Sei sulla pagina 1di 6

TOPIC 8

One Population Hypothesis Testing 6
Steps of Hypothesis Testing Procedure
S.1
State null & alternative hypotheses
S.2
Select the significance level
S.3
Formulate a decision rule
S.4
Compute the test statistic
S.5
Make a conclusion of the test
S.6
Make the related decision    Three possible statistical hypotheses to test a population mean one-tail test

(directional test)

H 0 : μ≤μ 0 H 1 : μ > μ 0

H 0 : π≤π 0 H 1 : π > π 0

H 0 : μ≥μ 0 H 1 : μ < μ 0

H 0 : π≥π 0 H 1 : π < π 0

two-tail test (Nondirectional test)

H 0 : μ = μ 0 H 1 : μ≠μ 0

H 0 : π = π 0 H 1 : π≠π 0 Three possible statistical hypotheses to test a population proportion   Introduction

Hypothesis testing is a procedure based on sample data and probability theory to determine whether there is enough statistical evidence in favor of a certain belief about a parameter.   S.1 State null & alternative hypotheses

A pair of competing statistical hypotheses consists of

Null Hypothesis (H 0 ): A statement about the value of a population parameter.

Alternative Hypothesis (H 1 ): A statement that is accepted if the sample data provide evidence that the null hypothesis is false. EXAMPLE 1:

For each of the following pairs of null and alternative hypotheses, determine whether the pair would be appropriate for a hypothesis test. If a pair is deemed inappropriate, explain why.

a) H 0 : μ > 90 ; H 1 : μ ≤ 90

b) H 0 : μ

75 ; H 1 : μ > 85

d) H 0 : π ≥ 0.48 ; H 1 : π > 0.52

e) H 0 : p 0.65 ; H 1 : π > 0.65

c) H 0 :

x

= 58 ; H

1

: x

58

1 EXAMPLE 2:

For each of the following statements, formulate appropriate null and alternative hypotheses.

a) The average college student spends no more than RM500 per semester at the university’s bookstore.

b) The proportion adult drinks 2 cups of coffee per day is more than 0.62.

c) The average SAT score for entering freshmen is at least 1200.

d) The proportion candidate passing on the qualifying exam differs from 0.33.  S.3 Formulate a decision rule Decision Rule : A statement of the specific conditions under which the H 0 is rejected.

Critical Value :

- the dividing point between the region where the H 0 is rejected and the region where it is not rejected  One-tail Test:
(right tail)
Level of Significance, α=0.01
Sampling distribution of the
test statistic Z
Decision Rule :
Reject H 0 if
Z > 2.3263
Rejection
Region
(α = 0.01)

2.3263

Critical Value: Z α  S.2 Select the significance level significance level (α) measure the prob. of rejecting H 0 when it is true

in practice, significance level is set at 1% (0.01), 5% (0.05) and 10% (0.10) Critical Value :

- it is determined by the type of alternative hypothesis, sampling distribution of the test statistic and level of significance (α)

Rejection Region : The rejection region is a range of values such that if the test statistic falls into that range, the H 0 is rejected in favor of the H 1 .  One-tail Test:
(left tail)
Level of Significance, α=0.10
Sampling distribution of the
test statistic Z
Decision Rule :
Rejection
Region
(α = 0.10)
Reject H 0 if
Z < -1.2816

-1.2816

Critical Value: -Z α Two-tail Test:

Level of Significance, α=0.05 Sampling distribution of the test statistic Z Decision Rule for Z-Test and t-Test Reject H 0 if
Alternative
Hypothesis
Rejection
Region
(α/2 = 0.025)
Rejection
Region
(α/2 = 0.025)
Z -Test
t -Test
> Z
> Z α
t
> t α , n-1
< Z
< -Z α
t
< -t α , n-1
-1.96
1.96
Z
or
t
≠ > Z α/2
> t α/2 , n-1 or
Critical Value: -Z α/2
Critical Value: Z α/2
Z
t
< -Z α/2
< -t α/2, n-1
Decision Rule:
Reject H 0 if Z > 1.96 or Z < -1.96
S.4 Compute the test statistic
Type of Inference
Test Statistic
Test Statistic
Testing the μ when
σ 2 known
x
−μ
Z =
σ
n
A value, determined from sample
information, used to determine whether
or not to reject the H 0 .
Testing the μ when
σ 2 unknown
x −μ
t =
s
n
p −π
Z =
Testing the π
π −π
(1
)
n
S.5 Make a conclusion of the test
If the test statistic falls in the rejection
region, H 0 is rejected.
If the test statistic does not fall in the
rejection region, H 0 is not rejected.
Failure to reject a null hypothesis DOES
NOT constitute proof that it is true.
Therefore, we never say that we accept
the null hypothesis   S.6 Make the related decision If H 0 is rejected, we conclude that there is enough statistical evidence to infer that the H is true .

1

If H 0 is not rejected, we conclude that there is not enough statistical evidence to infer that the H 1 is true. Types of Errors

2 types of errors may occur when deciding whether to reject H 0 based on the stat. value.

a) Type I error (α):

Reject H 0 when it is true.

b) Type II error (β):

Do not reject H 0 when it is false.

The Type I error can be directly controlled. It is actually the level of significance. EXAMPLE 3:

Following a major earthquake, the city engineer must determine whether the stadium is structurally sound for an upcoming athletic event. If the null hypothesis is the stadium is structurally sound”, and the alternative hypothesis is “the stadium is not structurally sound”, which type of error (Type I or Type II) would the engineer least like to commit?  Do not reject H 0
Reject H 0
H 0 is true
Correct decision
Type I error
(α)
H 0 is false
Type II error
(β)
Correct decision
(1 – β)

Power of a test (1-β):

It represents the probability of rejecting the null hypothesis when it is false. Testing the Population Mean when the Population Variance is Known

EXAMPLE 4:

When a machine is set properly, it produces nails having a mean length of 1.325 inches, with a standard deviation of 0.0396 inches. For a sample of 80 nails, the mean length is 1.3229 inches. Using the 0.05 level of significance, examine whether the machine is adjusted properly. Testing Hypotheses and Confidence Interval Estimators

Confidence interval estimator can be used to conduct the two-sided alternative tests of hypotheses at the α level of significance

If the hypothesized value under null hypothesis falls into the (1–α)100% confidence interval estimate, H 0 is not rejected Refer to EXAMPLE 4:

The 95% confidence interval estimate of μ is

LCL = 1.3142, UCL = 1.3316

We have 95% confident that the population mean length is somewhere between 1.3142 and 1.3316 inches. Since μ = 1.325 falls within the interval estimate, we conclude at 5% that the mean length do not differ from 1.325 inches.

The conclusion is the same as that reached at the α level of significance Assuming population is approximately normal distributed, and using the 0.05 level of significance, is the agent’s suspicion confirmed? Testing the Population Proportion

EXAMPLE 6:

A simple random sample of 300 items is

selecte

reveals that 4% of the sampled are defective. At the 0.10 level of significance, can we conclude that less than 7% of the items in the shipment are defective?

from a large shipment, an

testing Testing the Population Mean when the Population Variance is Unknown

EXAMPLE 5:

A scrap metal dealer claims that the mean of his cash sales is no more than \$80, but an Internal Revenue Service agent believes the dealer is untruthful. Observing a sample of 20 cash customers, the agent finds the mean purchase to be \$85, with a standard deviation of \$21.   Making Decision Using p-value Approach Calculation of the p-value

 The p-value of a test is the lowest level of significance at which the null hypothesis can be rejected. The calculation of p-value depends on the type of alternative hypothesis, and the sampling distribution of the test statistics Decision Rule: If the p-value is smaller than the significance level (α), H 0 is rejected.

Making Decision Using p-value

 Example 4: Example 5: Example 6:
 p-value Example 4: 2 P(Z > |Z * |) = 2 P(Z > 0.47) = 0.6384 (two – tail test) Example 5: P(t > t * ) = P(t > 1.065) = 0.1502 (from Excel) (right – tail test) Example 6: P(Z < Z * ) = P(Z < -2.04) = 0.0207 (left – tail test) Z * or t * is the value of the test statistic  6