Applying Anova Simulation

Aplying Anova Simulation

Applying ANOVA simulation

RES/342

Sept 26, 2011

Jim Patton

Applying ANOVA simulation

Praxidike Systems is a software company that is having difficulty determining why their projects are not being completed or delivered on time. The scenario begins by offering two non-parametric testing methods which include ANOVA and Kruskal-Wallis. Non-parametric testing methods have certain requirements to be function successfully. The major hypothesis of ANOVA testing consists of the normal population distribution, errors are independent, and each population has the same variance. The Kruskal-Wallis test does not require the assumption of a normal distribution and the data must be on an ordinal scale. This is usually a better choice if all the assumptions of ANOVA cannot be met.

A brief description of the scenarios and lessons learned from ANOVA and non-parametric testing

The scenario provides three examples of how these testing methods can be applied to real world situations. During the first phase of the scenario the choice applied was the Kruskal-Wallis test since all the assumptions of ANOVA could not be met. After investigating the data, it was evident that the level of aptitude of the software engineers was correlated with their production. In the second phase the two-way ANOVA test that scrutinized the variation between competency and project type. The test provided 3 null hypotheses, which led to similar results to those in phase one. Last, these testing methods are applied to analyze a situation in which a customer was unhappy because of a product defect. A two-way ANOVA test directed to the conclusion that a defect tracking system should be in place to prevent future occurrences. An independent analysis of each group concluded that this would be the most effective solution.

The concepts from the simulation that can be used at Ann’s bait and tackle are

Non-parametric and ANOVA testing methods can be applied to a variety of...