations are unknown to the analysts. These data are
valuable in assessing bias and precision for network
samples. |
Mean
The average obtained by dividing a sum by the number of its
addends.
where:
n = number of
values
xi = values
|
| Mean Bias The sum of the bias for
each sample divided by the total number of replicates (n). |
| Mean Percent Recovery The sum of
the percent recovery for each sample divided by the number
of replicates (n). |
| Method Detection Limit MDL
The minimum concentration of an analyte that can be reported
with 99 percent confidence that the value is greater than
zero (Glaser et al., 1981). |
| Paired Data t Test
where:
d =
the difference in each pair of values
= the average difference in the pairs of values. The
absolute value is generally used.
n =
the number of pairs of values
df = degrees of freedom
associated with t
sd
= standard deviation of the differences between the pairs
|
| Percent Bias The difference
between the mean value obtained by repeated analysis of a
homogeneous sample and the accepted true value expressed as
a percentage of the true value.
%Bias = 100 * [(Vm
! Vt
)/Vt
]
where:
Vm = mean measured value
Vt = true value |
| Precision The degree of agreement
of repeated measurements of a homogeneous sample by a
specific procedure, expressed in terms of dispersion of the
values obtained about the mean value. It is often reported
as the sample standard deviation (s). |
| Quality Assessment The system of
procedures that ensures that QC practices are achieving the
desired goal in terms of data quality. Included is a
continuous evaluation of analytical performance data. |
| Quality Assurance QA
An integrated system of activities involving planning, QC,
reporting, and remedial action to ensure that a product or
service meets defined standards of quality. |
| Quality Control QC
The system of procedures designed to eliminate analytical
error. These procedures determine potential sources of
sample contamination and monitor analytical procedures to
produce data within prescribed tolerance limits. |
| Quality Control Solution
QCS A solution containing known concentrations of
analytes used by the analysts to verify calibration curves
and validate sample data. The values obtained from the
analyses of these samples are used for calculation of bias
and precision and for the monthly control charts. |
| Relative Standard Deviation RSD The standard deviation expressed as a percentage.
RSD = 100 * (s/)
where:
s = sample
standard deviation
= mean value |
| Replicates
(Splits) Two aliquots of the same sample treated
identically throughout the laboratory analytical procedure.
Analyses of laboratory replicates are beneficial when
assessing precision associated with laboratory procedures
but not with collection and handling. |
| Sensitivity
The method signal response per unit of analyte. |
| Standard
Deviations The number representing the dispersion
of values around their mean.
where:
n =
number of values |
|
Standard Deviation Estimated from paired Measurements
The standard deviation may be estimated from the differences
of several sets of paired measurements using the equation
(Taylor, 1987):
where:
d
= difference of duplicate measurements
k
= number of sets of duplicate measurements |
| Variance
s2
The best measure of the dispersion of repeated results (precision)
or the measure of the dispersion of a series of results
around their average.
(Anderson, 1987).
where:
 ,
the difference between value 1 and 2 of pair
i
n = the number of pairs of data |
|
Web page was updated
01/20/2010
|
