Knowledge on statistics is increasingly important for professionals in modern business. For example, hypothesis testing is especially one of the critical topics for quality managers and team workers in Six-Sigma training programs. Delivering the knowledge of hypothesis testing sometimes is an issue for people who do not have strong learning capability. This paper proposes to add a direct statement before the conventional procedures of hypothesis testing that are popularly adopted in the current statistics textbooks and training materials used in U.S. business schools and industrial training programs. The effectiveness of this modified procedure was tested statistically. The results showed the proposed approach can be helpful for learners/trainees who are not strong in reasoning thinking to select the correct hypotheses and to reach the right conclusion in hypothesis testing of applications problems. The rationality of this approach can be justified by poka-yoke concept, a principle well recognized in operations management for fail-safing
[...] Since our focus is on how the learners perform in setting up the correct hypotheses for applications problems, we review the scoring so that only the part of performance on hypothesis construction is taken out and re-scored in a scale from 1 to 5. The results are given in Table 1. That is, the data is Table 1 is not the raw data, but the rescored data of partial performance in setting up hypotheses. This modification must be done because the score in the raw data assesses the general performance on the whole problem. [...]
[...] And as the test among selected higher performance learners suggests the effect of this addition is insignificant, the proposed Step 0 may be just effective for the learners whose learning capability is not strong enough. Whether this approach is effective in better schools/programs leaves a question to be tested. Nevertheless in some situations, the proposed approach can be helpful. Moreover, very likely the approach can also help the learners to conclude better. Hence, it may be worth introducing this approach to others. [...]
[...] When a learner did not get a right conclusion in solving an applications problem on hypothesis testing aided by computer software, very often it is because the form of H o and H a is wrongly constructed. The correct way to set up H o and H a is, of course, as addressed in the best seller Modern Business Statistics by Anderson et al. general, a hypothesis test about the value of a population mean µ takes one of the following three forms: H o : µ µ0 H o : µ µ0 H o : µ = µ0 H a : µ < µ0 H a : µ > µ0 H a : µ µ0 In many situations, the choice of H o and H a is not obvious and judgment is necessary to select the proper form. [...]
[...] Smith. (2005) Business Statistics, A decision Making Approach, 6th ed., Pearson/Prentice Hall, NJ. Hinckley, C.M. and Barkan, P. (1995). The role of variation, mistakes, and complexity in producing nonconformities. Journal of Quality Technology 27(3):242-249. Keller, G., B. Warrack. (2003) Statistics for Management and Economics, 1st ed. Brooks/Cole, CA. Levine, D. Stephan, T. Krehbiel, and M. Berenson (2002) Statistics for Managers [...]
[...] For pure education and training purpose, understanding principles may be even more important. Where to draw the line of the trade-off may depend on the type of the education/training programs and the type of the learners/trainees. References Anderson, D., D. Sweeney, T. Williams. (2003) Modern Business Statistics with Microsoft Excel, 1st ed. South-Western, OH. Bandyopadhyay, J. K. (1993). “Poka yokay systems to ensure zero defect quality manufacturing.” International Journal of Management 10 29-33. Bowerman, B., and R. O'Connell. (2003) Business Statistics in Practice, 3rd ed., McGraw-Hill/Irwin, New York, NY. [...]
using our reader.