Design of Experiments
by Lee Creighton, Mia Stephens, Ann Lehman, Ph.D., John Sall
JMP Start Statistics, 5th Edition
- Preface
- Preliminaries
- JMP Right In
- Data Tables, Reports, and Scripts
- Formula Editor Adventures
- What Are Statistics?
- Simulations
- Univariate Distributions: One Variable, One Sample
- Overview
- Looking at Distributions
- Describing Distributions of Values
- Statistical Inference on the Mean
- Practical Significance vs. Statistical Significance
- Examining for Normality
- Special Topic: Practical Difference
- Special Topic: Simulating the Central Limit Theorem
- Seeing Kernel Density Estimates
- Exercises
- The Difference Between Two Means
- Overview
- Two Independent Groups
- When the Difference Isn’t Significant
- Check the Data
- Launch the Fit Y by X Platform
- Examine the Plot
- Display and Compare the Means
- Inside the Student’s t-Test
- Equal or Unequal Variances?
- One-Sided Version of the Test
- Analysis of Variance and the All-Purpose F-Test
- How Sensitive Is the Test? How Many More Observations Are Needed?
- When the Difference Is Significant
- Normality and Normal Quantile Plots
- Testing Means for Matched Pairs
- Two Extremes of Neglecting the Pairing Situation: A Dramatization
- A Nonparametric Approach
- Exercises
- Comparing Many Means: One-Way Analysis of Variance
- Fitting Curves through Points: Regression
- Categorical Distributions
- Categorical Models
- Overview
- Fitting Categorical Responses to Categorical Factors: Contingency Tables
- Two-Way Tables: Entering Count Data
- If You Have a Perfect Fit
- Special Topic: Correspondence Analysis— Looking at Data with Many Levels
- Continuous Factors with Categorical Responses: Logistic Regression
- Surprise: Simpson's Paradox: Aggregate Data versus Grouped Data
- Generalized Linear Models
- Exercises
- Multiple Regression
- Fitting Linear Models
- Overview
- The General Linear Model
- Kinds of Effects in Linear Models
- Coding Scheme to Fit a One-Way anova as a Linear Model
- Regressor Construction
- Interpretation of Parameters
- Predictions Are the Means
- Parameters and Means
- Analysis of Covariance: Continuous and Categorical Terms in the Same Model
- The Prediction Equation
- The Whole-Model Test and Leverage Plot
- Effect Tests and Leverage Plots
- Least Squares Means
- Lack of Fit
- Separate Slopes: When the Covariate Interacts with a Categorical Effect
- Two-Way Analysis of Variance and Interactions
- Optional Topic: Random Effects and Nested Effects
- Exercises
- Design of Experiments
- Bivariate and Multivariate Relationships
- Exploratory Modeling
- Control Charts and Capability
- Mechanics of Statistics
- Analyze and Graph Menu Commands
- Answers to Selected Exercises
- Chapter 4, "Formula Editor Adventures"
- Chapter 7, "Univariate Distributions: One Variable, One Sample"
- Chapter 8, "The Difference Between Two Means"
- Chapter 9, "Comparing Many Means: One-Way Analysis of Variance"
- Chapter 10, "Fitting Curves through Points: Regression"
- Chapter 11, "Categorical Distributions"
- Chapter 12, "Categorical Models"
- Chapter 13, "Multiple Regression"
- Chapter 14, "Fitting Linear Models"
- Chapter 15, "Design of Experiments"
- Chapter 16, "Bivariate and Multivariate Relationships"
- Chapter 17, "Exploratory Modeling"
- Chapter 18, "Control Charts and Capability"
- References and Data Sources
- Technology License Notices
- Index
15
Design of Experiments
Overview
Designed experiments are an important methodology in science and engineering. We learn by trial and error, i.e. experimentation, and experimental design makes that learning process efficient and reliable.
Design of Experiments (DOE) is probably the single most powerful technique you can use for product and process development, refinement, problem solving, scientific inquiry, and optimization.
The DOE platform in JMP provides tools for creating designed experiments and saving them in a JMP data table. JMP supports two ways to make a designed experiment:
• Build a new design that matches the description of your engineering problem and remains within your budget for time and material. This method uses the Custom Design platform and is recommended for creating all experimental designs.
• Choose a pre-formulated design from a list of designs. These lists of designs include Screening, Response Surface, Full Factiorial, Mixture, Space Filling, Taguchi and other specialized designs.
This chapter shows you how JMP generates and analyzes common experimental designs using the Custom Design platform. This introductory DOE chapter shows examples of screening, response surface, and split plot designs.
See the JMP DOE Guide in the online help for details about designs not covered in this chapter.
-
No Comment
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.