Statistical Analysis

Abstract

The basic requirements of quality data are a high degree of reliability, minimum uncertainty, and preferably no bias and error. The basic parameters to assess the quality of data are accuracy and precision. Even a highly precise result may be inaccurate if the calibrating standards are deteriorated due to improper pH of the standard solution or otherwise due to the relative nature of most analytical techniques. The uncertainties propagate in the determinations causing error. The dispersion of data (scatter) is adjudged by the spread of a set of observations, i.e., range, its standard deviation (measure of the precision), coefficient of variation and variance. The error introduced due to independent (and random) sources in an analysis may be combined as the sum of the variances. Regression equations (minimum uncertainty equation for calculation) of the variables x and y are evaluated. Analytical figures of merit are reported with the desired confidence level using an appropriate number of significant figures. Accreditation and follow-up at regular prescribed intervals should be implemented subsequent to requisite measures of analytical quality assurance (AQA) and control (AQC) to ensure the quality of data generated.