Geoscience Reference
In-Depth Information
Fig. 5.8
Histogram of the
absolute differences of duplicate
sample pairs
Fig. 5.9
Scatter plot of the
original vs. the duplicate samples
pulp and coarse duplicates; the differences between the
expected values of the standard and the control samples;
histograms and basic statistics of the percentages of devi-
ation of pulp and coarse duplicates; etc., see, for example,
Fig.
5.8
.
• A scatterplot of the original versus duplicate samples
(pulp and coarse) along with correlation statistics. This
plot provides a visual and numerical analysis of the cor-
relation of the two variables, see Fig.
5.9
.
• Control graphs for standards versus time including lines
representing the expected value of the standard and the
upper and lower limits. These graphs are also useful for
duplicate samples because it may help identify periods
when the quality of the laboratory work is poor. It is often
seen that the laboratory work performed immediately
after holidays or long weekends is of lower quality. The
quality may also be poor when the laboratory is working
at capacity or overloaded.
• Another time-dependent control chart, for both standards
and duplicates, is to plot a moving average of results
that would include 20 to 40 control samples at a time,
i.e., several batches. This control chart would be useful
to detect longer-term trends, which are sometimes dis-
guised or difficult to detect when looking at individual
batches.
• Another useful graph is the cumulative frequency of the
absolute value of Eq. 5.1, see Fig.
5.10
. It shows if the
control samples comply with the prescribed 90 % of the
