Experiment Protocol Process
Experiment Protocol Process
Administrative
Protocal Name
Protocol ID Number
Version History
Background
Description of the Process
Motivations
Instructions from Management
Voice of the Customer (VOC)
References to Process Information
Calibration to establish accuracy
GR&R study to demonstrate measurement reliability
First Principles
Process documentation, SOPs
DFMEA & PFMEA
SPC records
Process capability studies
Corrective Actions
Problem to be Solved
Review of Previous Investigations
Experiment Design
Input-Process-Output (IPO) Diagram
Response(s)
CTQs
KPOVs
Ordinary POVs
Factors
Controlled
Uncontrolled
Can't be measured
Can be measured
Type
Qualitative
Fixed Levels
Random Levels
Quantitative
Chosen Design
Design Family
Low resolution designs, e.g. screening experiments for many PIVs
Intermediate resolution designs, e.g. two-level factorials and fractional factorials
High resolution designs, e.g. response surface designs
Variables Matrix
Design Matrix
Alternative Designs Considered
Sample Size Calculation
Using the intended analysis/model
Standard deviation estimate
Effect size estimate and power
Randomization Plan
Validate by analyzing the run order as the response
Blocking Plan
Create the data collection worksheets
Run number
KPIVs
CTQs
Covariates
Notes/comments
Validate the design by analyzing a fictional response
Experiment Procedure
Personel required
Equipment and hardware required
Material required
Safety plan
Plan for managing interruption to production process
Experimental procedure
Data recording process
Contingency Plan
Review
Management Review
Technical Review
Customer/Process Owner Review
Institutional Review Board (IRB)
Animal Review Board (ARB)
Statistical Analysis
Software to be used and version number
Analysis Method
ANOVA/Regression
Regression
Binary Logistic Regression
Poisson Regression
Regression with Life Data
Other
Models
Full Model
Post-Occam Model
Alternative Models
Tests of Assumptions
Requirements of Residuals
Homoscedasticity
Normality
Freedom from Outliers
Goodness of Fit
Contingencies
Model Acceptance Criteria
Contingency Model
Model Confirmation Study
Design of
Analysis of
Acceptance Criteria
Model Application
Maximize
Minimize
Hit target value
Simultaneous requirements on CTQs
Reduce process variation
Robust design
Cost Analysis
Cost of the experiment
Adminstrative cost
Cost per cell of the experiment design matrix
Cost per replicate observation
Cost of analysis and interpretation
Cost to implement the recommended actions
Risks of not doing the experiment
Experiment Report Format
Administrative Information
Findings (abstract or executive summary)
Introduction
Background information
Goal
Experiment Design
Variables matrix
Design Matrix
Sample size, randomization, and blocking plan
Alternative designs considered
Experiment Administration
Experiment Procedure
Experimental Data
Deviations from Protocol
Analysis
Software and Version Number
Graphical analysis
Statistical analysis
Full Model
Post-Occam Model
Assumption Checks
Normality
Homoscedasticity
Goodness of Fit
Contingency analysis
Interpretation of the model
Evaluation against decision criteria
Discussion
Success or failure of the experiment
Consequences of deviations from protocol
Recommendations for future experiments
