Numerical Methods of Data Analysis

The proper analysis of experimental data requires careful consideration of the numerical techniques used. Real data are subject to experimental error which can have an effect on results derived from the analysis. Often, this analysis involves fitting a curve to experimental data over the whole range or over part of the range in which experimental observations have been made. When thermodynamic data are involved, the relationship between the independent and dependent variable is usually not known. Then, arbitrary functions such as polynomials in the independent variable are often used in the data analysis. This type of data analysis requires consideration of the level of error in both variables, and of the effects of the error on derived results. 

In this appendix, methods of numerical analysis used for interpolation, integration, and differentiation of numerical data are discussed. Since these procedures involve curve fitting, regression analysis is also considered. 

Emphasis is placed on methods in which a minimum of assumptions are made regarding the functional relationship between the independent variable x and the dependent variable y. 

---

General information and useful definitions concerning statistical and numerical methods of data analysis may by found in the monograph of Btirgi and Dunitz28 and in earlier... 

---