Accuracy and. Precision

The accuracy of an analytical method may be defined as the extent to which its results coincide with the truth. If the mean of a large number of measurements closely coincides with the truth, the method may be said to be accurate. Accuracy in this sense is really freedom from bias. Bias is primarily a characteristic of the procedure, although it may be influenced by the operator. 

The precision of a method is the level of agreement that can be obtained between replicate results. It is a measure of the ability of the method to discriminate between similar samples. For example, if in the hands of a particular operator a method could give replicates with a spread of say 0.3%, it would be able to discriminate between samples differing by about that amount, the exact level of discrimination depending on the number of replicate measurements. There is an inherent limiting value for the discriminating power of every method, upon which no analyst, however skilled, can improve, but the skill of the operator is the main factor. There can be no doubt that a careful, skilled operator can consistently achieve better precision than a careless, unskilled one. 

for example, a method which in the hands of a skilled operator gave replicate results scattered over a range of say 2% relative, i.e. 2% of the measured value or one part in fifty, but with a mean lying very close to the true value, would be accurate but not precise. A method whose results scattered over a range of 0.1% but with a mean 2% too low would be precise but not accurate. A method capable of giving results with a low degree of scatter and with a mean within say 0.2% of the truth would be both accurate and precise. 

ISO 5725 [5] defines precision as the closeness of agreement between mutually independent test results obtained under stipulated conditions , and identifies two aspects of it, which it terms repeatability and reproducibility . Repeatability is superficially similar to what is here called precision, being a measure of the agreement between replicate results obtained by a single operator on the same sample, using the same apparatus in the same laboratory over a short period of time. But this concept seems to be based on the assumption that all operators are equal, and they are manifestly not. Reproducibility is a measure of the level of agreement between the results obtained by different operators on the 

In the author s opinion, the most important single factor influencing the precision of a method is the amount of care used in carrying it out. 

Accuracy and precision, precision and accuracy. .. same thing, right Chemists everywhere gasp in horror, reflexively clutching their pocket protectors — accuracy and precision are different  

Accuracy Accuracy describes how closely a measurement approaches an actual, true value. 

1 Precision Precision, which we discuss more in the next section, describes how close repeated measurements are to one another, regardless of how close those measurements are to the actual value. The bigger the difference between the largest and smallest values of a repeated measurement, the less precision you have. 

The two most common measurements related to accuracy are error and percent error  

Error Error measures accuracy, the difference between a measured value and the actual value  

The accuracy and precision of the analytical methods were determined by the average and standard deviation of individual method recoveries of the fortitied-control samples in 50 different matrices (see Tables 1 and 2). These methods were also demonstrated to be very rugged based on the results of accuracy and precision for a variety of crop and animal matrices. 

The extraction efficiencies using a blender and a shaker were compared and both methods gave similar results. A corn sample treated with radiolabeled carfentrazone-ethyl and collected from a metabolism study was used for comparison. Multiple samples can be extracted simultaneously if extraction is performed by shaking. In addition, since the extraction procedures in the residue study closely followed the extraction scheme in the metabolism study, the resulting extraction efficiencies from both studies were almost identical. 

During the initial partition with hexane and water, the aqueous pH must not exceed 8. Carfentrazone-ethyl is extremely unstable under alkaline conditions and will rapidly degrade to C-Cl-PAc. At times, the workup of the crop samples, including the fortification step, should be completely separated for carfentrazone-ethyl and the acid metabolites, to avoid any possible interference from the parent compound. 

Pyridine and BF3 in methanol are hazardous and must be used only in a well-ventilated hood. A solvent partition after acylation helps remove residual pyridine from the sample. Material Safety Data Sheets for the derivatizing agents should be reviewed and kept readily available. 

The injection standards of carfentrazone-ethyl must be in acetonitrile. Other solvents (e.g., ethyl acetate) lead to poor chromatography following injection of matrix samples. This can lead to apparent enhanced recoveries of analyte in the fortified samples. 

The levels of accuracy and precision determine the quality of a measurement. The data are as good as random numbers if these parameters are not specified. Accuracy is determined by analyzing samples of known concentration (evaluation samples) and comparing the measured values to the known. Standard reference materials are available from regulatory agencies and commercial vendors. A standard of known concentration may also be made up in the laboratory to serve as an evaluation sample. 

Effective use of evaluation samples depends on matching the standards with the real-world samples, especially in terms of their matrix. Take the example of extraction of pesticides from fish liver. In a real sample, the pesticide is embedded in the liver cells (intracellular matter). If the calibration standards are made by spiking livers, it is possible that the pesticides will be absorbed on the outside of the cells (extracellular). The extraction of 

Precision is measured by making replicate measurements. As mentioned before, it is known to be a function of concentration and should be determined at the concentration level of interest. The intrasample variance can be determined by splitting a sample into several subsamples and carrying out the sample preparation/analysis under identical conditions to obtain a measure of RSD. For example, several aliquots of homogenized fish liver can be processed through the same extraction and analytical procedure, and the RSD computed. The intersample variance can be measured by analyzing several samples from the same source. For example, different fish from the same pond can be analyzed to estimate the intersample RSD. 

The precision of the overall process is often determined by the extraction step rather than the analytical step. It is easier to get high-precision analytical results it is much more difficult to get reproducible extractions. For example, it is possible to run replicate chromatographic runs (GC or HPLC) with an RSD between 1 and 3%. However, several EPA-approved methods accept extraction efficiencies anywhere between 70 and 120%. This range alone represents variability as high as 75%. Consequently, in complex analytical methods that involve several preparative steps, the major contributor to variability is the sample preparation. 

The accuracy of interference microscopy depends on the calibration standards used. For measurements in the secondary wall, these values are 1.604 for Nb and a range of values for Na depending on the specimen. The variation in Na arises presumably from variation in microfibril angle (Hermans 1946). The need to measure Na for each sample is one disadvantage of this technique that increases the amount of work necessary to perform the measurement. If Na is not measured for each sample, an error of 11% is introduced compared to the usual error of 2% (Donaldson 1985a). Na does not vary between earlywood 

Accuracy of measurements in the middle lamella depends not only on Nb, but also on the value of Nd for this region. Donaldson (1985a) measured the refractive index of the middle lamella in differentiating tracheids before lignifica-tion and in primary xylem, where the middle lamella is unlignified, and obtained 

It is worth discussing some aspects of precision that relate to the way in which data are processed. Typically, m.e.p. is estimated as the average of ten measurements on a single site. These values often deviate from a normal distribution and should not be used as replicates (sample sizes of 100 or more are normally distributed). To obtain replicates, independent determinations of o.p.d. are made on different sections, using the mean value in each case. 

Barer R (1953) Determination of dry mass, thickness, solid and water concentration in living cells Nature 172 1098 

Donaldson LA (1987) S3 lignin concentration in radiata pine tracheids Wood Sci Technol 21 227-234 

In any form of analysis, accuracy and precision are required otherwise, the analytical data are suspect and cannot be used with any degree of certainly. This is especially true of analytical data used for commercial operations where the material is sold on the basis of purity. Being a complex material, one may wonder about the purity of coal, but in this sense the term purity refers to the occurrence (or lack thereof) of foreign constituents within the organic coal matrix. Such foreign constituents (impurities) are water, pyrite, and mineral matter. Therefore, at this point, it is advisable to note the differences inherent in the terms accuracy and precision. 

The word accuracy is used to indicate the reliability of a measurement or an observation, but it is, more specifically, a measure of the closeness of agreement between an experimental result and the true value. Thus, the accuracy of a test method is the degree of agreement of individual test results with an accepted reference value. 

Accuracy is often expressed inversely in terms of the standard deviation or variance and includes any systematic error or bias. Accuracy includes both the random error of precision and any systematic error. The effect of systematic error on the standard deviation is to inflate it. In the measurement of coal quality for commercial purposes, accuracy expressed in this manner is generally of less interest than is systematic error itself. When systematic error is reduced to a magnitude that is not of practical importance, accuracy and precision can become meaningful parameters for defining truly representative sampling and for interpretation of the results of various test methods. 

Estimation of the limits of accuracy (deviation from a true or theoretical value) is not ordinarily attempted in coal analysis. Precision, on the other hand, is determined by means of cooperative test programs. Both repeatability, the precision with which a test can be repeated in the same laboratory, usually but not always by the same analyst using the same equipment and following the prescribed method(s), and reproducibility, the precision expected of results from different laboratories, are determined. Values quoted in test methods are the differences between two results that should be exceeded in only 5 out of 100 pairs of results, equal to 2-Jl times the standard deviation of a large population of results. 

The specification of repeatability and reproducibility intervals, without specification of a statistical confidence level, weakens the precision and accuracy 

Which estimate of goodness is more important Well, both of them are important. It is important to you and your patient that the reading on a pulse oximeter accurately represents the patient s blood Por Likewise, you won t have confidence in a pulse oximeter if it gives a different reading every time you look at it (assuming the patient s condition has not changed). 

The accuracy of a measurement can be improved by making replicate measurements and taking the average. All measurements have some inherent uncertainty. But if the uncertainties are random, then half of the measurements should be too big and half too small. The errors will cancel each other out. Accuracy is assessed by calculating the percent error, which is given by the formula  

Notice the quotation marks around true value. A method is verified by testing it against known standards, and in this case, you do know the true value, because you prepared the system. However, in most cases, you will never know what the true value is. The patient certainly has some value for serum cholesterol level, but you don t know what that is. 

The smaller the ratio of the standard deviation to the average value, the better the precision. However, there is no absolute standard that defines good precision or bad precision. In general, you should become increasingly suspicious of measurements as the ratio of standard deviation to the average value increases. For an experienced analytical chemist, working with a well-known system, this ratio will typically be less than 0.1%, but this level of precision is unlikely in a clinical setting. 

For example, imagine measuring the (arterial) PQ2 of a patient three times and obtaining these values 105, 96.0, and 102 mmHg (in that order). What is the average value Does the data seem precise What is the standard deviation of the measurements What is the percent error  

It is very important to understand the definitions of accuracy and precision and to recognize the difference between precision and accuracy. Accuracy is a measure of how close a measured analytical result is to the true answer. For most analytical work, the true answer is not usually known. 

It is important for students to realize that the inability to obtain the correct answer does not necessarily mean that the analyst uses poor laboratory techniques or is a poor chemist. Many canses 

The quantitative analysis of any particular sample should generate results that are precise and accurate. The results should be reproducible, reliable, and truly representative of the sample. Unfortunately, some degree of error is always involved in analytical determinations, as discussed in Section 1.3.3. 

The lab information presented in this book is certainly no substitute for the actual lab work, but it can serve to refresh your memory of those labs you have done and give you a basic idea of any labs you may not have had time to complete. 

With most infrared spectrometers, it is possible to determine absorbance with a precision of 0.5 to 1.0%. However, the accuracy of the quantitative determination can vary from 1 to 10%, depending on the system analyzed [41], 

Students frequently have misconceptions about the definitions of accuracy and precision based on popular uses of the words. The term accuracy refers to how close a measured value is to the true value of a quantity. In other words, it describes how close you got. Precision describes how close your measurements are to each other, not how close they are to the actual value. That is, if your values are close together, they are considered precise, even if they are nowhere near the true value (accurate). 

The distinction between accuracy and precision becomes extremely important when you consider real-world measurements. For example, if you try to measure the distance between two objects using a meter stick, the accuracy of your answer is limited. The smallest markings on the meter stick are millimeters. Therefore, you can definitely know the exact number of millimeters, and you can even estimate one more place after the decimal. 

However, your value can be no more accurate than that as long as you are using a meter stick. It is possible that you may obtain a more sophisticated device, like a sonic range-finder, that can provide you with a more accurate measurement. The important thing to remember is that your data can be no more accurate than the device with which you are measuring. This concept is the foundation for the use of significant figures in calculations. 

At this point, we digress slightly to make some observations about the accuracy and precision of experimental data. Since we, as engineers, continuously make use of data that represent measurements of various 

Name Symbol Formula Notation Significance Application 

Archimedes number /VAr a, PtfifApd3 A/Ar - Pf = fluid density Ap = solid density — fluid density (Buoyant x inertial)/ (viscous) forces Settling particles, fluidization 

Bingham number Nm N - T°° Moo / r0 = yield stress = limiting viscosity (Yield/viscous) stresses Flow of Bingham plastics 

Bond number number Nbo N -APd29 /vBo — T o- = surface tension (Gravity/surface tension) forces Rise or fall of drops or bubbles 

International collaborative studies have also been carried out to investigate the accuracy and precision of such bioassays when applied to complex mixtures (e.g., Claxton et al.., 1992a, 1992b Lewtas et al., 1992 May et al., 1992). The conclusions from one of the major studies (Claxton et al., 1992b) are representative of such intercomparisons  

The performance of measurement systems has been traditionally defined in terms of accuracy and precision. Accuracy can be defined as a measure of how close a result is to the actual value and precision is thought of as the uncertainty of the result, which we could identify with the standard uncertainty. Modem usage in the context of quality of analytical results tends to avoid these terms. This is because there has been a more fundamental appreciation of the actual measurement process. For example, accuracy or, perhaps we should say, inaccuracy, involves bias within a measurement process as well as statistically determined factors that cause the result to be different from the true result What, at one time, we would have blithely termed precision is now discussed as repeatability, the variability of a method when applied to measurements on a single sample within a laboratory, and reproducibility, which applies to measurements of that sample when appUed by different laboratories using different instruments operated by different operators. 

The lUPAC recommendations on this matter have been published by Currie (1995). On a day-to-day basis, there is little harm in applying the traditional usage. However, when producing formal documentation I would recommend that the lUPAC usage should be adopted. 

Accuracy is a concept that encompasses getting the answer right (sometimes known as trueness) with acceptable uncertainty (i.e., with good precision). The relationship between accuracy and precision is shown in figure 1.6 where high precision is represented by the closeness of the cluster of hits on a target and high accuracy is 

We have discussed uncertainty of a measurement result in terms of a possible spread of values. In the RACI competition it appears that 

Note that the pdf is a function of. v—the values that can be taken by the data. A probability density function is defined in terms of its area the probability of finding a result between two values of. v (say x and x2) is the area under the pdf between x and. v2. The shape of this pdf is the familiar bell-shaped curve shown overlaying the histogram of figure 1.3. 

We shall see in the next chapter that this analysis implies knowledge of /x and a, and these statistics are not always available, and are certainly not the same as the mean and sample standard deviation of a small data set. 

Assuming infinite mass accuracy, we should be able to identify the molecular formula of any ion merely on the basis of its exact mass - the emphasis is on infinite mass accuracy (Chap. 3.5.1). In reality we are dealing with errors in the order of one to several ppm depending on the type of instrument and the mode of its operation. 

Unequivocal formula assignment by accurate mass alone only works in a range up to about m/z 500 depending on the particular restrictions [41]. Obviously, for ions of larger m/z the number of hits rapidly increases beyond a reasonable limit. Even at a high mass accuracy of 1 ppm and with the particular case of peptides the elemental composition can only be unambiguously identified up to about 800 u 

The situation becomes more complicated as more elements and fewer limitations of their number must be taken into account. In practice, one must try to restrict oneself to certain elements and a maximum and/or minimum number of certain isotopes to assure a high degree of confidence in the assignment of formulas. Isotopic patterns provide a prime source of such additional information. Combining the information from accurate mass data and experimental peak intensities with calculated isotopic patterns allows to significantly reduce the number of potential elemental compositions of a particular ion [46,47]. 

A repeated assay on the same sample yields a range of values that can be expressed as the mean, x, +/- the standard deviation cr. For a precise assay a is small compared with the mean (x) and it is desirable that the j/x (the % error) should be less that 5%. If the scatter is small then the reproducibility is good and the assay has good resolution (the ability to distinguish between two close values). Elements that affect precision are random errors such as pipetting, weighing etc. 

An accurate assay is one where the measured value of x (x) is close to the true value, X, i.e. X - x is small If we are measuring a reaction spectrophoto-metrically to an end point there are a number of possible systematic errors. For example, the reaction may not go to completion if a concentration of reagent is too low, or the time allowed, while appropriate for a high concentration of analyte, is too short at lower concentrations. A trivial, hut occasionally important source of inaccuracy can occur by working away firom the wavelength of maximum absorbance (X ). This is particularly applicable to compounds vwth narrow absorption bands and is further exacerbated by use of wide slit widths on the spectrophotometer (see Chapter 1). 

In several experiments accuracy may be less important than precision. For example, if during the purification of an enzyme a protein assay is used based on albumin as a standard, the colour change observed for 1 mg of albumin may be different to that observed with 1 mg of the enzyme under study. However, as long as the error is systematic and not random, the method can still be used to monitor the increasing specific activity of the enzyme preparation at different stages of the purification. 

In most cases, however, accuracy is critical. In measuring the stoichiometries of ligand binding to proteins it is essential to know the true molecular weight and concentration of the protein. We shall therefore now discuss a number of practical aspects that should allow the absolute calibration of assays for macromolecules, particularly proteins and DNA 

Assay volumes can be scaled down to a final volume of 1 ml or below if semi-microcells (0.2-0.4 cm wide 1-cm pathlength) are available. However, it is important in this case to use a narrow slit width for the light beam, and in some cases it may be necessary to blacken the sides of the cuvette by insulation tape or even by using a permanent marker pen. Semi-microcuvettes with blackened sides can be purchased directly from manufacturers (e.g. Hellma). 

2% according to the table footnote, so the determination by Analyst 2 is not accurate. Analyst 3 is both inaccurate and imprecise. It is very unlikely that an imprecise determination will be accurate. Precision is required for accuracy, but does not guarantee accuracy. 

It is important for students to realize that the inability to obtain the correct answer does not necessarily mean that the analyst uses poor laboratory techniques or is a poor chemist. Many causes contribute to poor accuracy and precision, some of which we will discuss in this chapter as well as in later chapters. Careful documentation of analytical procedures, instrument operating conditions, calculations, and final results are crucial in helping the analyst recognize and ehminate errors in analysis. 

The selection of entries for inclusion in a particular study on the basis of structural precision depends on the purpose of the study. In cases where structural changes are likely to be several times larger than the experimental uncertainties, then precision criteria should not be restrictive. If, however, we are looking for very small geometrical variations, as in structure-reactivity correlations, then only the most precise structures are satisfactory. Indeed, the degree of experimental precision and the degree of precision with which the chemical fragment should be defined (see Section 3.7.4) are interrelated criteria in structure correlation work. 

It is extremely difficult to estimate exactly the accuracy of the determination since the major contribution to the systematic error probably has its source in the sampling procedure itself (see Chapter 1). When attention is paid to all the sources of systematic errors (see Section 4.3), most of which result in an increased oxygen content, a field precision of 0.005 mL/L can be achieved using 100 mL samples and photometric endpoint detection and 0.03 mL with 50 mL samples and visual (starch) endpoint detection. The precision is about 25 % less for oxygen contents below 2 mL/L. If a good quality iodate standard is used for calibration the analytical accuracy is equal to the precision. 

Grasshoff, K. (1981), in Marine Electrochemistry Whitfield, M., Jagner, D., (Eds.). Chichester, John Wiley Sons, 1981 pp.. 327-420. 

Millero, EJ., Sohn, M.L. (1992), Chemical Oceanography. Boca Raton CRC Press. 

---

The raw data collected during the experiment are then analyzed. Frequently the data must be reduced or transformed to a more readily analyzable form. A statistical treatment of the data is used to evaluate the accuracy and precision of the analysis and to validate the procedure. These results are compared with the criteria established during the design of the experiment, and then the design is reconsidered, additional experimental trials are run, or a solution to the problem is proposed. When a solution is proposed, the results are subject to an external evaluation that may result in a new problem and the beginning of a new analytical cycle. 

Analytical chemists converse using terminology that conveys specific meaning to other analytical chemists. To discuss and learn analytical chemistry you must first understand its language. You are probably already familiar with some analytical terms, such as "accuracy and "precision, but you may not have placed them in their appropriate analytical context. Other terms, such as "analyte and "matrix, may be less familiar. This chapter introduces many important terms routinely used by analytical chemists. Becoming comfortable with these terms will make the material in the chapters that follow easier to read and understand. 

The successful application of an external standardization or the method of standard additions, depends on the analyst s ability to handle samples and standards repro-ducibly. When a procedure cannot be controlled to the extent that all samples and standards are treated equally, the accuracy and precision of the standardization may suffer. For example, if an analyte is present in a volatile solvent, its concentration will increase if some solvent is lost to evaporation. Suppose that you have a sample and a standard with identical concentrations of analyte and identical signals. If both experience the same loss of solvent their concentrations of analyte and signals will continue to be identical. In effect, we can ignore changes in concentration due to evaporation provided that the samples and standards experience an equivalent loss of solvent. If an identical standard and sample experience different losses of solvent. 

Caro s acid, H2SO5 is used as a titrant for determining Fe +. Directions are given for exploring the method of end point detection, the titrant s shelf-life, the method s accuracy and precision, and the susceptibility of the method to interference from other species. 

The accuracy and precision of FIA are comparable to that obtained by conventional methods of analysis. The precision of a flow injection analysis is influenced by variables that are not encountered in conventional methods, including the stability of the flow rate and the reproducibility of the sample s injection. In addition, results from FIA may be more susceptible to temperature variations. These variables, therefore, must be carefully controlled. 

A final component of a quality control program is the certification of an analyst s competence to perform the analysis for which he or she is responsible. Before an analyst is allowed to perform a new analytical method, he or she may be required to successfully analyze an independent check sample with acceptable accuracy and precision. The check sample should be similar in composition to samples that the analyst will routinely encounter, with a concentration that is 5 to 50 times that of the method s detection limit. 

Accurate, precise isotope ratio measurements are important in a wide variety of applications, including dating, examination of environmental samples, and studies on drug metabolism. The degree of accuracy and precision required necessitates the use of special isotope mass spectrometers, which mostly use thermal ionization or inductively coupled plasma ionization, often together with multiple ion collectors. 

An isotope ratio is frequently measured 10 to 20 times for each sample to obtain high accuracy and precision. 

Accurate, precise isotope ratio measurements are used in a variety of applications including dating of artifacts or rocks, studies on drug metabolism, and investigations of environmental issues. Special mass spectrometers are needed for such accuracy and precision. 

Compressed Tablets. This popular type of dosage form offers convenience, stabiUty, accuracy and precision, and good bioavadabihty of active ingredients. After the best formulation has been estabflshed, compressed tablets can be manufactured at high rates of speed on advanced equipment. Tablets can be made to achieve rapid dmg release or to produce delayed, repeat, or prolonged therapeutic action (Controlled release technology, pharmaceutical). ... 

Test Methods. In addition to that provided by proper sampling and rephcation of analysis, a test method also has a significant impact on the accuracy and precision of the results. Preferred methods are those which are accepted in the chemical industry such as those from the American Society of Testing Materials (ASTM), Association of Official Analytical Chemists (AO AC), or from compendia such as the United States Pharmacopoeia (USP) or the Pood Chemical Codex (FCC) (36). The use of such methods eliminates the need for method vahdation. 

Method Transfer. Method transfer involves the implementation of a method developed at another laboratory. Typically the method is prepared in an analytical R D department and then transferred to quahty control at the plant. Method transfer demonstrates that the test method, as mn at the plant, provides results equivalent to that reported in R D. A vaUdated method containing documentation eases the transfer process by providing the recipient lab with detailed method instmctions, accuracy and precision, limits of detection, quantitation, and linearity. 

To solve a flow problem or characterize a given fluid, an instmment must be carefully selected. Many commercial viscometers are available with a variety of geometries for wide viscosity ranges and shear rates (10,21,49). Rarely is it necessary to constmct an instmment. However, in choosing a commercial viscometer a number of criteria must be considered. Of great importance is the nature of the material to be tested, its viscosity, its elasticity, the temperature dependence of its viscosity, and other variables. The degree of accuracy and precision required, and whether the measurements are for quaUty control or research, must be considered. The viscometer must be matched to the materials and processes of interest otherwise, the results may be misleading. 

Determination of accuracy and precision should be made by analysis of repHcate sets of analyte samples of known concentration from equivalent matrix. At least three concentrations representing the entire range of the caUbration should be studied one near the minimum (MAQ), one near the middle, and one near the upper limit of the standard curve. 

BeryUium aUoys ate usuaUy analyzed by optical emission or atomic absorption spectrophotometry. Low voltage spark emission spectrometry is used for the analysis of most copper-beryUium aUoys. Spectral interferences, other inter-element effects, metaUurgical effects, and sample inhomogeneity can degrade accuracy and precision and must be considered when constmcting a method (17). 

Graphite is frequently, although incorrectly, analyzed by the proximate method used for coal in which the volatile material is deterrnined by strongly beating the sample in a covered or luted cmcible. Some oxidation of the graphite always occurs so that the value obtained for volatile matter is high and thus the "fixed carbon" is too low. The method lacks both accuracy and precision. 

Known samples should also be run to verify the accuracy and precision of the routine methods to be used during the unit test. Poor quality will manifest itself as poor precision, measurements inconsistent with plant experience or laboratory history, and disagreement among methods. Plotting of laboratory analysis trends wiU help to determine whether calibrations are drifting with time or changing significantly. Repeated laboratory analyses will establish the confidence that can be placed in the results. 

The presence of errors within the underlying database fudher degrades the accuracy and precision of the parameter e.stimate. If the database contains bias, this will translate into bias in the parameter estimates. In the flash example referenced above, including reasonable database uncertainty in the phase equilibria increases me 95 percent confidence interval to 14. As the database uncertainty increases, the uncertainty in the resultant parameter estimate increases as shown by the trend line represented in Fig. 30-24. Failure to account for the database uncertainty results in poor extrapolations to other operating conditions. 

We can conclude that transmittance near infrared based methodology provides an accuracy and precision agree with those required by CIPAC for the pesticide analysis in commercially available formulations. 

Accuracy and precision of the methods were checked by the added-found method and statistic treatment of the data of determinations (RSD ranged from 0.025 to 0.046). 

The sensitivity, accuracy, and precision of solid-sample analysis have been greatly improved by coupling LA with ICP-OES-MS. The ablated species are transported by means of a carrier gas (usually argon) into the plasma torch. Further atomization, excitation, and ionization of the ablated species in the stationary hot plasma result in a dramatic increase in the sensitivity of the detection of radiation (LA-ICP-OES) or of the detection of ions (LA-ICP-MS). 

The chromatography literature contains a vast amount of dispersion data for all types of chromatography and, in particular, much of the data pertains directly to GC and LC. Unfortunately, almost all the data is unsuitable for validating one particular dispersion equation as opposed to another. There are a number of reasons for this firstly, the necessary supporting data (e.g., diffusivity data for the solutes in the solvents employed as the mobile phase, accurate distribution and/or capacity factor constants (k")) are not available secondly, the accuracy and precision of much of the data are inadequate, largely due to the use of inappropriate apparatus with high extracolumn dispersion. 

Katz et fl/.[l] searched the literature for data that could be used to identify the pertinent dispersion equation for a packed column in liquid chromatography. As a result of the search, no data was found that had been measured with the necessary accuracy and precision and under the sufficiently diverse solute/mobile phase conditions required to meet the second criteria given above. It became obvious that a... 

---