Evaluation, calibration exercise

It is essential to identify and separate these two types of errors to avoid confusion. If numerical errors are not isolated, they may lead to undesirable spurious model calibration exercises. It is, therefore, necessary to devise systematic methods to quantify numerical errors. The basic idea behind error analysis is to obtain a quantitative measure of numerical errors, to devise corrective measures to ensure that numerical errors are within tolerable limits and the results obtained are almost independent of numerical parameters. Having established adequate control of numerical errors, the simulated results may be compared with experimental data to evaluate errors in physical modeling. The latter process is called model validation. Several examples of model validation are discussed in Chapters 10 to 14. In this section, some comments on error analysis are made. 

As palaeoceanographic reconstructions based on bulk fossil chemistry proliferate, comparability of measurements between laboratories has become an important issue. A recent calibration exercise involving thirteen laboratories by Rosenthal et al. (2004) evaluated the reproducibility of analyses of synthetic standard solutions within and between laboratories. The study found that intralaboratory instrumental precisions were generally better than 0.5% for both Mg/Ca and Sr/Ca measurements, but interlahoratory precisions (r.s.d) were significantly worse (up to 3.4% and 1.8%, respectively). This could be a result of differences between cahbration standards used by laboratories, or because the circulated standard solutions had become contaminated in the interval between their... 

Writing or reviewing articles on NIRS involves researching the literature. To communicate the results of a NIRS exercise effectively up to 16 pieces of information should be included. These are summarized in Table 7.1.1. Very few papers contain all of this information, or even enough for a full evaluation of the efficiency of the calibration exercise. Items 11,14, and 15 (RPD) can give a good approximate evaluation of a calibration. Minimum acceptable values for the RPD and RER are 3.0 and 10.0, respectively. Some of the papers that are cited in Tables 7.1.4 and 7.1.5 have been found to have RPD values as low as 1.05. If the RPD is 1.0 this means that the SEP is equal to the SD of the reference data of the validation samples set (whether cross-validation or a test set is used), and the calibration model is of no practical value in analysis. 

The variation of sensitivity between different sensors was also checked. Calibration curves with five different sensors were performed. A Relative Standard Deviation of 13, 13 and 42% of calibration slopes (sensitivity) were obtained for Cu, Pb and Cd respectively. These variations should have limited consequence on bias and precision when the standard addition method is used. However, for Cd, variations in the limit of quantification between two electrodes could be expected. Finally, the accuracy of the method was evaluated by the measurement of a SWIFT reference material used during the 2nd SWIFT-WFD Proficiency Testing exercise (Table 4.2.2). The reference value was chosen as the consensus value of the selected data population obtained after excluding the outliers. The performances of the device were estimated according to the Z-score (Z) calculation. Based on this score, results obtained with the SPEs/PalmSens method were consistent with those obtained by all methods for Pb and Cu ( Z < 2) while the result was less satisfactory for Cd (2 < Z < 3). 

A method performance study may also be used to evaluate the performance of a new analytical method or an automated form of an already existing method [1,19]. Under certain circumstances a method performance study may be called intercalibration study. This should only concern exercises where the calibration of methods is investigated e.g. studies on calibrants, calibrant solutions, calibration procedures etc. Such studies are more frequent in physical measurements for metrological purposes. 

Students have to do a series of short exercises in practical sessions, using the application developed, shown in Fig. 1, so that they can familiarize themselves with the graphical interface, study component errors, evaluate the mechanism performance, analyze the influence of modifying design parameters and analyze the improvement in mechanism accuracy after calibration. To do this, a first practical session introduces this application. 

Care has to be exercised when planning the instrumentation and analysis of either a scale model test or a full scale test. It should be ensured that adequate and correctly calibrated instrumentation and test devices are provided so that the test results may be documented and evaluated in order to verily the test results. At the same time, it is necessary to ensure that the instrumentation, test devices and electrical connections do not interfere with the model in a way that would invalidate the test results. 

Pfannkoch et al. [11] examined the utility of a number of small molecules as test probes for HPSEC using the requirement that a molecule must have minimal interaction with the support. The dipeptide glycyltyrosine functioned adequately when aqueous mobile phases with neutral pH were used. An inexpensive and stable protein that is available in high purity, such as ovalbumin, is generally well suited for macromolecular plate height determinations. It should be noted that plate height is also a function of operational parameters such as mobile phase velocity, viscosity, and temperature therefore, care must be exercised to perform evaluations under identical conditions. The simple evaluations described earlier will not yield any information about pore volume or slope of the calibration curve. 

---