10th International
Conference on
Teaching Statistics
8 – 13 July 2018
Kyoto, Japan
Contributed paper list


Contributed Paper C287

 

Design and Statistical Analysis of Catapult Experiments


Presenter

Siong Lin Ho (Singapore)

Abstract

This paper centers on designing innovative catapult experiments to simulate the scenario of “product design and development” or an “industrial manufacturing process” for teaching Statistics and data analytics. Empirical modeling via statistical design of experiments is a powerful methodology used to determine the most important variables and for identifying an optimal combination of process or design variables that can significantly impact output performances. In our catapult model, we demonstrated the use of 2k and 3k factorial design techniques. From statistical analysis of data using Analysis of Variance (ANOVA), F-tests and p-values, the significant factors that maximize the projectile distance and the optimal combination of factors’ levels are identified. Regression models with linear and quadratic components are constructed for prediction. The measurement systems variation is briefly discussed.