Course Title: Statistical Methods for Process Improvement:

Part 4: Using Data to Make Decisions

Davis M. Woodruff, PE, CMC

Course Outline

This 2 hour course presents an overview of using small samples to draw conclusions about process or other changes at a defined confidence level. It provides the participant with tools and techniques for using data to make decisions with small samples as well as larger samples. To get the most benefit, the participant should have completed Parts 1 (PDH 207) and 2 (PDH 208) of Statistical Methods for Process Improvement or other statistical methods courses. While some theory is essential to understanding this is not a theoretical statistics course and is intended as a practical course for using small sample statistics.

This course will also provide help organizations that are either ISO 9001, 14001, 13485, TS 16949 or AS 9100 registered or seeking registration to meet the requirements for process and product monitoring and measurement as well as data analysis and continual improvement found in the standards.

This course discusses the following topics:

• Sample Statistics
• An Overview of Hypothesis Testing
• Errors in Testing
• Procedure for Hypothesis Testing
• Testing Variances
• Testing Means
• The Universal Statistic, Chi-Squared

Practical illustrations are used throughout the course and these process improvement tools may be adapted to any business or business model.

This course includes a multiple choice quiz at the end, which is designed to enhance the understanding of the course materials.

Learning Objective

At the conclusion of this course, the participant will:

• Understand the basics of hypothesis testing;
• Understand the two errors in hypothesis tests and what they mean;
• Be able to clearly state the test hypotheses;
• Know how to establish a confidence level for a specific test;
• Be able to compare variances of two populations;
• Be able to compare means of samples and populations;
• Know how to use a paired test for means for similar processes;
• Be able to use the Chi Squared statistic in a variety of situations; and
• Understand how to interpret test results correctly.

Intended Audience

Any professional who is involved in process or product monitoring or measurement  or who needs to understand how to apply statistical methods to sustain effective process improvements. Engineers, consultants and managers interested in understanding process or product monitoring and measurement as a part of the CI process and to more effectively manage your business using facts will benefit from this course on small sample statistics.

Benefit to Attendees

Course participants will learn how to use simple, powerful and practical statistical methods for process improvement that can guide fact based decision making thought the effective use of statistical inference and hypothesis testing using small samples.

Course Introduction

Previously, the subject of descriptive statistics was discussed (PDH course 207). The purpose is fairly self-explanatory -- description of a group of data.  Now we turn our attention to an area with a different focus -- statistical inference.  With statistical inference, the objective is to infer conclusions about populations by using information obtained from samples.  Statistical inference involves mathematical procedures which allow us to draw these conclusions. This course presents an overview of the subject and gives practical information and examples

Often times in business and industry one has small samples to work with in determining the effects of process improvements or changes. Also, there are times when it is necessary to compare results of small samples from different plants or facilities to determine if differences exist. For example, when evaluating raw material suppliers one may need to compare two or more suppliers. Small sample statistics could be the tools needed for these comparisons.

In Part 4 of Statistical Methods for Process Improvement we will learn about correctly stating the hypothesis we are attempting to accept or reject and how to interpret what the information is telling us.

The mathematical methods used in statistical inference are referred to as hypothesis tests.  A hypothesis is a statistical statement made for the purpose of rejecting or not rejecting (accepting).  Hypotheses can be stated about means, variances or shapes of distributions. Some examples of hypotheses are:

• The rejection rate for resin #125 in Demopolis is the same as St. Paul.
• The mean dimension is 2.105 inches.
• The variability on this product is no different in Greenville or Hayneville.

The test is simply a means of deciding whether the hypothesis is false or not, depending upon the sample observations. The key is asking the right question and selecting the correct methods for testing the hypothesis.

This 2 hour course presents a practical and common sense approach to understanding and using process control charts as part 3 a this five part series of Statistical Methods for Process Improvement. These are simple, yet powerful tools for process improvement.

Course Content

In this lesson, you are required to download and study the following course content in PDF format:

Course Summary

This five part series of courses provides the information necessary to apply fundamental statistical concepts and methods for process improvement. Statistics can be theoretical and boring. In fact, many engineers dreaded taking “stats” in college, but now find that practical statistics are essential in today’s work place. This course, part 4, as well as the other 4 parts will provide an understanding of how to really use statistics for process improvement. This is not a course in probability theory or theoretical statistics.

A suggested reading list and detailed glossary is included with each of these courses. Also, there are several articles posted on www.daviswoodruff.com that can be downloaded and used along with these courses.