Why a Manager Needs to Know About Statistics

 To Know How to Properly Present Information

 To Know How to Draw Conclusions about Populations

Based on Sample Information

 To Know How to Improve Processes

 To Know How to Obtain Reliable Forecasts

 Why We Need Data
 To Provide Input to Survey
 To Provide Input to Study
 To Measure Performance of Ongoing Service or
Production Process
 To Evaluate Conformance to Standards
 To Assist in Formulating Alternative Courses of
Action
 To Satisfy Curiosity
 Binomial Probability Distribution

 ‘n’ Identical Trials

 E.g., 15 tosses of a coin; 10 light bulbs taken from
a warehouse
 2 Mutually Exclusive Outcomes on Each Trial
 E.g., Heads or tails in each toss of a coin;
defective or not defective light bulb
 Trials are Independent
 The outcome of one trial does not affect the
outcome of the other
 Binomial Probability Distribution
(continued)

 Constant Probability for Each Trial

 E.g., Probability of getting a tail is the same each
time we toss the coin
 2 Sampling Methods
 Infinite population without replacement
 Finite population with replacement
Binomial Probability Distribution Function

n!
P X   p 1  p 
X n X

X ! n  X  !
P  X  : probability of X successes given n and p
X : number of "successes" in sample  X  0,1, , n
p : the probability of each "success"
n : sample size Tails in 2 Tosses of Coin
X P(X)
0 1/4 = .25
1 2/4 = .50
2 1/4 = .25
Poisson Distribution

Siméon Poisson
 Poisson Distribution
 Discrete events (“successes”) occurring in a given
area of opportunity (“interval”)
 “Interval” can be time, length, surface area, etc.
 The probability of a “success” in a given “interval” is
the same for all the “intervals”
 The number of “successes” in one “interval” is
independent of the number of “successes” in other
“intervals”
 The probability of two or more “successes” occurring
in an “interval” approaches zero as the “interval”
becomes smaller
 E.g., # customers arriving in 15 minutes
 E.g., # defects per case of light bulbs
 The Normal Distribution

 “Bell Shaped” f(X)

 Symmetrical
 Mean, Median and
Mode are Equal X

 Inter-quartile Range
Equals 1.33 s Mean
Median
 Random Variable Mode
Has Infinite Range
 The Mathematical Model
1  (1/ 2)  X    / s 
f X  
2

e
2s
f  X  : density of random variable X
  3.14159; e  2.71828
 : population mean
s : population standard deviation
X : value of random variable    X   
 Many Normal Distributions
There are an Infinite Number of Normal Distributions

Varying the Parameters s and , We Obtain

Different Normal Distributions
 The Standardized Normal
Distribution
When X is normally distributed with a mean  and a
X   follows a
standard deviation s, Z 
s
standardized (normalized) normal distribution with a
mean 0 and a standard deviation 1.

f(Z) f(X) s

sZ 1

X
Z  0 Z
 Finding Probabilities
Probability is
the area under
the curve! P c  X  d   ?

f(X)

X
c d
 Standardizing Example
X  6.2  5
Z   0.12
s 10
Normal Distribution Standardized
Normal Distribution
s  10 sZ 1

6.2 X 0.12 Z
 5 Z  0
 Example:
P  2.9  X  7.1  .1664
X  2.9  5 X  7.1  5
Z   .21 Z   .21
s 10 s 10
Normal Distribution Standardized
Normal Distribution
s  10 .0832 sZ 1
.0832

2.9 7.1 X 0.21 0.21 Z

 5 Z  0
 Example:
P  2.9  X  7.1  .1664(continued)
Cumulative Standardized Normal
Distribution Table (Portion)
Z  0 sZ 1
Z .00 .01 .02
.5832
0.0 .5000 .5040 .5080
0.1 .5398 .5438 .5478

0.2 .5793 .5832 .5871 0

Z = 0.21
0.3 .6179 .6217 .6255
 Example:
P  2.9  X  7.1  .1664(continued)
Cumulative Standardized Normal
Distribution Table (Portion)
Z  0 sZ 1
Z .00 .01 .02 .4168
-0.3 .3821 .3783 .3745
-0.2 .4207 .4168 .4129

-0.1 .4602 .4562 .4522 0

Z = -0.21
0.0 .5000 .4960 .4920
 Recovering X Values for Known
Probabilities
Normal Distribution Standardized
Normal Distribution
s  10
.6179 sZ 1
.3821

X 0.30
 5 ? Z  0
Z

X    Zs  5  .3010  8
 More Examples of Normal
Distribution Using PHStat
A set of final exam grades was found to be normally
distributed with a mean of 73 and a standard deviation of 8.
What is the probability of getting a grade no higher than 91
on this exam?

X N  73,8 2
 P  X  91  ? s 8
Mean 73
Standard Deviation 8

Probability for X <= X

X Value 91
Z Value 2.25
  73 91
P(X<=91) 0.9877756
Z
0 2.25
 More Examples of Normal
Distribution Using PHStat
(continued)

What percentage of students scored between

65 and 89?
X N  73,82  P  65  X  89  ?

Probability for a Range

From X Value 65
To X Value 89
Z Value for 65 -1
Z Value for 89 2 X
P(X<=65) 0.1587
P(X<=89) 0.9772 65   73 89
P(65<=X<=89) 0.8186 Z
-1 0 2
 More Examples of Normal
Distribution Using PHStat
(continued)

The middle 50% of the students scored

between what two scores?
X N  73,82  P  a  X  b  .50
Find X and Z Given Cum. Pctage.
Cumulative Percentage 25.00%
Z Value -0.67449
X Value 67.60408 .25 .25

Find X and Z Given Cum. Pctage.

X
Cumulative Percentage 75.00% 67.6   73 78.4
Z Value 0.67449 Z
X Value 78.39592 -0.67 0 0.67
 Assessing Normality

 Not All Continuous Random Variables are

Normally Distributed
 It is Important to Evaluate How Well the Data
Set Seems to Be Adequately Approximated by
a Normal Distribution
 Assessing Normality (continued)
 Construct Charts
 For small- or moderate-sized data sets, do the
stem-and-leaf display and box-and-whisker plot
look symmetric?
 For large data sets, does the histogram or polygon
appear bell-shaped?
 Compute Descriptive Summary Measures
 Do the mean, median and mode have similar
values?
 Is the interquartile range approximately 1.33 s?
 Is the range approximately 6 s?
 Assessing Normality
(continued)

 Observe the Distribution of the Data Set

 Do approximately 2/3 of the observations lie
between mean  1 standard deviation?
 Do approximately 4/5 of the observations lie
between mean  1.28 standard deviations?
 Do approximately 19/20 of the observations lie
between mean  2 standard deviations?
 Evaluate Normal Probability Plot
 Do the points lie on or close to a straight line with
positive slope?
 Assessing Normality
(continued)

 Normal Probability Plot

 Arrange Data into Ordered Array
 Find Corresponding Standardized Normal Quantile
Values
 Plot the Pairs of Points with Observed Data Values
on the Vertical Axis and the Standardized Normal
Quantile Values on the Horizontal Axis
 Evaluate the Plot for Evidence of Linearity
 Normal Probability Plot
Left-Skewed Right-Skewed
90 90
X 60 X 60
30 Z 30 Z
-2 -1 0 1 2 -2 -1 0 1 2

Rectangular U-Shaped
90 90
X 60 X 60
30 Z 30 Z
-2 -1 0 1 2 -2 -1 0 1 2
 Unbiasedness (  X   )
f X 
Unbiased Biased

 X X
 Effect of Large Sample
For sampling with replacement:
As n increases, s X decreases
f X  Larger
sample size
Smaller
sample size

 X
 When the Population is Normal
Population Distribution
Central Tendency s  10
X  

Variation   50
s Sampling Distributions
sX 
n n4 n  16
sX 5 s X  2.5

 X  50 X
 When the Population is
Not Normal
Population Distribution
Central Tendency s  10
X  

Variation   50
s Sampling Distributions
sX 
n n4 n  30
sX 5 s X  1.8

 X  50 X
 Level of Significance and the
Rejection Region

H0:  3.5 a Critical

H1:  < 3.5 Value(s)

Rejection 0
Regions a
H0:   3.5
H1:  > 3.5
0
a/2
H0:  3.5
H1:   3.5
0
 One-Tail Z Test for Mean
( s Known)

 Assumptions
 Population is normally distributed
 If not normal, requires large samples
 Null hypothesis has  or  sign only
 s is known
 Z Test Statistic

X  X X 
Z 
sX s/ n
 Rejection Region
H0: 0 H0: 0
H1:  < 0 H1:  > 0
Reject H0 Reject H0
a a

0 Z 0 Z
Z must be significantly Small values of Z don’t
below 0 to reject H0 contradict H0 ; don’t
reject H0 !
 Reject and Do Not Reject
Regions
H 0 :   368
Reject

.05
Do Not Reject
X
 X   X  368 372.5
0 1.645 Z
1.5
0

H1 :   368
 Connection to Confidence
Intervals
For X  372.5, s  15 and n  25,
the 95% confidence interval is:
372.5  1.96  15 / 25    372.5  1.96  15 / 25
or
366.62    378.38
We are 95% confident that the population mean is
between 366.62 and 378.38.
If this interval contains the hypothesized mean (368),
we do not reject the null hypothesis.
It does. Do not reject.
 t Test: s Unknown
 Assumption
 Population is normally distributed
 If not normal, requires a large sample
 s is unknown
 t Test Statistic with n-1 Degrees of Freedom
X 
 t
S/ n
 Example Solution: One-Tail
H0: 368 Test Statistic:
H1:  368
X   372.5  368
a = 0.01 t   1.80
S 15
n = 36, df = 35 n 36
Critical Value: 2.4377
Reject Decision:
Do Not Reject at a = .01.
.01
Conclusion:
Insufficient Evidence that
0 2.437 t35
7
1.80
True Mean is More Than 368.
 p -Value Solution
(p-Value is between .025 and .05)  (a = 0.01)
Do Not Reject.

p-Value = [.025, .05]

Reject

a = 0.01

0 t35
1.80 2.4377
Test Statistic 1.80 is in the Do Not Reject Region
Potential Pitfalls and Ethical Issues
 Data Collection Method is Not Randomized to Reduce
Selection Biases
 Treatment of Human Subjects are Manipulated Without
Informed Consent
 Data Snooping is Used to Choose between One-Tail
and Two-Tail Tests, and to Determine the Level of
Significance
Potential Pitfalls and Ethical Issues (continued)

 Data Cleansing is Practiced to Hide

Observations that do not Support a Stated
Hypothesis
 Fail to Report Pertinent Findings
 One-Way Analysis of Variance
F Test
 Evaluate the Difference Among the Mean Responses of 2
or More (c ) Populations
 E.g., Several types of tires, oven temperature settings

 Assumptions
 Samples are randomly and independently drawn
 This condition must be met

 Populations are normally distributed

 F Test is robust to moderate departure from normality

 Populations have equal variances

 Less sensitive to this requirement when samples are of equal

size from each population

Why ANOVA?
 Could Compare the Means One by One using Z or t Tests for
Difference of Means
 Each Z or t Test Contains Type I Error
 The Total Type I Error with k Pairs of Means is 1- (1 - a) k

 E.g., If there are 5 means and use a = .05

 Must perform 10 comparisons

 Type I Error is 1 – (.95)

10 = .40

 40% of the time you will reject the null hypothesis of equal means

in favor of the alternative when the null is true!

Hypotheses of One-Way ANOVA
 H 0 : 1  2   c
 All population means are equal
 No treatment effect (no variation in means among groups)

 H1 : Not all i are the same

 At least one population mean is different (others may be the
same!)
 There is a treatment effect
 Does not mean that all population means are different
 One-Way ANOVA
(No Treatment Effect)
H 0 : 1  2   c
H1 : Not all i are the same
The Null
Hypothesis is
True

1   2  3
 One-Way ANOVA
(Treatment Effect Present)
H 0 : 1  2   c
H1 : Not all i are the same The Null
Hypothesis is
NOT True

1   2  3 1  2  3
 One-Way ANOVA
(Partition of Total Variation)
Total Variation SST

Variation Due to Variation Due to Random

= Group SSA + Sampling SSW
Commonly referred to as: Commonly referred to as:
 Among Group Variation  Within Group Variation
 Sum of Squares Among  Sum of Squares Within
 Sum of Squares Between  Sum of Squares Error
 Sum of Squares Model  Sum of Squares Unexplained
 Sum of Squares Explained
 Sum of Squares Treatment
 Total Variation
c nj

SST   ( X ij  X ) 2

j 1 i 1

X ij : the i -th observation in group j

n j : the number of observations in group j
n : the total number of observations in all groups
c : the number of groups
c nj

 X
j 1 i 1
ij

X  the overall or grand mean

n
 Total Variation (continued)

  X   X 
2 2 2
SST  X 11  X 21 X nc c X
Response, X

Group 1 Group 2 Group 3

 One-Way ANOVA
F Test Statistic
 Test Statistic
MSA
F


MSW
 MSA is mean squares among
 MSW is mean squares within
 Degrees of Freedom
 df1  c 1

df 2  n  c
 One-Way ANOVA
Summary Table

Degrees Mean
Source of Sum of F
of Squares
Variation Squares Statistic
Freedom (Variance)
Among MSA =
c–1 SSA MSA/MSW
(Factor) SSA/(c – 1 )
Within MSW =
n–c SSW
(Error) SSW/(n – c )
SST =
Total n–1
SSA + SSW
Features of One-Way ANOVA F Statistic
 The F Statistic is the Ratio of the Among Estimate
of Variance and the Within Estimate of Variance
 The ratio must always be positive
 df1 = c -1 will typically be small
 df2 = n - c will typically be large
 The Ratio Should be Close to 1 if the Null is True
 If the Null Hypothesis is False
 The numerator should be greater than the denominator
 The ratio should be larger than 1
 One-Way ANOVA F Test
Example
As Production Manager, you
want to see if 3 filling machines Machine1 Machine2 Machine3
have different mean filling 25.40 23.40 20.00
times. 26.31 21.80 22.20
You assign 15 similarly trained 24.10 23.50 19.75
& experienced workers, 5 per 23.74 22.75 20.60
machine, to the machines. 25.10 21.60 20.40
At the .05 significance level, is
there a difference in mean
filling times?
 One-Way ANOVA
Example: Scatter Diagram

Machine1 Machine2 Machine3 27

25.40 23.40 20.00 26 •
26.31 21.80 22.20 •
25 • X1
24.10 23.50 19.75 24 •
23.74 22.75 20.60 • ••
25.10 21.60 20.40
23
• X2 •
X
22 ••
21
X 1  24.93 X 2  22.61 •• X3
20 ••
X 3  20.59 X  22.71 19
 Residual Analysis for
Homoscedasticity

Y Y

X X
SR SR

X X

Heteroscedasticity Homoscedasticity
 One-Way ANOVA Example
Computations
Machine1 Machine2 Machine3 X 1  24.93 nj  5
25.40 23.40 20.00
26.31 21.80 22.20 X 2  22.61 c3
24.10 23.50 19.75 X 3  20.59 n  15
23.74 22.75 20.60
25.10 21.60 20.40 X  22.71

SSA  5  24.93  22.71   22.61  22.71   20.59  22.71 

 2 2 2
 
 47.164
SSW  4.2592  3.112  3.682  11.0532
MSA  SSA /(c -1)  47.16 / 2  23.5820
MSW  SSW /( n - c)  11.0532 /12  .9211
 One-Way ANOVA Example
Solution
Test Statistic:
H0: 1 = 2 = 3
H1: Not All Equal MSA 23.5820
a = .05 F   25.6
df1= 2 df2 = 12 MSW .9211

Critical Value(s):
Decision:
Reject at a = 0.05.
a = 0.05 Conclusion:
There is evidence that at
least one  i differs from
0 3.89 F the rest.
 Two-Way ANOVA

 Examines the Effect of:

 Two factors on the dependent variable
 E.g., Percent carbonation and line speed on soft
drink bottling process
 Interaction between the different levels of these
two factors
 E.g., Does the effect of one particular percentage of
carbonation depend on which level the line speed is
set?
 Two-Way ANOVA (continued)

 Assumptions
 Normality
 Populations are normally distributed
 Homogeneity of Variance
 Populations have equal variances
 Independence of Errors
 Independent random samples are drawn
 Two-Way ANOVA
Total Variation Partitioning

Variation Due to SSA +

Factor A d.f.= r-1

SSB +
Variation Due to
d.f.= c-1
Total Variation Factor B

Variation Due to SSAB +

SST
= Interaction d.f.= (r-1)(c-1)
d.f.= n-1
Variation Due to SSE
Random Sampling d.f.= rc(n’-1)
 Two-Way ANOVA
Total Variation Partitioning

r  the number of levels of factor A

c  the number of levels of factor B
n  the number of values (replications) for each cell
'

n  the total number of observations in the experiment

X ijk  the value of the k -th observation for level i of
factor A and level j of factor B
 Features of Two-Way ANOVA
F Test
 Degrees of Freedom Always Add Up
 rcn’-1=rc(n’-1)+(c-1)+(r-1)+(c-1)(r-1)
 Total=Error+Column+Row+Interaction
 The Denominator of the F Test is Always the
Same but the Numerator is Different
 The Sums of Squares Always Add Up
 Total=Error+Column+Row+Interaction
 Purpose of Regression Analysis

 Regression Analysis is Used Primarily to Model

Causality and Provide Prediction
 Predict the values of a dependent (response)
variable based on values of at least one
independent (explanatory) variable
 Explain the effect of the independent variables on
the dependent variable
 Types of Regression Models
Positive Linear Relationship Relationship NOT Linear

Negative Linear Relationship No Relationship

 Simple Linear Regression Model

 Relationship between Variables is Described

by a Linear Function
 The Change of One Variable Causes the Other
Variable to Change
 A Dependency of One Variable on the Other
 Simple Linear Regression Model
(continued)

Population regression line is a straight line that

describes the dependence of the average value
(conditional mean) of one variable on the other
Population Random
Population Error
Slope
Y Intercept Coefficient

Yi      X i   i
Dependent
Population Independent
(Response)
Variable Regression  Y |X (Explanatory)
Line Variable
(Conditional Mean)
 Simple Linear Regression Model
(continued)

Y (Observed Value of Y) = Yi      X i   i

 i = Random Error 

Y | X      X i
 (Conditional Mean)
X
Observed Value of Y
 Linear Regression Equation
Sample regression line provides an estimate of
the population regression line as well as a
predicted value of Y
Sample
Sample
Slope
Y Intercept

Yi  b0  b1 X i  ei
Coefficient

Residual

Ŷ  b0  b1 X (Fitted Regression Line, Predicted Value)

Simple Regression Equation
 Linear Regression Equation
(continued)

 b0 and b1 are obtained by finding the values

of b0 and b that minimize the sum of the
1
squared residuals

 
n n
  ei2
2
Yi  Yˆi
i 1 i 1

 b0 provides an estimate of  
 b1 provides an estimate of 
 Linear Regression Equation
(continued)

Yi  b0  b1 X i  ei Yi      X i   i
Y b1
i 
ei
Y | X      X i
 Yˆi  b0  b1 X i
b0 X
Observed Value
 Interpretation of the Slope
and Intercept

   E Y | X  0 is the average value of Y

when the value of X is zero

change in E Y | X 
 1  measures the
change in X
change in the average value of Y as a result
of a one-unit change in X
 Interpretation of the Slope
and Intercept (continued)

 b  Eˆ Y | X  0  is the estimated average

value of Y when the value of X is zero

change in Eˆ Y | X 
 b1  is the estimated
change in X
change in the average value of Y as a result
of a one-unit change in X
 Simple Linear Regression:
Example
You wish to examine Annual
Store Square Sales
the linear dependency Feet ($1000)
of the annual sales of 1 1,726 3,681
produce stores on their 2 1,542 3,395
sizes in square footage. 3 2,816 6,653
Sample data for 7 4 5,555 9,543
stores were obtained. 5 1,292 3,318
Find the equation of 6 2,208 5,563
the straight line that 7 1,313 3,760
fits the data best.
 Scatter Diagram: Example
12000
Annua l Sa le s ($000)

10000

8000

6000

4000

2000

0
0 1000 2000 3000 4000 5000 6000

S q u a re F e e t
Excel Output
 Simple Linear Regression
Equation: Example

Yˆi  b0  b1 X i
 1636.415  1.487 X i

From Excel Printout:

C o e ffi c i e n ts
I n te r c e p t 1 6 3 6 .4 1 4 7 2 6
X V a ria b le 1 1 .4 8 6 6 3 3 6 5 7
 Interpretation of Results:
Example

Yˆi  1636.415  1.487 X i

The slope of 1.487 means that for each increase of
one unit in X, we predict the average of Y to
increase by an estimated 1.487 units.

The equation estimates that for each increase of 1

square foot in the size of the store, the expected
annual sales are predicted to increase by $1487.
 Residual Analysis

 Purposes
 Examine linearity
 Evaluate violations of assumptions
 Graphical Analysis of Residuals
 Plot residuals vs. X and time
 Residual Analysis for Linearity
Y Y

X
e X
e
X
X

Not Linear
 Linear
 Residual Analysis for
Homoscedasticity

Y Y

X X
SR SR

X X

Heteroscedasticity Homoscedasticity
 Residual Analysis: Excel Output
for Produce Stores Example
Observation Predicted Y Residuals
1 4202.344417 -521.3444173
2 3928.803824 -533.8038245
3 5822.775103 830.2248971
Excel Output 4 9894.664688 -351.6646882
5 3557.14541 -239.1454103
6 4918.90184 644.0981603
7 3588.364717 171.6352829
Residual Plot

0 1000 2000 3000 4000 5000 6000

Square Feet
 Residual Analysis for
Independence
 The Durbin-Watson Statistic
 Used when data is collected over time to detect
autocorrelation (residuals in one time period are
related to residuals in another period)
 Measures violation of independence assumption
n

 i i1
( e  e ) 2
Should be close to 2.
D i 2
n If not, examine the model
e
i 1
2
i for autocorrelation.
 Sample Observations from
Various r Values
Y Y Y

X X X
r = -1 r = -.6 r=0
Y Y

X X
r = .6 r=1
 Features of rand r

 Unit Free
 Range between -1 and 1
 The Closer to -1, the Stronger the Negative
Linear Relationship
 The Closer to 1, the Stronger the Positive
Linear Relationship
 The Closer to 0, the Weaker the Linear
Relationship
 t Test for Correlation
 Hypotheses
 H0: r = 0 (no correlation)
 H1: r  0 (correlation)
 Test Statistic
rr
t where
  r 2

n2
n

 X i  X Yi  Y 
r  r2  i 1
n n

 X X  Y  Y 
2 2
i i
i 1 i 1
 Example: Produce Stores
From Excel Printout r
Is there any R e g r e ssi o n S ta ti sti c s
M u lt ip le R 0 .9 7 0 5 5 7 2
evidence of linear R S q u a re 0 .9 4 1 9 8 1 2 9
relationship between A d ju s t e d R S q u a re 0 . 9 3 0 3 7 7 5 4
annual sales of a S t a n d a rd E rro r 6 1 1 .7 5 1 5 1 7
store and its square O b s e rva t io n s 7
footage at .05 level
of significance? H0: r = 0 (no association)
H1: r  0 (association)
a  .05
df  7 - 2 = 5
 Example: Produce Stores
Solution

Decision:
rr .9706 Reject H0.
t   9.0099
 r 2
1  .9420
Conclusion:
n2 5
There is evidence of a
Critical Value(s): linear relationship at 5%
level of significance.
Reject Reject
The value of the t statistic is
.025 .025 exactly the same as the t
statistic value for test on the
-2.5706 0 2.5706 slope coefficient.
 Estimation of Mean Values
Confidence Interval Estimate for Y | X  X i
:
The Mean of Y Given a Particular Xi
Size of interval varies according
Standard error to distance away from mean, X
of the estimate
1 (Xi  X ) 2
Yˆi  tn 2 SYX  n
(Xi  X )
n 2
t value from table
i 1
with df=n-2
 Prediction of Individual Values
Prediction Interval for Individual Response
Yi at a Particular Xi

Addition of 1 increases width of interval

from that for the mean of Y

1 (Xi  X ) 2
Yˆi  tn 2 SYX 1  n
(Xi  X )
n 2

i 1
 Interval Estimates for Different
Values of X
Confidence
Prediction Interval Interval for the
for a Individual Yi Mean of Y
Y

X
X a given X
 Example: Produce Stores
Data for 7 Stores:
Annual
Store Square Sales Consider a store
Feet ($000) with 2000 square
1 1,726 3,681 feet.
2 1,542 3,395
3 2,816 6,653
4 5,555 9,543 Regression Model Obtained:
5 1,292 3,318 
6 2,208 5,563
Yi = 1636.415 +1.487Xi
7 1,313 3,760
 Estimation of Mean Values:
Example
Confidence Interval Estimate for Y | X  X i

Find the 95% confidence interval for the average annual

sales for stores of 2,000 square feet.

Predicted Sales Yi = 1636.415 +1.487Xi = 4610.45 ($000)
X = 2350.29 SYX = 611.75 tn-2 = t5 = 2.5706

1 ( X i  X )2
Yˆi  tn 2 SYX  n  4610.45  612.66
(Xi  X )
n 2

i 1

3997.02  Y |X  X i  5222.34
 What is a Time Series?

 Numerical Data Obtained at Regular Time

Intervals
 The Time Intervals Can Be Annually,
Quarterly, Daily, Hourly, Etc.
 Example:
Year: 1994 1995 1996 1997 1998
Sales: 75.3 74.2 78.5 79.7 80.2
 Time-Series Components

Trend Cyclical

Time-Series

Seasonal Random
 Trend Component

 Overall Upward or Downward Movement

 Data Taken Over a Period of Years

Sales

Time
 Cyclical Component

 Upward or Downward Swings

 May Vary in Length
 Usually Lasts 2 - 10 Years

Sales
 Seasonal Component

 Upward or Downward Swings

 Regular Patterns
 Observed Within 1 Year
Sales

Summer

Winter
Spring Fall
Time (Monthly or Quarterly)
 Random or Irregular Component

 Erratic, Nonsystematic, Random, “Residual”

Fluctuations
 Due to Random Variations of
 Nature
 Accidents
 Short Duration and Non-Repeating
 Example: Quarterly Retail Sales
with Seasonal Components
Quarterly with Seasonal Components

25

20

15
Sales

10

0
0 5 10 15 20 25 30 35

Time
 Example: Quarterly Retail Sales with
Seasonal Components Removed

Quarterly without Seasonal Com ponents

25

20

15
Sales

Y(t)

10

0
0 5 10 15 20 25 30 35
Tim e
 Multiplicative Time-Series Model

 Used Primarily for Forecasting

 Observed Value in Time Series is the Product
of Components
 For Annual Data:
Ti = Trend
Yi  Ti Ci I i
 For Quarterly or Monthly Data: Ci = Cyclical

Yi  Ti Si Ci I i Ii = Irregular
Si = Seasonal
 Moving Averages

 Used for Smoothing

 Series of Arithmetic Means Over Time
 Result Dependent Upon Choice of L (Length of
Period for Computing Means)
 To Smooth Out Cyclical Component, L Should
Be Multiple of the Estimated Average Length
of the Cycle
 For Annual Time Series, L Should Be Odd
 Moving Averages
(continued)

 Example: 3-Year Moving Average

Y1  Y2  Y3
 First average: MA(3) 
3
Y2  Y3  Y4
 Second average: MA(3) 
3
 Moving Average Example
John is a building contractor with a record of a total
of 24 single family homes constructed over a 6-year
period. Provide John with a 3-year moving average
graph. Year Units Moving
Ave
1994 2 NA
1995 5 3
1996 2 3
1997 2 3.67
1998 7 5
1999 6 NA
 Moving Average Example
Solution

Year Response Moving

Ave Sales
L=3
1994 2 NA 8
1995 5 3 6
1996 2 3 4
1997 2 3.67 2
1998 7 5 0
94 95 96 97 98 99
1999 6 NA
No MA for the first and last (L-1)/2 years
 Exponential Weight: Example
Ei  WYi  (1  W ) Ei 1
Year Response Smoothing Value Forecast
(W = .2, (1-W)=.8)
1994 2 2 NA
1995 5 (.2)(5) + (.8)(2) = 2.6 2
1996 2 (.2)(2) + (.8)(2.6) = 2.48 2.6
1997 2 (.2)(2) + (.8)(2.48) = 2.384 2.48
1998 7 (.2)(7) + (.8)(2.384) = 3.307 2.384
1999 6 (.2)(6) + (.8)(3.307) = 3.846 3.307
 Exponential Weight:
Example Graph

Sales
8
Data
6

2 Smoothed

0
94 95 96 97 98 99 Year
 Linear Trend Model
Use the method of least squares to obtain the
linear trend forecasting equation:

Year Coded X Sales (Y)

95 0 2
96 1 5
Yˆi  b0  b1 X i
97 2 2
98 3 2
99 4 7
00 5 6
 Linear Trend Model
(continued)
Linear trend forecasting equation:
Yˆi  b0  b1 X i  2.143  .743X i
8
Excel Output
7
Co efficien ts
6
I n te r c e p t 2.14285714
5
X V a ria b le 1 0 .7 4 2 8 5 7 1 4
Sales

4
3 Projected to
2 year 2001
1
0
0 1 2 3 4 5 6
X
 The Quadratic Trend Model
Use the method of least squares to obtain
the quadratic trend forecasting equation:
Year Coded X Sales (Y)
95 0 2
96
97
1
2
5
2
ˆ
Yi  b0  b1 X i  b2 X i
2

98 3 2
99 4 7
00 5 6
 The Quadratic Trend Model
(continued)

ˆ
Yi  b0  b1 X i  b2 X i  2.857  .33X i  .214 X i
2 2

Excel Output 8
Coefficients 7

In te rce p t 2.85714286 6

X V a ria b le 1 -0.3285714 Sales 5

4
Projected to
X V a ria b le 2 0.21428571 year 2001
3
2
1
0
0 1 2 3 4 5 6
X
 The Exponential Trend Model
After taking the logarithms, use the method of least
squares to get the forecasting equation:
ˆ
Yi  b0b1
Xi
or log Yˆi  log b0  X1 log b1
Year Coded X Sales (Y) C o e f f ic ie n t s
In t e r c e p t 0 .3 3 5 8 3 7 9 5
95 0 2
X V a ria b le 10 . 0 8 0 6 8 5 4 4
96 1 5
97 2 2 Excel Output of Values in Logs
98 3 2 a n t ilo g (. 3 3 5 8 3 7 9 5 ) = 2.17
a n t ilo g (. 0 8 0 6 8 5 4 4 ) = 1.2
99 4 7
00 5 6 ˆ
Yi  (2.17)(1.2) Xi
 Autoregressive Modeling

 Used for Forecasting

 Takes Advantage of Autocorrelation
 1st order - correlation between consecutive values
 2nd order - correlation between values 2 periods
apart
 Autoregressive Model for p-th Order:
Yi  A0  AY
1 i 1  A2Yi  2   ApYi  p   i
Random
Error
 Autoregressive Model:
Example
The Office Concept Corp. has acquired a number of
office units (in thousands of square feet) over the
last 8 years. Develop the 2nd order autoregressive
model.
Year Units
93 4
94 3
95 2
96 3
97 2
98 2
99 4
00 6
 Autoregressive Model:
Example Solution
Develop the 2nd order Year Yi Yi-1 Yi-2
table 93 4 --- ---
94 3 4 ---
Use Excel to estimate a 95 2 3 4
regression model 96 3 2 3
Excel Output 97 2 3 2
Coefficients 98 2 2 3
I n te rc e p t 3.5 99 4 2 2
X V a ri a b l e 1 0.8125 00 6 4 2
X V a ri a b l e 2 -0 . 9 3 7 5

Yˆi  3.5  .8125Yi 1  .9375Yi 2

 Autoregressive Model Example:
Forecasting

Use the 2nd order model to forecast number

of units for 2001:
Yˆi  3.5  .8125Yi 1  .9375Yi 2
Yˆ2001  3.5  .8125Y2000  .9375Y1999
 3.5  .8125  6  .9375  4
 4.625
 Autoregressive Modeling Steps

1. Choose p : Note that df = n - 2p - 1

2. Form a Series of “Lag Predictor” Variables
Yi-1 , Yi-2 , … ,Yi-p
3. Use Excel to Run Regression Model Using All
p Variables
4. Test Significance of Ap
 If null hypothesis rejected, this model is selected
 If null hypothesis not rejected, decrease p by 1
and repeat