ABSTRACT
The key factor in any research that uses new equipment is the repeatability and accuracy of this equipment. In most performance analysis papers, the researchers are presenting systems that have been specifically designed for that experiment. It was the exception (Hughes et al. 1989; Wilson and Barnes 1998), rather than the rule that most papers presenting new systems produce evidence of systematized testing of the reliability of these new systems. In a number of studies, parametric techniques were used with data that were non-parametric, although, in some cases, the means of the data sets appeared ordinal, they were often means of nominal data and therefore the use of a parametric test put the conclusions at risk.
