Measurement Decision Risk Analysis Course Outline

State-of-the-Art Measurement Analysis Software, Training and Consulting Services

ŹCelebrating over 20 YRS of Excellence and Innovation 1987-2008

Training Measurement Decision Risk Analysis Course Outline

Measurement Decision Risk Analysis

Course Outline

This outline applies to a 3-day course of instruction presented by Integrated Sciences Group. This course covers the principles of measurement decision risk analysis and provides hands-on training in the use of ISG's AccuracyRatio software. A course segment is devoted to discussion of specific analysis problems relevant to the students' workplace. Click class schedule for our 2008 open-enrollment training locations and dates.

Course Objectives

This course provides a conceptual background in measurement decision risk analysis and hands-on training in the use of AccuracyRatio software. After completion of the course, students will know how to:

define risk baselines
set analysis options
specify equipment parameter bias uncertainties
estimate true in-tolerance percentages from observed percentages
establish and use tolerance guardbands
meet accuracy ratio requirements
compensate for cases where nominal (e.g., 4:1) requirements are not met
select measuring or test equipment for optimal cost vs. accuracy tradeoff
evaluate, interpret and report measurement decision risks
develop minimum-cost risk management programs, policies and strategies.

Day One

Part 1 - Measurement Decision Risk Concepts

Part 1 covers basic concepts of probability and statistics required for an understanding of measurement decision risk. The concepts also serve as a foundation for measurement uncertainty analysis and calibration interval analysis.

What is Measurement Decision Risk?

Why Compute Risks?

Factors Affecting Risk

Uncertainty and Risk Analysis Facts and Fallacies

Part 2 - Uncertainty Analysis Review

Because measurement decision risk analysis involves estimating the uncertainties of measurement, some discussion of this topic is required. Accordingly, Part 2 consists of a brief review of the uncertainty analysis methods and techniques that are covered in our four-day Uncertainty/SPC course.

Uncertainty Analysis and Risk Management

Preliminaries

Error Distributions

Error Sources

Error and Uncertainty

Combining Uncertainties

Estimating Uncertainty

Type A Uncertainty Estimates

Type B Uncertainty Estimates

Type B Formats

The General Uncertainty Analysis Process

Expanded Uncertainty

Reporting Uncertainty

Using Uncertainty Estimates

Day Two

Part 3 - Probability Relations

Computing measurement decision risk requires the use of probability functions. A knowledge of what these functions represent is necessary for understanding differences between types of decision risk and for distinguishing between risks computed at the program management level and risks associated with specific measurement results.

Preliminaries

Definition of Probability
Joint Probability
Conditional Probability
General Rules
Bayes' Relation

Parameter Tolerancing

Matching Attributes Risk Analysis

False Accept Risk

Risk Variables
Probability Relations
False Accept Risk 1
False Accept Risk 2
False Accept Risks 1 and 2

False Reject Risk

Risks with Guardbands

Bayesian Analysis

   Risk Analysis for a Measured Variable
   A priori Knowledge
   Post-Test Knowledge
   Bias Estimates
   Bias Uncertainty Estimates
   Subject Parameter In-Tolerance Probability
   Bayesian False Accept Risk

Part 4 - Computing Risk

Part 4 of the course involves computing false accept and false reject risks from the standpoint of determining risk requirements and assessing risk levels for test or calibration program management and for decision making in response to specific measurement results. Also considered are equipment renewal strategies (adjust or not adjust) and the development of test guardbands.

Equivalent Accuracy Ratios

Relevant Quantities

Including Process Uncertainties

Estimating Risk

Day Three

Part 5 - Compensating Measures

On occasion, computed risks are not suitable for a given application. In these cases, compensation may be possible. Various metrics associated with testing or calibration can be adjusted to ameliorate risks to match requirements. These metrics are discussed in Part 5.

Controlling In-Tolerance Rates

Controlling Process Uncertainties

Pareto Analysis

Multiple Independent Measures

Sequential Testing

Using Guardbands

   Test Guardbands
   Compensating for Risks
   Minimizing Costs
   Compensating for Measurement Uncertainty
   Compensating for Perception Error
   Implications for Interval Analysis
   Guardband Types

Part 6 - Optimizing Risks

False accept risk and false reject risk each cost money. Part 6 examines two approaches to modeling these costs and adjusting them to achieve the minimum total cost.

The Test and Calibration Support Hierarchy

Management Variables

A Simple Approach
A Detailed Cost Model Approach

Part 7 - Example Problems

In Part 7, risk analysis problems related to the student's workplace are analyzed.

What to Bring

PC Notebook
A PC laptop running Windows 98 or higher is required.

Study Materials
Pens, student workbooks and software will be provided.

Analysis Problems
Any measurement decision risk analysis problems that are of interest to your workplace.

Integrated Sciences Group
14608 Casitas Canyon Road Bakersfield, CA 93306
1-800-400-7866 FAX 1-661-872-3669 Tech Support 1-661-872-1683

Page Updated March 23, 2008