Statistical Strategies for Managers

Course Outline

Module 1: Statistical Thinking (2 hours)

• The Role of Management
• The Data Difference and Why Use Statistics
• Translation of an Engineering Problem to a Statistical Problem
• How do you Know Something and how Statistics can Help
• How Statistics can Give a Competitive Advantage
• Statistical Application Opportunities for Business including the Major Application of How Engineers/Researchers Do their Daily Jobs
• Manager Exercise: Which problems are statistical problems?
• Principles of Statistical Thinking
• Game: Testing Your Statistical Intuition
• Risk Management through Diversification
• Types of Variation—Common vs Special Cause
• Common Cause Problem Solving Roadmap
• Descriptive Statistics
  • Center: Mean vs Median
  • Spread: Range vs Stdev
    • Why Variances are Important
  • Shape: Normal and Skewed Distributions
    • Histograms, Probability Plots, Box Plots, Interval Plots
  • Graphical Pitfalls and How to Avoid Them
  • Manager Exercise: Questions to Ask when Reviewing Graphics.
  • Why center and spread are not a complete description.
• What Managers Can Do
• Summary of engineering or business objective, statistical technique or strategy, how to check validity and questions to ask.
• Self Test: 11 Questions

Module 2: Control Methods (2 hours)

• Control Method Selection
• Why use SPC?
• Where SPC is Applied
• Data Type and Analysis Matrix
• How Averages Behave
• Xbar/R Charts, Individual/MR Charts
  • Interpretation of Minitab Output
  • What Controls the Width of the Control Limits.
• 4 states of a Process
• Naive Data Analysis
• Manager Exercise: Critique on Management Action?
• Process Set Up Optimal Strategy
• Implementation Strategy
• Identifying Key Process Input and Output Variables (KPIV’s, KPOV’s)
• Selecting the Right Control Chart
• Chart Design for Subgroup Size, Frequency, Subgrouping, OOC Rules
• Control Chart Errors and What They Mean
• How to Tell if a Chart is Effective
• Out-of-Control-Action-Plan (OCAP)
• How to Search for Assignable Cause
• Four Problems with SPC, Special Topics and How to Fix Them
  • Batch Processing and the I/MR/R and I/R/I Charts
  • Non-Normality: Percentile Limits
  • Correlated Data: 3 Sigma Limits, EWMA
  • Fixed Effects: Individual Charts
  • Short Run SPC
  • Mixtures: Zero Inflated Distributions
  • Machines Running Different Products
  • Multivariate Control with Correlated Variables
• Manager Exercise: Management use of SPC
• Appendix: Ship-To-Control
• Summary of engineering or business objective, statistical technique or strategy, how to check validity and questions to ask
• Self Test: 10 Questions

Module 3: Measurement System Analysis (2 hours)

• Why Measurement Studies are Important
• Sources of Measurement Variation
• Order to Collect Data
• Game: Why Knowing Precision is Important
• Measurement Terminology
  • Measurement Error
  • Bias, Accuracy and Precision
  • Assessing Resolution
  • Repeatability and Reproducibility
  • Bias and Linearity
  • Measurement Metrics: Percent of Tolerance; Percent of Variation; and How People Cheat on Gauge R&R
• Gauge R&R
  • Interpretation of Minitab Output and What to Look For
  • Strategy to Reduce Repeatability by 30% or More with No Capital Investment
  • Expanded Gauge R&R
  • One-Sided Specs
  • Gauge Studies for Combined Repeatability and Reproducibility
  • Destructive and Non-Repeatable Testing
  • Special Methods to Reduce Nuisance Variation
  • Manager Exercise: Spotting Problems in a Variables Gauge R&R Study
  • Attribute Agreement Analysis
    • Percent Agreement
    • Kappa Statistics
    • Overkill and Escape Assessment
    • How People Cheat on Attribute MSA Studies
• Bias and Linearity Studies
  • Type I Gauge Studies
  • Bias and Linearity Tests and Calibration Curves
  • Engineering Delta Test
• Tool Matching
  • Correlation Criteria
  • Orthogonal Regression Criteria
  • Paired t-Test
• Summary of engineering or business objective, statistical technique or strategy, how to check validity and questions to ask.
• Self Test: 12 Questions

Module 4: Decision Making with Data (4 hours)

• Problems with Observational Data
• Data Cleaning
• Statistical Reasoning
  • Game: ESP Testing
  • 4 Steps in a Statistical Test
  • 2 Decision Errors
  • 2 Ways to reduce Missed Signals
  • Statistical Significance: p-Values
  • Practical Significance
  • 5 Conditions to Accept a Conclusion from Data
  • Confidence Intervals for Managers
• One, Two and Three or More Sample Comparisons
  • t-Test
  • Paired t-Test
  • Game: Testing Golf Putting Performance
  • ANOVA
  • Applications
  • Multifactor ANOVA
• Modeling Continuous Inputs vs Continuous Outputs
  • Simple Linear Regression Analysis
  • Multiple Regression
• Modeling Discrete Inputs on a Discrete Output
  • Chi Square Analysis
• Modeling Continuous Inputs on a Continuous Output
  • Logistic Regression
• Multivariate Methods
  • PCA
  • PLS
  • Discriminant Analysis
  • Recursive Partitioning
• Summary of engineering or business objective, statistical technique or strategy, how to check validity and questions to ask.
• Self Test: 18 Questions

Module 5: Planning, Executing and Analyzing Evaluations (Design of Experiments) (4 hours)

• Brief History of DOE
• Terminology for DOE
• 6 Purposes for DOE
• Delusion of Experiment vs Design of Experiments
• One-Factor-at-a-Time Experimentation
• Main Effects
• Interaction
• Principles of DOE
• Replication
• Randomization
• Importance of DOE
• One Factor Experiments
  • Simple Comparative Experiments
  • Comparing Proportions
  • One Continuous Factor Optimization
  • Multi Factor Experiments
  • General Full Factorials
  • Two Level Factorials
  • Blocking
  • Unreplicated Full Factorials
  • Idea of Robustness
  • Center points
  • Actions from a DOE
  • Fractional Factorials
• Optimization Strategy
  • CCD and Box Behnken Designs for Optimization
  • Multiple Response Optimization with Desirability Functions
  • Confirmation Runs
• Planning a DOE
• Game: Optimizing a Process with Noise
• What can go Wrong in a DOE and Special Situations
  • Hard-to-Vary Factors
  • Experiments with Factors at Multiple Process Steps
  • Botched Runs
  • D and I Optimal Designs
  • Covariates
  • Robust Design
  • Ruggedness Testing
  • EVOP
  • Mixture Designs
• Manager Exercises
• Appendices:
• Troubleshooting Guide for Non-Significant DOE’s
• 60 Ways to Mess Up a DOE
• Summary of engineering or business objective, statistical technique or strategy, how to check validity and questions to ask.
• Self Test: 20 Questions

Module 6: Qualification and Verification Methods (2 hours)

• Benefits of Knowing Statistical Methods for Qualification and Verification Testing
• Two Important Considerations in Qualification and Verification Testing
• Verification, Validation and Qualification Distinctions
• Translating Technical Requirements into Statistical Requirements
• Selecting a Statistical Test
• Concepts of Sample Size and Risk
• Number of Lots for Testing
• Review of Statistical Testing
  • Steps, Decision Errors, p-Values, 5 Conditions to Accept a Conclusion
• Tests Comparing Groups for Expected Performance
  • Comparing Means with t-Tests
  • Comparing Means with the Equivalence Test
  • Comparing Means with Different Test Conditions (Blocks)
  • Comparing Percent Conformance (Yield)
  • Minimal Sample Size Tests: Sequential Analysis Methods
    • History, Concept and Definition
    • When Sequential Analysis is Appropriate
      • Testing Means
      • Testing Proportions
      • Game: Sequential Analysis
    • Comparing Variation for Two or More Variances
    • Multiple Test Comparisons
  • Comparing Metal Levels to Medians using Binomial
    • Censored Data Methods
      • Partially Censored Data Comparison in Minitab
      • Totally Censored Data Comparison using Fisher’s Exact Test
  • Change FMEA
• Manager Exercise: Recommend a test and sample size for the following situations
• Summary of engineering or business objective, statistical technique or strategy, how to check validity and questions to ask.
• Self Test: 10 Questions