Applied Statistics Manual

Joel Smith, Matthew A. Barsalou

Description

This book was written to provide guidance for those who need to apply statistical methods for practical use. While the book provides detailed guidance on the use of Minitab for calculation, simply entering data into a software program is not sufficient to reliably gain knowledge from data. The software will provide an answer, but the answer may be wrong if the sample was not taken properly, the data was unsuitable for the statistical test that was performed, or the wrong test was selected. It is also possible that the answer will be correct, but misinterpreted. This book provides both guidance in applying the statistical methods described as well as instructions for performing calculations without a statistical software program such as Minitab.
One of the authors is a professional statistician who spent nearly 13 years working at Minitab and the other is an experienced and certified Lean Six Sigma Master Black Belt. Together, they strive to present the knowledge of a statistician in a format that can be easily understood and applied by non-statisticians facing real-world problems.
Their guidance is provided with the goal of making data analysis accessible and practical. Rather than focusing on theoretical concepts, the book delivers only the information that is critical to success for the practitioner. It is a thorough guide for those who have not yet been exposed to the value of statistics, as well as a reliable reference for those who have been introduced to statistics but are not yet confident in their abilities.
Supplemental files available! If you are an instructor who would like to conduct training with this book, please visit this "https://asqassets.widencollective.com/portals/sybdffda/(H1550)AppliedStatisticsManualAGuidetoImprovingandSustainingQualitywithMinitab" access:
Course descriptions for one or two semester university courses
Chapter descriptions for standalone sessions
A data file containing data sets used in the book

Sujets

• Sciences & Techniques
• Statistiques
• Statistiques
• Techniques
• Automatisation et Contrôle
• Contrôle des processus
• Automatisation
• Automate
• Sciences et techniques
• Seis Sigma
• Statistics
• Probability
• Mathematics
• Minitab
• Six Sigma
• Probabilités et statistiques

Informations

• Publié par: ASQ Quality Press
• Date de parution: 19 décembre 2018
• Nombre de lectures: 0
• EAN13: 9781953079060
• Langue: English
• Poids de l'ouvrage: 19 Mo

Extrait

Applied Statistics Manual
A Guide to Improving and Sustaining Quality with Minitab
Matthew A. Barsalou and Joel Smith
ASQ Quality Press
Milwaukee, Wisconsin
 
American Society for Quality, Quality Press, Milwaukee 53203
© 2019 by ASQ. Printed in 2018.
All rights reserved.
Library of Congress Cataloging-in-Publication Data
Names: Barsalou, Matthew A., 1975– author. | Smith, Joel, 1980– author.
Title: Applied statistics manual : a guide to improving and sustaining quality with Minitab / Matthew A. Barsalou, Joel Smith.
Description: Milwaukee : ASQ Quality Press, [2018] | Includes bibliographical references and index.
Identifiers: LCCN 2018036196 | ISBN 9780873899758 (hardcover : alk. paper)
Subjects: LCSH: Statistics. | Social sciences—Statistical methods.
Classification: LCC HA29 .B367 2018 | DDC 005.5/5—dc23 LC record available at https://lccn.loc.gov/2018036196
No part of this book may be reproduced in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Portions of information contained in this publication/book are printed with permission of Minitab Inc. All such material remains the exclusive property and copyright of Minitab Inc. All rights reserved.
Director, Quality Press and Programs: Ray Zielke
Managing Editor: Paul Daniel O’Mara
Sr. Creative Services Specialist: Randy L. Benson
ASQ Mission: The American Society for Quality advances individual, organizational, and community excellence worldwide through learning, quality improvement, and knowledge exchange.
Attention Bookstores, Wholesalers, Schools, and Corporations: ASQ Quality Press books, video, audio, and software are available at quantity discounts with bulk purchases for business, educational, or instructional use. For information, please contact ASQ Quality Press at 800-248-1946, or write to ASQ Quality Press, P.O. Box 3005, Milwaukee, WI 53201–3005.
To place orders or to request a free copy of the ASQ Quality Press Publications Catalog, visit our website at http://www.asq.org/quality-press .
 
This book is dedicated to my wife Silvana, who on many evenings shared my attention with a computer screen and never once complained.
Joel Smith
This book is dedicated to my son, Leander G. Barsalou.
Matthew A. Barsalou
 
Preface
The book Applied Statistics Manual: A Guide to Improving and Sustaining Quality with Minitab was written to provide guidance for those who need to apply statistical methods for practical use. While the book provides detailed guidance on the use of Minitab for calculation, simply entering data into a software program is not sufficient to reliably gain knowledge from data. The software will provide an answer, but the answer may be wrong if the sample was not taken properly, the data was unsuitable for the statistical test that was performed, or the wrong test was selected. It is also possible that the answer will be correct, but misinterpreted. This book provides both guidance in applying the statistical methods described as well as instructions for performing calculations without a statistical software program such as Minitab.
One of the authors is a professional statistician who spent nearly 13 years working at Minitab and the other is an experienced and certified Lean-Six Sigma Black Belt. Together, they strive to present the knowledge of a statistician in a format that can be easily understood and applied by nonstatisticians facing real-world problems.
Their guidance is provided with the goal of making data analysis accessible and practical. Rather than focusing on theoretical concepts, the book delivers only the information that is critical to success for the practitioner. It is a thorough guide for those who have not yet been exposed to the value of statistics, as well as a reliable reference for those who have been introduced to statistics but are not yet confident in their abilities.
ACKNOWLEDGMENTS
Matthew Barsalou would like to thank Artur Wilk and Jean-Baptiste Selfa for providing helpful feedback and suggestions.
Portions of information contained in this publication are printed with permission of Minitab. All such material remains the exclusive property and copyright of Minitab. All rights reserved.
 
Introduction
The book is divided into chapters based on common statistical concepts and often times aligns with the menus in Minitab. Certain chapters require an understanding of information from previous chapters, so readers using the book as a foundation should review material sequentially.
When learning how to apply statistics, it is important to first establish a fundamental understanding of data and distributions. The first chapter explains critical statistical concepts such as the difference between samples and populations as well as sampling methods and types of data. Chapter 2 introduces concepts related to descriptive statistics, such as measures of central tendency, measures of variability, and measures of shape. Chapter 3 describes Exploratory Data Analysis, which is an important first step in analyzing data. Chapter 4 describes probability statistics and various continuous and discrete distributions.
Once the fundamentals are understood, a variety of tools are accessible to practitioners to help them make decisions based on data. Chapter 5 explains hypothesis testing—the basis for statistical decision-making—for means, proportions, and variances. This chapter describes how to perform the hypotheses tests using both the p-value method and the critical value method. The critical value method is taught more often for those performing tests by hand, while the p-value method is the method typically taught to those with access to Minitab or other statistical software. Chapter 6 describes how to determine sample sizes and confidence intervals for the previously described hypothesis tests.
Chapter 7 builds on the basic hypothesis tests learned in Chapter 5 but pertains to the distribution of data and includes the normal probability plot, transformations, and individual distribution identification. Chapter 8 then describes nonparametric tests and their uses.
To this point in the book, hypothesis tests have focused on attributes of single variables or the relationship between two variables. The next chapters expand on this to enable many variables to be modeled simultaneously. Chapter 9 teaches analysis of variance (ANOVA), including the General Linear Model (GLM), and Chapter 10 covers regression.
A critical concept to understand is that of measurement error, even when performing basic tests. Once the reader is familiar with modeling, various tools for assessing measurement error can be understood. Chapter 11 presents various types of measurement system analysis (MSA).
Assessing the quality of a product or process requires understanding how it behaves. The subsequent chapter describes different types of process capability studies used to quantify the quality of a process or product. Chapter 13 discusses statistical process control (SPC) and includes control charts for variable data, continuous data with weighted averages, and attribute data.
To establish likely causation between factors in a study and a response of interest, a controlled experiment must be performed. Chapter 14 concerns design of experiments (DOE) and covers factorial designs, response surface, and mixture designs.
Chapter 15 then covers reliability, including distribution identification, distribution overview, and distribution analysis. Other quality tools that Minitab can create, such as the cause and effect diagram, Pareto chart, and run chart, are covered in Chapter 16.
Appendix A presents statistical terminology written in layman’s terms, and Appendix B presents a quick guide to common hypothesis tests. Appendix C includes various statistical tables. Appendix D contains control chart constants, and Appendix E presents a tour of Minitab.
 
Chapter 1
Sampling and Types of Data
SAMPLES AND POPULATIONS
A population consists of every potential unit of something, and a sample is a subset of the population (Moore 1991). The population is conceptual and might be thought of as an infinite quantity of units that we draw from, and statistics is the science of using a sample to estimate characteristics of that population. A sample from an item with a limited production run might consist of 8 randomly selected items (as shown in Figure 1.1 ) or all 441 parts produced.

Total production size can be incredibly large. There may be over one million units in existence for an item that has been in production for a period of years. It would be unrealistic to measure every one of these items, so a smaller sample would be used instead. A properly selected sample provides an opportunity to make inferences about the parent population.
Even measuring every single part produced is typically considered a sample of the theoretical population, which can only be estimated. It would be incorrect to assume that the characteristics of a population are known even after 100% sampling. To understand why this is erroneous, consider a perfectly fair six-sided die where each side has equal probability of landing up when rolled. The population of rolls of this die consists of an infinite number of potential rolls and has the characteristic of each number occurring with probability 1/6 on any given roll. We are unlikely to ever have our sample indicate precisely a 1/6 chance of each number occurring even if we record every single roll of the die. This means that even sampling 100% of available units is still providing a theoretical subset of the population.
As another example, think of a perfectly stable process where each part has a length pulled from a single distribution. The part comes from a single distribution because the process is perfectly stable. A stable process it is not undergoing changes over time and the true population me