2.5 Basic Analytical Properties
Limits of confidence are established at two different probability levels:
0.086
P = 0.10 (90% confidence), r = 2.02 32.10 ± 2.02· .g
0.086
P = 0.05 (95 % confidence), r = 2.57 32.10 ± 2.57· .g
32.10 ± 0.08%
32.10 ± 0.10%
65
Based on this criterion, the datum 31.96% (xq) will fall outside the limits at both P = 0.10 and
P = 0.05, so it should be rejected.
If the outlier is rejected and the remaining four data are used to recalculate statistics, then
the following resu lts are obtained:
X = 32.14% s = 0.034% n = 4
As can be seen, discarding the outlier results in a larger mean and a significantly smaller
relative standard deviation.
2.5.2 Sensitivity
This analytical property can be ascribed to analytical methods (processes) and
defined as "the ability to discriminate among similar concentrations of the analyte" or "the ability to detect (qualitative analysis) or determine (quantitative
analysis) small amounts of an analyte in a sample?' The more similar the analyte concentrations that can be distinguished or the lower those that can be
detected or determined, the higher the sensitivity.
Sensitivity is a basic property that sustains accuracy; in fact, no method can
be accurate unless it allows one to discriminate low concentrations in the detection or determination of an analyte. Mathematically, sensitivity can be expressed
via various parameters that are described below.
__ Box2.11
Let us illustrate a qualitative approach to sensitivity with a specific example: the task was to
determine the relative sensitivity levels of three different methods A, Band C that were used
to discriminate the total hydrocarbon index (THI) of water, a relevant parameter with a view
to assessing water pollution.
The three methods were individually applied to two standard samples of known concentrations: (1) 0.12 and (2) 0.11 )lg/L.
Method A: Neither sample contained hydrocarbons.
Method B: Both samples contained hydrocarbons and had an identical global index: 0. 1 )lg/L.
Method C: Both samples contained hydrocarbons but possessed a different global index:
(1) 0.125 and (2) 0.116 )lg/L.
Which method is the most reliable? No doubt, the sensitivity of method A is inadequate to
determine THI. Method B can only detect the presence of hydrocarbons, whereas method C
can additionally discriminate between samples of similar composition. Consequently, method
C possesses the highest sensitivity and method A the lowest. Also, method C will give slightly overestimated values.
Limits of confidence are established at two different probability levels:
0.086
P = 0.10 (90% confidence), r = 2.02 32.10 ± 2.02· .g
0.086
P = 0.05 (95 % confidence), r = 2.57 32.10 ± 2.57· .g
32.10 ± 0.08%
32.10 ± 0.10%
65
Based on this criterion, the datum 31.96% (xq) will fall outside the limits at both P = 0.10 and
P = 0.05, so it should be rejected.
If the outlier is rejected and the remaining four data are used to recalculate statistics, then
the following resu lts are obtained:
X = 32.14% s = 0.034% n = 4
As can be seen, discarding the outlier results in a larger mean and a significantly smaller
relative standard deviation.
2.5.2 Sensitivity
This analytical property can be ascribed to analytical methods (processes) and
defined as "the ability to discriminate among similar concentrations of the analyte" or "the ability to detect (qualitative analysis) or determine (quantitative
analysis) small amounts of an analyte in a sample?' The more similar the analyte concentrations that can be distinguished or the lower those that can be
detected or determined, the higher the sensitivity.
Sensitivity is a basic property that sustains accuracy; in fact, no method can
be accurate unless it allows one to discriminate low concentrations in the detection or determination of an analyte. Mathematically, sensitivity can be expressed
via various parameters that are described below.
__ Box2.11
Let us illustrate a qualitative approach to sensitivity with a specific example: the task was to
determine the relative sensitivity levels of three different methods A, Band C that were used
to discriminate the total hydrocarbon index (THI) of water, a relevant parameter with a view
to assessing water pollution.
The three methods were individually applied to two standard samples of known concentrations: (1) 0.12 and (2) 0.11 )lg/L.
Method A: Neither sample contained hydrocarbons.
Method B: Both samples contained hydrocarbons and had an identical global index: 0. 1 )lg/L.
Method C: Both samples contained hydrocarbons but possessed a different global index:
(1) 0.125 and (2) 0.116 )lg/L.
Which method is the most reliable? No doubt, the sensitivity of method A is inadequate to
determine THI. Method B can only detect the presence of hydrocarbons, whereas method C
can additionally discriminate between samples of similar composition. Consequently, method
C possesses the highest sensitivity and method A the lowest. Also, method C will give slightly overestimated values.