Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
Statistics
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
2
Where We’ve Been
Calculated point estimators of population parameters
Used the sampling distribution of a statistic to assess the reliability of an estimate through a confidence interval
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
3
Where We’re Going
Test a specific value of a population parameter
Measure the reliability of the test
8.1: The Elements of a Test of Hypotheses
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
4
Confidence IntervalWhere on the number line do the data point us?(No prior idea about the value of the parameter.)
Hypothesis TestDo the data point us to this particular value?(We have a value in mind from the outset.)
µ? µ?
µ0?
8.1: The Elements of a Test of Hypotheses
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
5
Null Hypothesis: H0
•This will be supported unless the data provide evidence that it is false• The status quo Alternative Hypothesis: Ha
•This will be supported if the data provide sufficient evidence that it is true• The research hypothesis
8.1: The Elements of a Test of Hypotheses
If the test statistic has a high probability when H0 is true, then H0 is not rejected.
If the test statistic has a (very) low probability when H0 is true, then H0 is rejected.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
6
8.1: The Elements of a Test of Hypotheses
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
7
8.1: The Elements of a Test of Hypotheses
Reality ↓ / Test Result → Do not reject H0 Reject H0
H0 is true Correct!Type I Error: rejecting a true null hypothesisP(Type I error) = α
H0 is falseType II Error: not rejecting a false null hypothesisP(Type II error) = β
Correct!
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
8
8.1: The Elements of a Test of Hypotheses
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
9
Note: Null hypotheses are either rejected, or else there is insufficient evidence to reject them. (I.e., we don’t accept null hypotheses.)
8.1: The Elements of a Test of Hypotheses
• Null hypothesis (H0): A theory about the values of one or more parameters
• Ex.: H0: µ = µ0 (a specified value for µ)
• Alternative hypothesis (Ha): Contradicts the null hypothesis
• Ex.: H0: µ ≠ µ0
• Test Statistic: The sample statistic to be used to test the hypothesis• Rejection region: The values for the test statistic which lead to rejection of
the null hypothesis• Assumptions: Clear statements about any assumptions concerning the
target population• Experiment and calculation of test statistic: The appropriate calculation for
the test based on the sample data• Conclusion: Reject the null hypothesis (with possible Type I error) or do
not reject it (with possible Type II error)
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
10
8.1: The Elements of a Test of Hypotheses
Suppose a new interpretation of the rules by soccer referees is expected to increase the number of yellow cards per game. The average number of yellow cards per game had been 4. A sample of 121 matches produced an average of 4.7 yellow cards per game, with a standard deviation of .5 cards. At the 5% significance level, has there been a change in infractions called?
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
11
8.1: The Elements of a Test of Hypotheses
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
12
H0: µ = 4Ha: µ ≠ 4Sample statistic: = 4.7α= .05
Assume the sampling distribution is normal.
Test statistic:
Conclusion: z.025 = 1.96. Since z* > z.025 , reject H0.(That is, there do seem to be more yellow cards.)
94.10064.
47.4* 0
xs
xz
8.2: Large-Sample Test of a Hypothesis about a Population Mean
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
13
The null hypothesis isusually stated as an equality …
… even though the alternative hypothesis can be either an equality or an inequality.
8.2: Large-Sample Test of a Hypothesis about a Population Mean
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
14
8.2: Large-Sample Test of a Hypothesis about a Population Mean
Lower Tailed Upper Tailed Two tailed
α = .10 z < - 1.28 z > 1.28 | z | > 1.645
α = .05 z < - 1.645 z > 1.645 | z | > 1.96
α = .01 z < - 2.33 z > 2.33 | z | > 2.575
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
15
Rejection Regions for Common Values of α
8.2: Large-Sample Test of a Hypothesis about a Population Mean
One-Tailed Test H0 : µ = µ0
Ha : µ < or > µ0
Test Statistic:
Rejection Region: | z | > zα
Two-Tailed Test
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
16
x
xz
0
H0 : µ = µ0
Ha : µ ≠ µ0
Test Statistic:
Rejection Region: | z | > zα/2
x
xz
0
Conditions: 1) A random sample is selected from the target population.2) The sample size n is large.
8.2: Large-Sample Test of a Hypothesis about a Population Mean
The Economics of Education Review (Vol. 21, 2002) reported a mean salary for males with postgraduate degrees of $61,340, with an estimated standard error (s ) equal to $2,185. We wish to test, at the α = .05 level, H0: µ = $60,000.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
17
8.2: Large-Sample Test of a Hypothesis about a Population Mean
The Economics of Education Review (Vol. 21, 2002) reported a mean salary for males with postgraduate degrees of $61,340, with an estimated standard error (s) equal to $2,185. We wish to test, at the α = .05 level, H0: µ = $60,000.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
18
H0 : µ = 60,000
Ha : µ ≠ 60,000
Test Statistic:
Rejection Region: | z | > z.025 = 1.96
613.
185,2
000,60340,61
0
z
z
xz
x
Do not reject H0
8.3:Observed Significance Levels: p - Values
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
19
Suppose z = 2.12. P(z > 2.12) = .0170.
Reject H0 at the α= .05 level Do not reject H0 at the α= .01 level
But it’s pretty close, isn’t it?
8.3:Observed Significance Levels: p - Values
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
20
The observed significance level, or p-value, for a test is the probability of observing the results actually observed (z*) assuming the null hypothesis is true.
The lower this probability, the less likely H0 is true.
)|*( 0HzzP
Let’s go back to the Economics of Education Review report (= $61,340, s = $2,185). This time we’ll test H0: µ = $65,000.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
21
H0 : µ = 65,000
Ha : µ ≠ 65,000
Test Statistic:
p-value: P(|z| > 1.675) = .047 * 2 =0.094
675.1
185,2
000,65340,61
0
z
z
xz
x
8.3:Observed Significance Levels: p - Values
8.3:Observed Significance Levels: p - Values
Reporting test results Choose the maximum tolerable value of α If the p-value ≤ α, reject H0
If the p-value > α, do not reject H0
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
22
8.4: Small-Sample Test of a Hypothesis about a Population Mean
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
23
If the sample is small and is unknown, testing hypotheses about µ requires the t-distribution instead of the z-distribution.
ns
xt
/0
One-Tailed Test H0 : µ = µ0
Ha : µ < or > µ0
Test Statistic:
Rejection Region: | t | > tα
Two-Tailed Test
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
24
ns
xt
/0
H0 : µ = µ0
Ha : µ ≠ µ0
Test Statistic:
Rejection Region: | t | > tα/2
ns
xt
/0
Conditions: 1) A random sample is selected from the target population.2) The population from which the sample is selected is
approximately normal.3) The value of tα is based on (n – 1) degrees of freedom
8.4: Small-Sample Test of a Hypothesis about a Population Mean
8.4: Small-Sample Test of a Hypothesis about a Population Mean
Suppose copiers average 100,000 between paper jams. A salesman claims his are better, and offers to leave 5 units for testing. The average number of copies between jams is 100,987, with a standard deviation of 157. Does his claim seem believable?
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
25
H0 : µ = 100,000
Ha : µ > 100,000
Test Statistic:
p-value: P(tdf=4 > 14.06) < .0005
8.4: Small-Sample Test of a Hypothesis about a Population Mean
Suppose copiers average 100,000 between paper jams. A salesman claims his are better, and offers to leave 5 units for testing. The average number of copies between jams is 100,987, with a standard deviation of 157. Does his claim seem believable?
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
26
06.145/157
000,100987,100/
0
t
t
ns
xt
H0 : µ = 100,000
Ha : µ > 100,000
Test Statistic:
p-value: P(|tdf=4| > 14.06) < .0005
8.4: Small-Sample Test of a Hypothesis about a Population Mean
Suppose copiers average 100,000 between paper jams. A salesman claims his are better, and offers to leave 5 units for testing. The average number of copies between jams is 100,987, with a standard deviation of 157. Does his claim seem believable?
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
27
06.145/157
000,100987,100/
0
t
t
ns
xt
Reject the null hypothesis based on the very low probability of seeing the observed results if the null were true.So, the claim does seem plausible.
One-Tailed Test H0 : p = p0
Ha : p < or > p0
Test Statistic:
Rejection Region: | z | > zα
Two-Tailed Test
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
28
p
ppz
ˆ
0ˆ
H0 : p = p0
Ha : p ≠ p0
Test Statistic:
Rejection Region: | z | > zα/2
p
ppz
ˆ
0ˆ
Conditions: 1) A random sample is selected from a binomial population. 2) The sample size n is large (i.e., np0 and nq0 are both ≥ 15).
8.5: Large-Sample Test of a Hypothesis about a Population Proportion
p0 = hypothesized value of p, , and q0 = 1 - p0 n
qpp
00ˆ
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
29
8.5: Large-Sample Test of a Hypothesis about a Population Proportion
Rope designed for use in the theatre must withstand unusual stresses. Assume a brand of 3” three-strand rope is expected to have a breaking strength of 1400 lbs. A vendor receives a shipment of rope and needs to (destructively) test it.
The vendor will reject any shipment which cannot pass a 1% defect test (that’s harsh, but so is falling scenery during an aria). 1500 sections of rope are tested, with 20 pieces failing the test. At the α = .01 level, should the shipment be rejected?
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
30
8.5: Large-Sample Test of a Hypothesis about a Population Proportion
The vendor will reject any shipment that cannot pass a 1% defects test . 1500 sections of rope are tested, with 20 pieces failing the test. At the α = .01 level, should the shipment be rejected?
H0: p = .01
Ha: p > .01
Rejection region: |z| > 2.33
Test statistic:
14.1
1500/)987)(.013(.
01.013.
ˆ
ˆ
0
z
z
ppz
p
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
31
8.5: Large-Sample Test of a Hypothesis about a Population Proportion
The vendor will reject any shipment that cannot pass a 1% defects test . 1500 sections of rope are tested, with 20 pieces failing the test. At the α = .01 level, should the shipment be rejected?
H0: p = .01
Ha: p > .01
Rejection region: |z| > 2.33
Test statistic:
14.1
1500/)987)(.013(.
01.013.
ˆ
ˆ
0
z
z
ppz
p
There is insufficient evidence to reject the null hypothesis based on the sample results.
There is insufficient evidence to reject the null hypothesis based on the sample results.
8.6: Calculating Type II Error Probabilities: More about β
To calculate P(Type II), or β, …
1. Calculate the value(s) of that divide the “do not reject” region from the “reject” region(s).
Upper-tailed test:
Lower-tailed test:
Two-tailed test:
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
32
n
szzx
n
szzx
n
szzx
n
szzx
xU
xL
x
x
2/02/00
2/02/00
000
000
8.6: Calculating Type II Error Probabilities: More about β
To calculate P(Type II), or β, …
1. Calculate the value(s) of that divide the “do not reject” region from the “reject” region(s).
2. Calculate the z-value of 0 assuming the alternative hypothesis mean is the true µ:
The probability of getting this z-value for the acceptance region is β.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
33
x
axz
0
8.6: Calculating Type II Error Probabilities: More about β
The power of a test is the probability that the test will correctly lead to the rejection of the null hypothesis for a particular value of µ in the alternative hypothesis. The power of a test is calculated as (1 - β ).
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
34
8.6: Calculating Type II Error Probabilities: More about β
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
35
The Economics of Education Review (Vol. 21, 2002) reported a mean salary for males with postgraduate degrees of $61,340, with an estimated standard error (s) equal to $2,185. We wish to test, at the α = .05 level, H0: µ = $60,000.
H0 : µ = 60,000
Ha : µ ≠ 60,000
Test Statistic: z = .613;
zα/2=.025 = 1.96
We did not reject this null hypothesis earlier, but what if the true mean were $62,000?
8.6: Calculating Type II Error Probabilities: More about β
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
36
The Economics of Education Review (Vol. 21, 2002) reported a mean salary for males with postgraduate degrees of $61,340, with s equal to $2,185.
We did not reject this null hypothesis earlier, but what if the true mean were $62,000?
xO,L= 60,000 - 1.96*2,185 = 55,717.4xO,U= 60,000 +1.96*2,185 = 64,282.6ZL = (55,717.4 – 62,000)/2,185 = -2.875ZU = (64,282.6 - 62,000)/2,185 = 1.045β = P(-2.875≤Z≤1.045) = 0.8511The power of this test is 1 – β = 1 – 0.8511 = 0.1489
8.6: Calculating Type II Error Probabilities: More about β
For fixed n and α, the value of β decreases and the power increases as the distance between µ0 and µa increases.
For fixed n, µ0 and µa, the value of β increases and the power decreases as the value of α is decreased.
For fixed α, µ0 and µa, the value of β decreases and the power increases as n is increased.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
37
8.6: Calculating Type II Error Probabilities: More about β
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
38
8.7: Tests of Hypotheses about a Population Variance
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
39
8.7: Tests of Hypotheses about a Population Variance
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
40
2
22 )1(
sn
The chi-square distribution is really a family of distributions, depending on the number of degrees of freedom.
But, the population must be normally distributed for the hypothesis tests on 2 (or ) to be reliable!
8.7: Tests of Hypotheses about a Population Variance
One-Tailed Test
Test statistic:
Rejection region:
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
41
2221
2
20
22
20
220
2
20
20
)1(
:
:
sn
H
H
a
Two-Tailed Test
Test statistic:
Rejection region:22/
222/1
2
20
22
20
2
20
20
)1(
:
:
or
sn
H
H
a
8.7: Tests of Hypotheses about a Population Variance
Conditions Required for a Valid
Large- Sample Hypothesis Test for 2
1. A random sample is selected from the target population.
2. The population from which the sample is selected is approximately normal.
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
42
8.7: Tests of Hypotheses about a Population Variance
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
43
Earlier, we considered the average number of copies between jams for a brand of copiers. The salesman also claims his copiers are more predictable, in that the standard deviation of jams is 125. In the sample of 5 copiers, that sample standard deviation was 157. Does his claim seem believable, at the α = .10 level?
8.7: Tests of Hypotheses about a Population Variance
Two-Tailed Test
Test statistic:
Rejection criterion:
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
44
48773.931.6
31.6125
)157)(15(
125:
125:
205.
2
2
22
2
20
aH
H
Earlier, we considered the average number of copies between jams for a brand of copiers. The salesman also claims his copiers are more predictable, in that the standard deviation of jams is 125. In the sample of 5 copiers, that sample standard deviation was 157. Does his claim seem believable, at the α = .10 level?
8.7: Tests of Hypotheses about a Population Variance
Two-Tailed Test
Test statistic:
Rejection criterion:
McClave, Statistics, 11th ed. Chapter 8: Inferences Based on a Single Sample: Tests of Hypotheses
45
48773.931.6
31.6125
)157)(15(
125:
125:
205.
2
2
22
2
20
aH
H
Earlier, we considered the average number of copies between jams for a brand of copiers. The salesman also claims his copiers are more reliable, in that the standard deviation of jams is 125. In the sample of 5 copiers, that sample standard deviation was 157. Does his claim seem believable, at the α = .10 level?
Do not reject the null
hypothesis.