Section B - Assessment of data

All measurement processes are subject to measurement errors that affect numerical data and which arise from a variety of sources. 

An absolute error is the numerical difference between a measured value and a true or accepted value. A relative error is the absolute error divided by the true or accepted value. 

Also known as systematic errors, or bias, these generally arise from determinate or identifiable sources causing measured values to differ from a true or accepted value. 

Also known as random errors, these arise from a variety of uncontrolled sources and cause small random variations in a measured quantity when 

Where several different measurements are combined to compute an overall analytical result, the errors associated with each individual measurement contribute to a total or accumulated error. 

---

Ba(C5Me5)2 has good solubility in aromatic hydrocarbons and ethers and is even slightly soluble in alkanes. Although decreases in sublimation temperatures have been achieved by introduction of sterically demanding substituents on the Cp ring and would portend widespread use of these compounds as precursors in MOCVD (see Section IV,B), it is difficult to quantitatively assess these data in that respect because data on rates of sublimation are lacking in the literature. 

In summarizing the intrinsic viscosity relations presented in this section, it must be admitted that they represent nothing more than rather small semi-empirical refinements of the Flory excluded volume theory and the Flory-Fox viscosity theory. For a large fraction of the existing body of experimental data, they offer merely a slight improvement in curve-fitting. But for polymers in good solvents it is believed that a more transcendental result has been achieved. The new equations permit more reliable assessment of unperturbed chain dimensions in such cases, and in some instances (e. g., certain cellulose derivatives see Section III B) they offer possible explanations of heretofore paradoxical solution behavior. 

A calculated isothermal section for the Si-B-C system at T = 1400 K was presented by Goujard et al. (1994) [75, 76]. The data were derived from JANAF [140] and assessments of the phases SiBi4, SiBg, SiB3 and the solid solution phase boron carbide. The isothermal section calculated using the data of Kasper (1996) [33] (Fig. 10b) deviates significantly. Experimental investigations at these temperatures suffer from slow reaction kinetics and decisions... 

The procedure adopted to portray the scope and utility of a linear free-energy relationship for aromatic substitution involves first a determination of the p-values for the reactions. These parameters are evaluated by plotting the values of log (k/ka) for a series of substituted benzenes against the values based on the solvolysis studies (Section IV, B). The resultant slope of the line is p, the reaction constant. The procedure is then reversed to assess the reliability and validity of the Extended Selectivity Treatment. In this approach the log ( K/ H) observations for a single substituent are plotted against p for a variety of reactions. This method assays the linear or non-linear response of each substituent to variations in the selectivity of the reagents and conditions. Unfortunately, insufficient data are available to allow the assignment of p for many reactions. It is more practical in these cases to adopt the Selectivity Factor S as a substitute for p and revert to the more empirical Selectivity Treatment for an examination of the behavior of the substituents. 

---