Probability and Statistics

LECTURE 8
HYPOTHESIS TESTING
PART II

Adapted from http://www.prenhall.com/mcclave

9-1

Outline

1. Explain the p-value concept

2. Explain the t-test for mean
3. Explain the z-test of proportion

9-2

p-Value

1.Probability of Obtaining a Test Statistic

More Extreme (or than Actual
Sample Statistic (Computed From The
Sample Data) Given H0 Is True
2.Used to Make Rejection Decision
 If p-Value < a, Reject H0
 If p-Value  a, Do Not Reject H0

9-3
 One-Tailed Z Test
p-Value Example
Does an average box of
cereal contain more than 365
grams of cereal? A random
sample of 25 boxes
showedX = 372.5. The
company has specified  to
be 15 grams. Find the p-
Value. Assume normal 368 gm.
population. Use significance
level of 0.05
9-4

One-Tailed Z Test
p-Value Solution
The sample mean
H0:  = 365
observed in our
Ha:   365
sample

x
Should we think that Ho is true
and our observed sample
mean (372.5) just happen by
chance?
9-5

One-Tailed Z Test
p-Value Solution

Use
alternative
hypothesis
to find
direction
Z
Z value of sample
statistic =2.5
9-6
 One-Tailed Z Test
p-Value Solution
p-Value = 0.0062 is small; therefore our
observed sample mean (372.5) is unlikely to
happen if Ho is true (chance is too low). But
remember that this value (372.5) did happen.
Therefore, evidence supports that Ho is false.
We compare p-value and a to make decision.
As p-value =0.0062 < a = 0.05, we say that
the test is significant and there is enough
evidence at a = 0.05 to infer that Ha is true.
9-7

Two-Tailed Z Test
p-Value Example
Does an average box of
cereal contain 368 grams
of cereal? A random
sample of 25 boxes
showedX = 372.5. The
company has specified 
to be 15 grams. Find the
p-Value. Assume normal 368 gm.
population. Use
significance level of 0.05.
9-8

Two-Tailed Z Test
p-Value Solution
H0:  = 368
Ha:   368

Z value of sample
9-9 statistic (observed)
 Two-Tailed Z Test
p-Value Solution
p-value is P(Z  -1.50 or Z  1.50) = .1336

Z value of sample
statistic
9 - 10

Two-Tailed Z Test
p-Value Solution
p-Value = .1336  a = .05
Do not reject Ho

9 - 11

One population tests

One
population

Mean Proportion

Z Test
 Z Test T Test

9 - 12
 T Test for 1 Population Mean
( Unknown)
1.Assumptions
 Population Standard Deviation Is Unknown
 Population Is Normally Distributed
2. t Test Statistic

9 - 13

T Test
Finding Critical T Values
Given: n = 3; a = .10; two tailed t test.

9 - 14

Two-Tailed t Test
Finding Critical t Values
Given: n = 3; a = .10 Critical Values of t Table
 (Portion)
df = n - 1 = 2 


a /2 = .05

a /2 = .05
9 - 15 
 Two-Tailed t Test
Example
Does an average box of
cereal contain 368
grams of cereal? A
random sample of 36
boxes had a mean of
372.5 & a standard
deviation of 12 grams.
Test at the .05 level.
368 gm.
Assume normal
population.
9 - 16

Two-Tailed t Test
Solution
Test Statistic:
H0:  = 368
Ha:   368
a = .05
df = 36 - 1 = 35
Critical Value(s): Decision:
Reject Ho at a = .05
Conclusion:
There is enough
evidence to conclude
9 - 17 that…

One-Tailed t Test
Example
Is the average capacity of
batteries at least 140
ampere-hours? A
random sample of 20
batteries had a mean of
138.47 & a standard
deviation of 2.66.
Assume a normal
distribution. Test at the
.05 level.
9 - 18
 One-Tailed t Test
Solution
Test Statistic:
H0:  = 140
Ha:  > 140
a = .05
df = 20 - 1 = 19
Critical Value(s): Decision:
Do not reject Ho at a = .05
Reject H 0
a0.05 Conclusion:
There is not enough
0 1.729 t evidence to conclude that …
9 - 19

Robustness of one-sample
T-procedures (Testing & CI)

• Results of one-sample t procedures

are most reliable when population is
normal (other conditions met)
• Procedures are robust if probability
calculations required are insensitive to
violations of assumptions

9 - 20

Robustness of one-sample
T-procedures (Testing & CI)

9 - 21
Practical guidelines for one-sample
T-procedures (Testing & CI)
• n < 15: Use t procedures if the data are
close to normal. If data are clearly non-
normal or if outliers are present, do not
use t.
• n  15: t can be used except in
presence of outliers or strong
skewness
• Large samples (n  40): t can be used
even for clearly skewed distributions
9 - 22without outliers

Practical guidelines for one-sample

T-procedures (Testing & CI)

Before using t procedures, use (a

combination of) appropriate graphical
methods (e.g. Q-Q plot, boxplot,
histogram, stem and leaf display,) to
check for skewness and outliers.

9 - 23

Z Test of Proportion

9 - 24
 Qualitative Data

1.Qualitative Random Variables Yield

Responses That Classify
 e.g., Gender (Male, Female)
2.Measurement Reflects # in Category
3.Nominal or Ordinal Scale
4.Examples
 Do You Own Savings Bonds?
 Do You Live On-Campus or Off-Campus?
9 - 25

Proportions

1.Involve Qualitative Variables

2.Fraction or % of Population in a Category
3.If Two Qualitative Outcomes, Binomial
Distribution
 Possess or Don’t Possess Characteristic

9 - 26

Proportions

1.Involve Qualitative Variables

2.Fraction or % of Population in a Category
3.If Two Qualitative Outcomes, Binomial
Distribution
Possess or Don’t Possess Characteristic
4.Sample Proportion (p)^

9 - 27
 Sampling Distribution
of Proportion
1. Approximated by
Normal Distribution Sampling Distribution
 and ^
P(P )
.3
.2
.1
^
.0 P
2. Mean .0 .2 .4 .6 .8 1.0

3. Standard Error
p  (1 p )
 p^  where p = Population Proportion
n
9 - 28

Standardizing Sampling
Distribution of Proportion
p^   p^ ^
p p
Z 
 p^ p  (1  p )
Sampling n Standardized
Distribution Normal Distribution
 P^ z= 1

^
 P^ P  Z= 0 Z
9 - 29

One-Sample Z Test
for Proportion
1.Assumptions
 Two Categorical Outcomes
 Population Follows Binomial Distribution
 Normal Approximation To The Sampling
Distribution Of Can Be Used
 and
[ p0 is the hypothesized population
proportion ]

9 - 30
 One-Sample Z Test
for Proportion

2.Z-test statistic for proportion

Hypothesized
population proportion

9 - 31

One-Proportion Z Test
Example
The present packaging
system produces 10%
defective cereal boxes.
Using a new system, a
random sample of 200
boxes had11 defects.
Does the new system
produce fewer defects?
Test at the .05 level.
9 - 32

One-Proportion Z Test
Solution
Test Statistic:
H0: p = .10
Ha: p < .10
a = .05
n = 200
Decision:
Critical Value(s):
Reject Ho at a = .05
Conclusion:
There is enough evidence
new system < 10% defective
9 - 33
 Conclusion

1. Explain the p-value concept

2. Explain the t-test for mean
3. Explain the z-test of proportion

9 - 34