From: To: ; Subject: Feedback on the Advanced Statistics Probability and Distributions from chingdp@appliedbiosystems.com Date: Monday, February 28, 2005 8:30 AM 01. Important: = The various types of discrete distributions available to handle estimating probability of a situation to take place or not take place, based on a set of samples taken from the larger population. All the discrete distributions were new to me and I did not know what they looked like, how they were different, and how to apply them to situations until I learned about them in this course. 02. Unclear: = I would like to know a more fundamental answer as to why the shape of the Binomial, Hypergeometric, and Poisson distributions look the way they do. Why does the discrete distribution bunch up near the beginning of the x-axis and trail off rapidly toward higher x values. I know you have presented equations for each distribution, without going into the theoretical details because it was beyond the scope of this course, and substituting values into the Excel formulas gives an answer. I would also like to know if you calculated the probability of an occurrence, how does that look on the discrete distribution curve. Is there a visual way to show a probability result directly on the discrete distribution curve? I am thinking of how you showed the relationship of the z-score to the highlighted proportion and location on the continuous normal distribution curve. 03. Apply: = I will try to apply these ideas to my workplace when the opportunity arises. Most of what is charted by some of my co-workers is manufacturing process control charts using normal probability distributions. I have not seen anyone use the Binomial, Hypergeometric, Poisson or Discrete Uniform distributions in a chemical manufacturing environment. One place that could benefit from using the Binomial or Hypergeometric distribution is to take MORE samples from different locations in a batch of chemical product and perform MULTIPLE replicate analyses. If the material passes product specification then no problem, but is it fails we will have better confidence to look for the cause of the failure. The additional testing would be difficult and expensive to implement for every batch and would require new protocols for additional in-process testing and approval of the test plan by management. 04. Expectations: = Yes. I learned a lot about probability and distributions and how these tools are to be used to analyze the odds of an outcome for discrete variable items and situations. 05. Rating: = 10 07. What else: = I enjoyed this course very much. It was very convenient to learn at my own pace so I can fit it in with my busy workload. I do miss the interaction with fellow classmates that could give more insights and experiences to handling the situations you used in the examples and exercises to teach us the concepts. I appreciate that Glen always gives excellent and timely feedback to my questions to help clarify my misunderstand of the course contents. I will definitely recommend this course to other co-workers if they need to learn statistics. Email = chingdp@appliedbiosystems.com SoupermailConf = /statistics/advanced/eStats-PD/review.txt Submit = s u b m i t y o u r r e v i e w ------------------------------- Remote Host: user-38ldt09.dialup.mindspring.com Remote IP: 209.86.244.9 User Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0) Referer: http://www.margaret.net/statistics/advanced/eStats-PD/review.htm -- No virus found in this incoming message. Checked by AVG Anti-Virus. Version: 7.0.300 / Virus Database: 266.5.0 - Release Date: 2/25/2005