Comparison of Experimental Assessment Methods to Estimate the Strength of Gear Teeth.

Regression type methods for analysing proportions, known as probit and logit analyses, are compared with the British Standard BS 3518 (part 5) recommendations for analysis of staircase designs. All these methods are found to give similar results. Different experimental strategies, involving fixed sample sizes for load levels chosen in advance, are compared with staircase designs. Recommendations for practice are made. Finally, the versatility of the logit analysis for investigating the effects of additional variables, such as additional treatments during manufacture, is demonstrated through a case study. [ABSTRACT FROM AUTHOR]