Analysts avoidance

The analyst can then pursue an experimental plan to measure characteristics of the material that may have been affected by the proposed cause(s). It is vitally important that the failure analyst avoid pursuing a proof of a presupposed cause. Instead, the data obtained from the measurements should be reviewed carefully to look for clues that are incompatible with the presupposed cause and may point to other potential causes. An example will help clarify this point. 

This method, due to Middleton (Analyst, 1935, 6o, 154), has the advantage over Lassaigne s method (pp. 321-326) that the use of metallic sodium is avoided it has the disadvantage, however, that the reagents are not so readily obtained pure, and the method requires rather more time. 

ETA breaks down an accident iato its contributing equipment failures and human errors (70). The method therefore is a reverse-thinking technique, ie, the analyst begias with an accident or undesirable event that is to be avoided and identifies the immediate cause of that event. Each of the immediate causes is examined ia turn until the analyst has identified the basic causes of each event. The fault tree is a diagram that displays the logical iaterrelationships between these basic causes and the accident. 

The critical role of analysts introduces a potential for bias that overrides all others—the an ysts evaluation of the plant information. Analysts must recognize that the operators methods, designers models, and control engineers models have merit but must so beware they can be misleading. If the analysts are not familiar with the unit, the explanations are seductive, particiilarly since there is the motivation to avoid antagonizing the operators and other engineers. 

Some of the problem areas mentioned are sometimes overblown by many analysts. That is, they sometimes overemphasize the importance of a particular behavioral characteristic. That characteristic might be important only in one small regime of structural response, and you must know that limitation on the validity of the characteristic. The designer s job, on the other hand, is to either avoid all those problem areas or to in some way overcome them. The situation is somewhat like having a mountain in front of you, and you must get to the other side. You either climb over that mountain, in which case you definitely recognize that it is there and solve the problem, or go around it, in which case you have simply avoided the mountain. In both cases, you must recognize that the mountain exists in order to properly deal with it. 

Select mobile phases for HPSEC based on their ability to dissolve the sample and their compatibility with the column. Zorbax PSM columns are compatible with a wide variety of organic and aqueous mobile phases (Table 3.4), but analysts should avoid aqueous mobile phases with a pH greater than 8.5. As mentioned earlier, select mobile phases that minimize adsorption between samples and silica-based packings. Sample elution from the column after the permeation volume indicates that adsorption has occurred. If adsorption is observed or suspected, select a mobile phase that will be more strongly adsorbed onto the silica surface than the sample. For example, N,N-dimethyl-formamide (DMF) is often used for polyurethanes and polyacrylonitrile because it eliminates adsorption and dissolves the polymers. When aqueous mobile phases are required, highly polar macromolecules such as Carbowax can be used to coat the silica surface and eliminate adsorption. Table 3.5 provides a list of recommended mobile-phase conditions for some common polymers. 

Some GPC analysts use totally excluded, rather than totally permeated, flow markers to make flow rate corrections. Most of the previously mentioned requirements for totally permeated flow marker selection still are requirements for a totally excluded flow marker. Coelution effects can often be avoided in this approach. It must be pointed out that species eluting at the excluded volume of a column set are not immune to adsorption problems and may even have variability issues arising from viscosity effects of these necessarily higher molecular weight species from the column. 

Most microprocessor-based systems permit a maximum of 256 APSs per database. This limit could be restrictive if the analyst wishes to establish a unique APS for each machine-train. To avoid this problem, APSs should be established for classes of machine-trains. For example, a group of bridle gearboxes that are identical in both design and application should share the same APS. 

To avoid unnecessary time and expenditure, an analyst needs some guide to the number of repetitive determinations needed to obtain a suitably reliable result from the determinations performed. The larger the number the greater the reliability, but at the same time after a certain number of determinations any improvement in precision and accuracy is very small. 

Independence of analysts and analyses in one organization is a fundamental question. It is important to have, even for the most reliable methods, more than one analyst/laboratory involved to avoid possible analyst/laboratory-specific biases. Certification by a single laboratory, without confirmation by another laboratory or method is risky. Measurement by a single definitive method is usually performed by two or more analysts working independently to minimize possible biases. Frequently, an accurately characterized back-up method is employed to corroborate the data. Some agencies feel that a certification campaign should not be based on a single measirrement procedure and therefore do not normally certify values on the basis of a definitive method applied in one laboratory. 

The concern of the independence of analysts in one organization still remains a fundamental question. Again it is important to have more than one analyst/laboratory involved to avoid possible analyst-specific and laboratory-specific biases. The characterization should be corroborated by additional methods or laboratories to provide additional assurance that the data are correct. 

One of the primary requirements for methods is that it be practicable [Section 512(b)(1)(G)]. A method that cannot be used in Federal laboratories has no value in the protection of the food supply. Method developers should avoid the use of rare or custom-made equipment, prohibitively expensive equipment, untested technologies, or reagents that are not commercially available. For a determinative procedure, an analysis should not exceed two working days, and methods should have a minimum sample throughput of at least six samples per analyst-day. 

Principles and Characteristics A sample can contain a great number of compounds, but analysts are usually interested only in the qualitative presence (and the quantitative amount) of a small number of the total compounds. Selectivity is an important parameter in analytical separations. The total analytical process clearly benefits from selectivity enhancement arising from appropriate sample preparation strategies. Selective separation of groups or compound classes can simplify a mixture of analytes before analysis, which in turn enhances analytical precision and sensitivity. Selective fractionation, in some cases, allows easier resolution of the compounds of interest, so analysts can avoid the extreme conditions of high-resolution columns. 

Organic materials, Sulfuric acid Analytical Methods Committee, Analyst, 1976, 101, 62-66 Advantages and potential hazards in the use of mixtures of 50% hydrogen peroxide solution and cone, sulfuric acid to destroy various types of organic materials prior to analysis are discussed in detail. The method is appreciably safer than those using perchloric and/or nitric acids, but the use of an adequate proportion of sulfuric acid with a minimum of peroxide is necessary to avoid the risk of explosive decomposition. The method is not suitable for use in pressure-digestion vessels (PTFE lined steel bombs), in which an explosion occurred at 80° C. 

In general, limit tests are quantitative or semi-quantitative tests particularly put forward to identify and control invariably small quantities of impurity that are supposed to be present in a pharmaceutical substance. Obviously the amount of any single impurity present in an official substance is usually small, and therefore, the normal visible-reaction-response to any test for that impurity is also quite small. Hence, it is necessary and important to design the individual test in such a manner so as to avoid possible errors in the hands of various analysts. It may be achieved by taking into consideration the following three cardinal factors, namely ... 

In order to be sure of good quality analytical data, the analyst is recommended to construct a Levich plot such as that shown in Figure 7.5 before starting a series of analyses analysts will avoid those rotation speeds represented on the graph as turbulent flow. In the jargon of the subject, the analyst may say that a laminar regime is maintained. 

In order to ensure that the solution flow is indeed laminar past a tubular electrode, we choose a solution of known and fixed concentration, and measure /iim as a function of Vf. A plot of log(/um) as y against log(Vf) as x should be linear with a gradient of 1/3. Deviation from this line at fast flow rates indicates those flow rates that should be avoided. The analyst now knows the range of flow rates to keep within. If the solution flow is not completely laminar, then the flow rate is amended until it is - amend in this context almost always means decrease o). ... 

The SHE is experimentally cumbersome, and is hazardous to use owing to the involvement of elemental hydrogen gas, while the values of h+,H2 can fluctuate quite badly during operation because of the cyclic nature of bubble formation. For these reasons, the SHE is avoided experimentally unless a secondary reference electrode requires calibration. We will not consider the SHE any further because it is so unlikely that an analyst would in fact wish to calibrate a new reference electrode. 

Figure 4 shows peak area precision vs. injection volume for a typical autosampler. Note that excellent peak area precision of 0.2% RSD was readily achievable for an injection volume >5 J,L. Precision levels are much poorer (0.5-1% RSD) for sampling volumes <5 J,L, attributable to the finite resolution of the sampling syringe and associated digital stepper motor. To obtain optimum peak area precision, the analyst must avoid potential problem situations such as an overly fast sample... 

The analyst should avoid creating impossible factor combinations. This occurs, for instance, when choosing both the batch number and the manufacturer of the capillary as factors in a robustness test by means of a two-level design. It is impossible to define two unique batch numbers that exist for both manufacturers. The way to examine both factors is by using nested designs. ... 

Most CZE separations are very sensitive to conductivity (e.g., salt concentration) in the run buffer. Therefore, to avoid introducing a high amount of salt from the sample injection, samples should be buffer exchanged with an appropriate low salt buffer prior to analysis. Centrifugal UF/DF devices are ideal for this purpose, as they are typically very reproducible and allow the analyst greater flexibility in controlling the final sample concentration. 

The research that we conducted, the Evergreen Project, was designed to avoid these problems. It was the largest study of its kind ever undertaken. We analyzed over 60,000 pages of information from 200 firms in multiple industries. In all, ten years of data were collected. The firms that were studied varied in size, and we complemented these broad extensive analyses with focused, in-depth exploration of issues of special interest. Both researchers and practitioners were involved. Fourteen prominent academics from Dartmouth, Harvard, Wharton, and other leading business schools were involved. Twenty-one practitioners implemented the study in coordination with the academic researchers. We interviewed journalists, executives, and Wall Street analysts. We let the answers come from the data, and then tested specific hypotheses to explore promising possibilities. And finally, we invented new statistical methodology that is appropriate to this level of complexity. 

---