How does it work?

Statistics can be applied in a number of scenarios:

Sampling – Sampling involves the selection of a part of some population or process so that inferences can be made about the whole.

Experimental design – Properly designed experiments with formally stated goals can save an organisation thousands of pounds relative to unstructured experimentation in which the outcome is frequently indeterminate.

Modelling - A description of the assumed structure of a set of observations that can range from a fairly imprecise verbal account to, more usually, a formalised mathematical expression of the process assumed to have generated the observed data.

Decision analysis - Decision analysis involves the structuring of a decision problem, e.g. by decision trees and influence diagrams, so that dependencies can be formally analysed.

Time series/longitudinal data - Frequently organisations collect data regularly over extended time periods. Such series may show trends, seasonality and more or less complex forms of internal dependence.

Calibration/assay - In many enterprises such as engineering and other laboratory-based industries, testing products or comparing samples / components is an essential part of the operation. Designing such tests is critical to ensure that appropriate precision and compatibility are achieved.

Quality control - Taguchi methods and Six-Sigma are concepts that are frequently bandied about, but what are the ideas behind them? One danger with the ‘branding’ of specific statistical ideas is that the method becomes paramount without an understanding of their origins or limitations.

Reliability - Reliability focuses on the ability of a product to perform its intended function without failure for a specified period of time under stated conditions.

Image Processing - Digital image processing and the analysis of information from images are methods that have become increasingly important in scientific research (particularly biological sciences) and many technical fields