Yukio TakahashiI've been using SAS since before JMP4. At my previous company, I provided statistical education using various versions of JMP while promoting JMP within the company. I have occasionally presented the results of this work at JMPer's meetings. After retirement, I established a self-employed consulting firm and continue to work there. Recently, I have come to realize that the ability to analyze real-world data statistically is declining collectively in many academic fields in Japan (with the exception of clinical statistics). For me, experimental design has been essential for learning statistical theory and applying it in practice. The topic of this lecture is related to the design of experimental sample size. The challenge I faced was to enable experimental researchers to estimate the minimum detectable sensitivity (MDD) for a new Latin square experiment with a fixed sample size using the square root of the mean sum of squares (RMSE) of residuals from previous Latin square experiments. The t-distribution function available in JMP scripts includes a non-central t-distribution function, making it difficult to say that this can be achieved using it. Fortunately, I was able to discover what could be done by using the DOE's "Sample Size Explorer" in a creative way, which was very encouraging.
Representative Director - BioStat Research Institute Co., Ltd.
