Design of Engineering Experiments

Chapter 2 – Some Basic Statistical Concepts

• Describing sample data
– Random samples
– Sample mean, variance, standard deviation
– P l ti
Populations versus samples
l
– Population mean, variance, standard deviation
– Estimatingg pparameters
• Simple comparative experiments
– The hypothesis testing framework
– The two-sample t-test
– Checking assumptions, validity

Chapter 2 Design & Analysis of Experiments 1

7E 2009 Montgomery
 Portland Cement Formulation (page 24)
• 2 different formulations referred to as 2 treatments or 2 levels of
the factor formulations. New formulation is able reduce cure time.
• However,, need to evaluate the bond strengthg of the formulations.

Chapter 2 Design & Analysis of Experiments 2

7E 2009 Montgomery
 Graphical View of the Data
Dot Diagram, Fig. 2.1, pp. 24

• Each observations is called a run. Each strength value differ so

there’s some fluctuation or noise or experimental error which is
generally
ll unavoidable.
id bl This
Thi implies
i li bond
b d strength
t th is
i a random
d
variable.

Chapter 2 Design & Analysis of Experiments 3

7E 2009 Montgomery
 If you have a large sample,
sample a
histogram may be useful

Chapter 2 Design & Analysis of Experiments 4

7E 2009 Montgomery
 Box Plots
Plots, Fig
Fig. 22.3,
3 pp
pp. 26

Chapter 2 Design & Analysis of Experiments 5

7E 2009 Montgomery
 The Hypothesis Testing Framework

• Statistical hypothesis testing is a useful

framework for many y experimental
p
situations
• Origins of the methodology date from the
early 1900s
• We will use a procedure known as the two-
two
sample t-test

Chapter 2 Design & Analysis of Experiments 6

7E 2009 Montgomery
 The Hypothesis Testing Framework

• Sampling
p g from a normal distribution
• Statistical hypotheses: H :   
0 1 2

H1 : 1   2
Chapter 2 Design & Analysis of Experiments 7
7E 2009 Montgomery
 Estimation of Parameters
1 n
y   yi estimates the population mean 
n i 1
1 n
S 
2

n  1 i 1
( yi  y ) estimates the variance 
2 2

Chapter 2 Design & Analysis of Experiments 8

7E 2009 Montgomery
 Summary Statistics (pg. 36)
Formulation 1 Formulation 2

“New recipe” “Original recipe”

y1  16.76 y2  17.04
S  0.100
1
2
S 22  0.061
S1  0.316 S 2  0.248
n1  10 n2  10

Chapter 2 Design & Analysis of Experiments 9

7E 2009 Montgomery
 How the Two
Two-Sample
Sample t-Test
t Test Works:
Use the sample means to draw inferences about the population means
y1  y2  16.76  17.04  0.28
Difference in sample means
St d d ddeviation
Standard i ti off th
the diff
difference in
i sample
l means
2
 y2 
n
This suggests a statistic:
y1  y2
Z0 
 12  22

n1 n2

Chapter 2 Design & Analysis of Experiments 10

7E 2009 Montgomery
 How the Two-Sample t-Test Works:
Use S and S to estimate  and 
1
2 2
2
2
1
2
2

y1  y2
The previous ratio becomes
2 2
S S

1 2
n1 n2
However, we have the case where      2
1
2
2
2

Pool the individual sample

p variances:
( n  1) S 2
 ( n  1) S 2
Sp  1
2 1 2 2
n1  n2  2
Chapter 2 Design & Analysis of Experiments 11
7E 2009 Montgomery
 How the Two-Sample
Two Sample tt-Test
Test Works:
The test statistic is
y1  y2
t0 
1 1
Sp 
n1 n2
• Values of t0 that are near zero are consistent with the null
h
hypothesis
th i
• Values of t0 that are very different from zero are consistent
with the alternative hypothesis
• t0 is a “distance” measure-how far apart the averages are
expressed in standard deviation units
• Notice the interpretation of t0 as a signal to noise ratio
signal-to-noise
Chapter 2 Design & Analysis of Experiments 12
7E 2009 Montgomery
 The Two-Sample (Pooled) t-Test
(n1  1) S12  (n2  1) S 22 9(0.100)
9(0 100)  9(0.061)
9(0 061)
S 
2
  0.081
n1  n2  2 10  10  2
p

S p  0.284

y1  y2 16.76  17.04
t0    2.20
1 1 1 1
Sp  0.284 
n1 n2 10 10

The two sample means are a little over two standard deviations apart
Is this a "large" difference?

Chapter 2 Design & Analysis of Experiments 13

7E 2009 Montgomery
William Sealy Gosset (1876, 1937)

Gosset's interest in barley cultivation led 
him to speculate that design of 
experiments should aim, not only at 
improving the average yield, but also at 
breeding varieties whose yield was 
insensitive (robust) to variation in soil and 
climate.
climate

Developed the t‐test (1908)

Gosset was a friend of both Karl Pearson 
f d fb h l
and R.A. Fisher, an achievement, for each 
had a monumental ego and a loathing for 
the other. 

Gosset was a modest man who cut short 
an admirer with the comment that “Fisher 
would have discovered it all anyway.”
ld h di d it ll ”

Chapter 2 Design & Analysis of Experiments 14

7E 2009 Montgomery
 The Two-Sample
p ((Pooled)) t-Test
• So far, we haven’t really
done any “statistics”
statistics t0 = -2.20
2.20

• We need an objective
basis for deciding how
l
large the
h test statistic
i i t0
really is
• In 1908,, W. S. Gosset
derived the reference
distribution for t0 …
called the t distribution
• Tables of the t
distribution – see
textbook appendix
Chapter 2 Design & Analysis of Experiments 15
7E 2009 Montgomery
 The Two-Sample
p ((Pooled)) t-Test
• A value of t0 between
–2.101 and 2.101 is t0 = -2.20
consistent with
equality of means
• It is ppossible for the
means to be equal and
t0 to exceed either
2.101 or –2.101,, but it
would be a “rare
event” … leads to the
conclusion that the
means are different
• Could also use the
P-value approach
Chapter 2 Design & Analysis of Experiments 16
7E 2009 Montgomery
 The Two-Sample (Pooled) t-Test

t0 = -2.20

• The P-value is the area (probability) in the tails of the t-distribution beyond -2.20 + the
probability beyond +2.20 (it’s a two-sided test)
• The P-value
P value is a measure of how unusual the value of the test statistic is given that the null
hypothesis is true
• The P-value the risk of wrongly rejecting the null hypothesis of equal means (it measures
rareness of the event)
• The P-value in our pproblem is P = 0.042

Chapter 2 Design & Analysis of Experiments 17

7E 2009 Montgomery
 Computer Two-Sample t-Test Results

Chapter 2 Design & Analysis of Experiments 18

7E 2009 Montgomery
 Checking Assumptions –
The Normal Probability Plot

Chapter 2 Design & Analysis of Experiments 19

7E 2009 Montgomery
 Importance of the t-Test
• Provides an objective framework for simple
comparative
p experiments
p
• Could be used to test all relevant hypotheses
in a two
two-level
level factorial design,
design because all
of these hypotheses involve the mean
response at one “side”
side of the cube versus
the mean response at the opposite “side” of
the cube
Chapter 2 Design & Analysis of Experiments 20
7E 2009 Montgomery
 Confidence Intervals (See pg. 44)
• Hypothesis testing gives an objective statement
concerning the difference in means,
means but it doesn’t
specify “how different” they are
• General form of a confidence interval
L    U where P ( L    U )  1  
• The 100(1-
100(1 α)% confidence interval on the
difference in two means:
y1  y2  t / 2,n1  n2  2 S p (1/ n1 )  (1/ n2 )  1   2 
y1  y2  t / 2,n1  n2  2 S p (1/ n1 )  (1/ n2 )

Chapter 2 Design & Analysis of Experiments 21

7E 2009 Montgomery
Chapter 2 Design & Analysis of Experiments 22
7E 2009 Montgomery
Comparing a single mean to a specified value
• Some experiments involve comparing only one population
mean  to a specified
p y, 0, then the hypotheses
value,, say, yp
are
H0 :  = 0
H1 :  = 0
• If the population is normal with known variance or if the
population is nonnormal but the sample size is large
enough so that central limit theorem applies, then
hypothesis tested using a direct application of the normal
di t ib ti The
distribution. Th test
t t statistic
t ti ti is
i
y  0
Z 0 
 n
DOX 6E Montgomery 23
DOX 6E Montgomery 24
DOX 6E Montgomery 25
DOX 6E Montgomery 26
Figure 2.13 (p. 50)
The reference
Th f di
distribution
t ib ti (t with
ith 9 degrees
d off freedom)
f d ) forf the
th hardness
h d testing
t ti
problem.
DOX 6E Montgomery 27
DOX 6E Montgomery 28
DOX 6E Montgomery 29
 Question to be attempted

Chapter 2
No. 2.1, 2.2, 2.3, 2.5, 2.7, 2.9, 2.10, 2.11, 2.17

DOX 6E Montgomery 30