Rejection of results

There are a variety of statistical tests that have been used to decide if a data point should be rejected, as well as some rules of thumb. The range chosen to guide the decision will limit all of these tests and guidelines. A large range will retain possibly erroneous results, while a very small range will reject valid data points. It is important to note that the outlier must be either the highest 

One rule of thumb is that if the outlier value is greater than 4o from the mean, it should be rejected. When calculating the mean and standard deviation, an outlier result should not be included in the calculation. After the calculation, the suspected result should be examined to see if it is more than 4o from the mean. If it is outside this limit, it should be ignored if it is within this limit, the value for o should be recalculated with this result included in the calculation. It is not permissible to reject more than one result on this basis. A suspected result should not be included in calculating o. If it is included, it will automatically fall within 4a because such a calculation includes this number. Other reference sources recommend an even smaller range for rejection, such as a 2.5a limit. 

A statistical test called the g-test can be used effectively for small data sets (see the reference by Rorabacher) to determine if a given data point should be rejected. The 

---

There are two criteria which are commonly used to gauge the rejection of results. Of these, the most convenient to use is based on the x interval... 

There are two criteria which are commonly used to gauge the rejection of results. Of these, the most convenient to use is based on the jf 2believe this limit is too wide, and regard the Q-test at a 90% confidence level to be a more acceptable alternative. A rejection quotient Q is defined as... 

Determination of phenol by C NMR spectroscopy has the advantage that each determination affords three independent results that can be averaged, allowing rejection of results with too large RSD. The method was applied to the determination of phenol in tars of the cumene process, and were correlated with those of H NMR, UVV spectroscopy and titration with bromine. RSD for single results was 0.8% . ... 

In 1979, the ISM frequency allocations were revised as a result of the World Administrative Radio Conference (WARC) (16). A considerable effort was made to increase the number and worldwide uniformity of ISM frequency allocations. Most of those proposals were rejected. The resulting allocations are Hsted in Table 1. 

Risk-Based Inspection. Inspection programs developed using risk analysis methods are becoming increasingly popular (15,16) (see Hazard ANALYSIS AND RISK ASSESSMENT). In this approach, the frequency and type of in-service inspection (IS I) is determined by the probabiUstic risk assessment (PRA) of the inspection results. Here, the results might be a false acceptance of a part that will fail as well as the false rejection of a part that will not fail. Whether a plant or a consumer product, false acceptance of a defective part could lead to catastrophic failure and considerable cost. Also, the false rejection of parts may lead to unjustified, and sometimes exorbitant, costs of operation (2). Risk is defined as follows ... 

As a result, most systems are limited, to achieving a mole fraction reject of 0.1 or less (see Membrane technology). 

Risk and uncertainty associated with each venture should translate, ia theory, iato a minimum acceptable net return rate for that venture. Whereas this translation is often accompHshed implicitly by an experienced manager, any formal procedure suffers from the lack of an equation relating the NRR to risk, as well as the lack of suitable risk data. A weaker alternative is the selection of a minimum acceptable net return rate averaged for a class of proposed ventures. The needed database, from a collection of previous process ventures, consists of NPV, iavestment, venture life, inflation, process novelty, decision (acceptance or rejection), and result data. 

The situation changes when the system pressure becomes high enough for refrigerant to condense in the condenser and reject the resulting latent heat to the environment. Further addition of heat to the adsorbate desorbs more refrigerant which condenses in the condenser and trickles down into the receiver. The system pressure stays approximately constant as desorption and condensation proceed. 

The second part of the requirement deals with inspection and test records, which are also covered in clause 4.10.5. The difference between these requirements is that clause 4.10.1 requires you to document the records to be established (in other words define ) in the quality plan or procedures and clause 4.10.5 requires you to produce the records defined in the quality plan or procedures. Your inspection and test procedures therefore need to specify or contain the forms on which you intend to record the results of the inspections and tests performed. The details are covered later in this chapter, but there are two types of record to be considered the record that shows which inspections and tests have been performed and the record that shows the results of these inspections and tests. One may be a route card, shop traveler, or document which acts as both a plan of what to do and a record of the progress made and the other may be a table of results with specified parameters and accept/reject criteria. 

Thus, the error from ignoring the bend-twist coupiing terms is about 24%, certainly not a negligible error. Hence, the specially orthotropic laminated plate is an unacceptable approximation to a symmetric angle-ply laminated plate. Recognize, however, that Ashton s Rayleigh-Ritz results are also approximate because only a finite number of terms were used in the deflection approximation. Thus, a comparison of his results with an exact solution would lend more confidence to the rejection of the specially orthotropic laminated plate approximation. 

As RO membranes become looser their salt rejection falls (see Section 31.8.1). Eventually a point is reached at which there is no rejection of salts, but the membrane still rejects particulates, colloids and very large molecules. The membrane pore size can be tailored to a nominal molecular weight cut-off. The resulting filtering process is called ultra-filtration. 

In order to use any of the results obtained by objective methods as the basis for the acceptance or rejection of a product, there must be available reliable information as to the relationship between the values obtained and organoleptic quality in terms of consumer acceptance and utility. Where standards are based upon measurement of such labile constituents as ascorbic acid or sugar, a knowledge of the normal values for good commercial practice is necessary. Such values have not yet been adequately established. This should constitute a useful field for research of inestimable value to the industry. 

A limitation to the more widespread use of membrane separation processes is membrane fouling, as would be expected in the industrial application of such finely porous materials. Fouling results in a continuous decline in membrane penneation rate, an increased rejection of low molecular weight solutes and eventually blocking of flow channels. On start-up of a process, a reduction in membrane permeation rate to 30-10% of the pure water permeation rate after a few minutes of operation is common for ultrafiltration. Such a rapid decrease may be even more extreme for microfiltration. This is often followed by a more gradual... 

Laboratory catalyst testing is sometimes done under conditions that are far removed from exhaust gas conditions, and can be a very unreliable guide to the utility of a catalyst. For instance, noble metals may rank below base metal oxides in oxidation activity at low temperatures, but the ranking reverses at high temperatures. These and other hazards were pointed out by Schlatter et al. (53). Laboratory catalyst testing is usually done by the catalyst manufacturers, resulting in the rejection of a vast majority of formulations. 

Now that we have a model, we must check its consistency with various experiments. Sometimes such inconsistencies result in the complete rejection of a model. More often, they indicate that we need to refine the model. In the present case, the results of careful experiments show that the collision model of reactions is not complete, because the experimental rate constant is normally smaller than predicted by collision theory. We can improve the model by realizing that the relative direction in which the molecules are moving when they collide also might matter. That is, they need to be oriented a certain way relative to each other. For example, the results of experiments of the kind described in Box 13.2 have shown that, in the gas-phase reaction of chlorine atoms with HI molecules, HI + Cl — HC1 I, the Cl atom reacts with the HI molecule only if it approaches from a favorable direction (Fig. 13.28). A dependence on direction is called the steric requirement of the reaction. It is normally taken into account by introducing an empirical factor, P, called the steric factor, and changing Eq. 17 to... 

One important reason for the emergence of Paracelsian alchemical illustration may have been its function as a psychological compensation for the rejection of Catholic imagery by the Protestants and the resulting spiritual and emotional insecurity experienced by many of the faithful. It is not merely coincidental that Paracelsian iconography should appear in Protestant areas where traditional Catholic icons had been destroyed. (The development of a specihcally Catholic interest in alchemy, however, was a phenomenon of the mid-seventeenth century developed in the work of Athanasius Kircher and other Jesuits.) ... 

---