Statistical Methods for Quality Engineering

Document Sample
Statistical Methods for Quality Engineering Powered By Docstoc
					Statistical Methods for Quality Engineering


The purpose of this training is to provide those who have the responsibilities of product
or process development and testing the means to make effective decisions concerning
product or process changes. The variation of the product or process performance must be
statistically addressed to make decisions with the appropriate confidence. Technical
personnel are consistently making changes to product and process designs and the
resultant performance changes need a statistical basis for moving ahead to the cost
assessment and release phases.

This course is also very good for the person seeking to attain the Certified Quality
Engineering status within the American Society for Quality.

This course is facilitated by Phil Ross - President of International Quality Services Inc

Course Description

This three-day course is aimed at the person(s) responsible for product and process
changes. Specifically, the person attending this class should have at least high school
mathematical and graphing skills, a good technical understanding of products and
processes in their work environment, and a good technical understanding of testing
methods and protocols.

Various continuous and discrete probability functions are covered with the normal
distribution receiving the most emphasis. Other distributions that are addressed are the
Weibull, Exponential, Binomial, Poisson, and Hypergeometric plus a section on
nonparametric comparisons. Various confidence intervals and tests of comparison for
these probability functions are covered. Some key tests of comparison are the Z test,
Student’s t tests, Chi-Square test, F test, and ANOVA for the normal distribution. The
attendee will work along with the instructor on many examples throughout the three days.
What The Participant Will Learn

The participant will learn:

       How to select the proper distribution model
       How to determine valid sample sizes
       How to design valid tests of comparison
       How to make effective decisions at stated confidence levels

Course Materials

The materials for the SMQE class are handout materials only (no textbook or manual
needed) which make up approximately 220 pages.

Statistical Methods for Quality Engineering
(three-day course)

1. Introduction
   1.1. training objectives
   1.2. statistical resources
2. Concept of variation
   2.1. common development questions
   2.2. histograms
   2.3. descriptive statistics
3. Distribution Models
   3.1. continuous
   3.2. discrete
   3.3. applications
4. Model Selection
   4.1. empirical distribution functions
   4.2. cumulative distribution functions
   4.3. normal probability paper
   4.4. small sample sizes and median ranks
   4.5. tests for normality
5. Parametric Evaluations and Tests
   5.1. normal and log-normal data
       5.1.1. Z confidence intervals and tests
       5.1.2. t confidence intervals and tests
       5.1.3. Chi-Square confidence intervals and tests
       5.1.4. K factor confidence intervals
       5.1.5. F tests
       5.1.6. Analysis of Variance
   5.2. Weibull distribution confidence intervals and tests
   5.3. Exponential distribution tests
   5.4. Poission distribution applications
   5.5. Binomial distribution applications
   5.6. Hypergeometric applications
6. Nonparametric Tests
   6.1. sign tests
   6.2. run tests
   6.3. rank tests

Shared By: