Judgment errors

Much more difficult to detect are data interpretation and judgment errors, such as the unrecognized false positive and false negative results and the incorrect interpretation of mass spectra and chromatographic patterns. The detection and correction of these errors is made possible through internal review by experienced analysts. 

BALs of 80-100 mg/dl denote legal intoxication in most states. BALs of 80-200 mg/dl show coordination problems and judgment errors, labile mood, and deterioration in cognition. 

Judgmental error (i.e., estimated value based on professional opinion). 

As discussed in Chapter 9, there is also "conscious competence." Sometimes poor judgment is used to intentionally take a risk. I often make judgment errors and take calculated risks on the tennis court. Sometimes, I rush the net when I should not or stroke the ball long... 

Judgment had to be exercised in data selection. For each fluid, all available data were first fit simultaneously and second, in groups of authors. Data that were obviously very old, data that were obviously in error, and data that were inconsistent with the rest of the data, were removed. 

Analysts should review the technical basis for uncertainties in the measurements. They should develop judgments for the uncertainties based on the plant experience and statistical interpretation of plant measurements. The most difficult aspect of establishing the measurement errors is estabhshing that the measurements are representative of what they purport to oe. Internal reactor CSTR conditions are rarely the same as the effluent flow. Thermocouples in catalyst beds may be representative of near-waU instead of bulk conditions. Heat leakage around thermowells results in lower than actual temperature measurements. 

Recommendation When all measurements were recorded by hand, operators and engineers could use their judgment concerning their validity. Now with most acqmred automatically in enormous numbers, the measurements need to be examined automatically. The goal continues to be to detect correctly the presence or absence of gross errors and isolate which measurements contain those errors. Each of the tests has limitations. The hterature indicates that the measurement test or a composite test where measurements are sequentially added to the measurement set are the most powerful, but their success is limited. If automatic analysis is required, the composite measurement test is the most direct to isolation-specific measurements with gross error. 

Now you can reconsider the material balance equations by adding those additional factors identified in the previous step. If necessary, estimates of unaccountable losses will have to be calculated. Note that, in the case of a relatively simple manufacturing plant, preparation of a preliminary material-balance system and its refinement (Steps 14 and 15) can usefully be combined. For more-complex P2 assessments, however, two separate steps are likely to be more appropriate. An important rule to remember is that the inputs should ideally equal the outputs - but in practice this will rarely be the case. Some judgment will be required to determine what level of accuracy is acceptable, and we should have an idea as to what the unlikely sources of errors are (e.g., evaporative losses from outside holding ponds may be a materials loss we cannot accurately account for). In the case of high concentrations of hazardous wastes, accurate measurements are needed to develop cost-effective waste-reduction options. It is possible that the material balance for a number of unit operations will need to be repeated. Again, continue to review, refine, and, where necessary, expand your database. The compilation of accurate and comprehensive data is essential for a successful P2 audit and subsequent waste-reduction action plan. Remember - you can t reduce what you don t know is therel... 

The paired comparisons method (NUREG/CR-3688) is a structured expert judgment method in which human errors are compared in pairs. By combining the judgments of the group of experts, the errors arc arranged in order of likelihood of occurrence, of the human errors considered, they can be used as "anchor points list. Documentation requirements are given in Table 4.5-8. 

Human reliability [lata NJUREG/CR-1278 was supplemented by judgment of system analysts and plant personnel. Human error probabilities were developed from NUREG/CR-12 8, human action time windows from system analysis and some recovery limes from analysis of plant specific experience. Data sources were WASH-1400 HEPs,Fullwood and Gilbert assessment ot I S power reactor Bxp., NUREG/ CR -127K. and selected acro ptice li.it.j... 

The uncertainties in human error rates may be within the stated uncertainty bounds, but such is not demonstrated from sparse experiments. Both the qualitative description of the human interaction logic and the quantitative assessment of those actions rely on the virtually untested judgment of experts. 

Embry, D. E. et al., SLIM-MAUD an Approach to Assessing Human Error Probabihues Using Structured Expert Judgment, BNL, March 1984. 

It should be clear that a complete FMEA approach is not practical for the evaluation of production facility safety systems. This is because (1) the cost of failure is not as great as for nuclear power plants or rockets, for which this technology has proven useful (2) production facility design projects cannot support the engineering cost and lead time associated with such analysis (3) regulatory bodies are not staffed to be able to critically analyze the output of an FMEA for errors in subjective judgment and most importantly, (4) there are similarities to the design of all production facilities that have allowed industry to develop a modified FME.A approach that can satisfy all these objections. 

Embrey, D. E., Kirwan, B., Rea, K., Humphreys, P., Rosa, E. A. (1984). SLIM-MAUD. An approach to Assessing Human Error Probabilities Using Structured Expert judgment Vols. I and II. Washington, DC NUREG/CR—3518 US Nuclear Regulatory Commission. 

Number and type of record The number of data points or tables of data presented in the resource or the number of events the data set reflects where available, the form in which the data are presented, such as failure rates or availability data, confidence intervals or error factors the raw data source used, sueh as surveys, plant records, tests, or judgment. 

Among tlie phenomena tliat lead to accidents, human error is tlie most unpredictable. Tliis section describes accidents due to errors of judgment. Even well trained people occasionally make such errors, and one must eitlier accept an occasional mistake or cliange tlie work situation to ininiinize or remove tlie opportunities for error. 

The material in this book vvas prepared in good faith and carefully reviewed and edited. The author and publisher, however, cannot be held liable for errors of any sort in these chapters. Furthermore, because the author has no means of checking the reliability of some of the data presented in the public literature, but can only examine it for suitability for the intended purpose herein, this information cannot be w arranted. Also because the author cannot vouch for the experience or technical capability of the user of the information and the suitability of the information for the user s purpose, the use of the contents must be at the best judgment of the user. 

It should be noted that there is some variation in reported chemical shifts for particular compounds in the literature, as would be expected. Usually, these variations are less than 2ppm, and they can usually be attributed to concentration and solvent effects (as well as simple experimental error ). When given a choice, data reported using CDC13 as solvent will be preferred, with chemical shifts being reported to the nearest parts per million (except occasionally when comparisons within a series from a common study are reported). When multiple values have been reported in the literature, the author will use his judgment regarding choice of the value to use in the book. 

Selection of the form of an empirical model requires judgment as well as some skill in recognizing how response patterns match possible algebraic functions. Optimization methods can help in the selection of the model structure as well as in the estimation of the unknown coefficients. If you can specify a quantitative criterion that defines what best represents the data, then the model can be improved by adjusting its form to improve the value of the criterion. The best model presumably exhibits the least error between actual data and the predicted response in some sense. 

---