692 P.J. Mart ́ın-Alvarez ́
0123456
Concentration
0.00
0.05
0.10
0.15
0.20
Relative area
Relative Area = .00589 + .030635*Concentration
r = .980
Limit of
detection = 0.9
Fig. 13.3Plot of the observed relative area versus the concentration of analyte in the standard
solution, the estimated calibration line with its 95% confidence interval, and a possible way of
calculating the limit of detection
table with thelack of fit testobtained with the STATGRAPHICS program (Statis-
tical Graphics Corp., http://www.statgraphics.com/)) using theCalibration Models
procedure in theSpecial Advanced Regressionmodule (the model appears to be
adequate for the observed data, becauseP>0.10). Figure 13.3 displays plots of the
observed relative area to the internal standard versus the concentration of analyte
in the standard solution, the estimated calibration line with its 95% confidence
interval, and a possible way of calculating thedetection limit(the upper value of
the 95% confidence interval for the estimatedXconcentration corresponding to the
upper value of the 95% confidence interval for the true mean value ofYatX= 0
(Sharaf et al. 1986). The calibration range for the analyte and the results of recovery
experiments to evaluate the occurrence of proportional systematic errors are often
included in the calibration results.
13.3 Multivariate Statistical Techniques
13.3.1 Introduction
Multivariate analysis can be defined as the set ofstatistical and mathematical meth-
odsto analyse, describe and interpretmultidimensional observations. Application