Transcript Document
If we can reduce our desire,
then all worries that bother us will disappear.
Design of Experiments I
Topic: Introduction
Prof. Shenghua(Kelly) Fan
California State University
Representative Sample
• (Simple) random sample: Each group of
units of the required size from the
population has the same chance to be the
selected sample.
Research studies
Observational
studies
Sample surveys
Randomized
experiments
Case control studies
Case Study: Baldness and Heart
Attacks
“A really bad hair day: Researchers link baldness
and heart attack.” “Men with typical male pattern
baldness … are anywhere from 30 to 300 % more
likely to suffer a heart attach than men with little or
no hair loss at all.” Newsweek, March 8, 1993.
Q: What type of study is it?
Which Type of Study?
• Ethical concerns
• Resource limitations
• Desired conclusion:
– cause and effect
– relationship
I want to slim
down.
Should I do
exercise or
limit my fat
intake?
Problem: What are the factors affecting
the taste of a soft drink beverage?
• Type of sweetener
• Ratio of syrup to water
• Carbonation level
• Temperature
• Others??
Iterative Nature of Experimentation
Conjecture
Conjecture
Design Analysis
Experiment
Conjecture
Design Analysis Design
Experiment
Increasing knowledge
Design = Analysis of conjecture & synthesis of experiment
Analysis = Interpretation of data & synthesis of new conjecture
The Phases/Objectives of
Experimentation
Increase knowledge about process under study.
Phase
Objective
# of
variables
Design
type
Screening
(which)
•Identify key variables
•Estimate effects
2-15;
Numerical/
Categorical
Factorial
design
Empirical
(how)
•Fit/test empirical model
•Determine local optimum
2-6;
Numerical
Response
surface design
1-5;
Numerical/
Categorical
Optimal design
Theoretical •Estimate parameters in
mechanistic model
(why)
•Model testing
What Is an Experiment?
• An inquiry in which an investigator
chooses the levels (values) of input or
independent variables and observes the
values of the output or dependent
variable(s).
The Six Steps of Experimental
Design
•
•
•
•
•
•
Plan the experiment.
Design the experiment.
Perform the experiment.
Analyze the data from the experiment.
Confirm the results of the experiment.
Evaluate the conclusions of the
experiment.
Plan the Experiment
• Identify the dependent or output
variable(s).
• Translate output variables to measurable
quantities.
• Determine the factors (input or
independent variables) that potentially
affect the output variables that are to be
studied.
• Identify potential combined actions
between factors.
Problem (Cont.): soft drink
beverage
•What is the output variable?
Taste of the drink; score 1 to 10 (from poor to good)
•What factors and at which levels should we study?
•
•
•
•
Type of sweetener
Ratio of syrup to
water
Carbonation level
Temperature
A, B
Low, High
Design the Experiment
Determine the levels of independent
variables (factors) and the number of
experimental units at each combination
of these levels according to the
experimental goal.
Experimental unit:
the unit we apply the factors on
to get the response.
Problem (Cont.): soft drink
beverage
•What combinations of factors should be studied?
All 2x2x2x2 combinations.
•How should we assign the studied combinations to
experimental units?
Assign equal number of units to each combination.
(unit: the “null” beverage or say the plain water)
More on Experimental Design
• Randomizing the condition assigned to a unit
– the type of treatments
– the order of treatments
• Adding control groups
– placebo; standard treatment; both
• Preventing bias: blinding
– double blind; single blind
• Reducing the variability/ Increasing the accuracy
– blocking; adding covariates (nuisance variables)
Complete Randomized Design
(CRD)
The treatments are randomly assigned to
(experimental) units.
Randomized Block Design
(RBD)
Within each block, conduct CRD using an
equal number of units.
Eg. Block factor: the age group
Basic Terminology
• Factors:
the controlled variables;
must be categorical
• Treatments: conditions constructed from
the factors
• Replications: observations or
measurements in the
observational study
Analysis of Variance (ANOVA)
Example:
Y = battery life
X = battery brand
Objective: the impact of X on Y
study how the value of Y changes
according to X
study how much of the variability
in Y is due to the different levels of X
ANOVA:
Analyze the variability in Y (output variable)
to find the reasons of it (input factors).
Example: Y = LIFETIME (HOURS)
3 replications
per level
1
BRAND
2
3
4
5
6
7
8
1.8 4.2 8.6 7.0 4.2 4.2 7.8 9.0
5.0 5.4 4.6 5.0 7.8 4.2 7.0 7.4
1.0 4.2 4.2 9.0 6.6 5.4 9.8 5.8
2.6 4.6 5.8 7.0 6.2 4.6 8.2 7.4 5.8
The total variability in life (Y)
How the variability in life is associated with brand (X):