# DESIGN OF EXPERIMENTS by pptfiles

VIEWS: 5 PAGES: 14

• pg 1
```									DESIGN OF
EXPERIMENTS
DOE

STATISTICAL EXPERIMENTS
By Mike Newtown
• Why experiment?
• To increase knowledge about a process
• What good is increased knowledge?
• Increase the output of the process
• Reduce variability
• Produce a robust product that is not
influenced by environmental factors
• Why Experimental Design?
• To obtain valid conclusions from
an experiment
• To yield the maximum amount of
useful information with the least
amount of experimentation
(experiments cost money)
• Types of scientific studies
• Experimental- variables of interest
can be controlled
• Observational- variables can not
be controlled (data analysis)
• Choose at least two factors that
influence the output of the
experiment
• Rank order the factors
• Determine levels of control
• Consider the resource available
to perform the test
Good Experiments
Should:
• Eliminate know source of bias
• Guard against unknown source
of bias
Variation

• Experimental Errors:
• Setting not exact enough to be
fixed exact same point each time
• Measurement Errors:
• Inability to obtain exactly the
same measurement on the
successive, identical test runs
Combat Variation

• Standardize procedure
• Collect larger number of data
points

Accuracy & Precision vs. Cost
Scientific Method

•   Investigate situation
•   Define problem clearly
•   Formulate hypotheses
•   Select appropriate experiment
•   Conduct experiment
•   Analyze results
•   Draw Conclusions
•   Take action on conclusions
Statistics

•   Mean
•   Range
•   Standard Deviation
•   Z tables

Are very important in analyzing
data
Hypothesis Testing

• In an experiment, we don’t test
to see if our ideals are right, we
test to see if they are wrong.
• Null Hypothesis (Ho) statement
of ideal as if nothing unusual
was occurring
• Alternative hypothesis (Ha)
Lets see how this works

Problem: Sub dough at Subway
Factor A: dough type (type 1 & 2)
Factor B: oven temperature
(125 and 140 Celsius)
Factor C: bake time (8 mins or 10
mins)

See board work