HYPOTHESIS TESTING OF

MEANS
POPULATION STANDARD DEVIATION KNOWN
HYPOTHESIS TESTING OF MEANS
• We assume the Population Standard Deviation is known for the next
few slides
• Two-Tailed Tests of Means: Testing in the Scale of the Original
Variable
Example: A manufacturer supplies rear axles for the U.S. Postal Service
mail trucks.
• These axels must be able to withstand 80,000 pounds per square inch
in stress tests.
• An excessively strong axle raises production costs significantly.
• From experience, the standard deviation of the strength of the axles is
4,000 𝑝𝑜𝑢𝑛𝑑𝑠 𝑝𝑒𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑖𝑛𝑐ℎ.
HYPOTHESIS TESTING OF MEANS
• The manufacturer selects a sample of 100 axles from production, tests
them and finds that the sample mean stress capacity is
79,600 𝑝𝑜𝑢𝑛𝑑𝑠 𝑝𝑒𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑖𝑛𝑐ℎ.
• If the axle manufacturer using a significance level 𝛼 of 0.05 in
testing, will the axle meet his stress requirements?
HYPOTHESIS TESTING OF MEANS
• Solution: We may write the data as:
𝜇𝐻0 = 80, 000 ← 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑧𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑀𝑒𝑎𝑛
𝜎 = 4,000 ← 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
𝑛 = 100 ← 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒
𝑥 = 79,600 ← 𝑆𝑎𝑚𝑝𝑙𝑒 𝑀𝑒𝑎𝑛
𝐻0 : 𝜇 = 80,000 ← 𝑁𝑢𝑙𝑙 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝐻1 : 𝜇 ≠ 80, 000 ← 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝛼 = 0.05 ← 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑓𝑜𝑟 𝑡𝑒𝑠𝑡𝑖𝑛𝑔
Hypothesis Testing of Means
• These seven data must be written down in the above sequence for each
Hypothesis Testing Example where we do a two-tailed test hypothesis.
• We know the population standard deviation and the population size is (𝑛 =
100) large enough to be treated as infinite (see table on next page). Thus,
we can use the normal distribution for testing.
• First, we calculate the standard error for the mean
𝜎
Standard error for the mean = 𝜎𝑥 =
𝑛
4,000
=
100
4,000
=
10
= 400 𝑝𝑜𝑢𝑛𝑑𝑠 𝑝𝑒𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑖𝑛𝑐ℎ
• Therefore, we examine a normal random variable with mean 80,000 and
standard deviation 400 𝑝𝑜𝑢𝑛𝑑𝑠 𝑝𝑒𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑖𝑛𝑐ℎ.
HYPOTHESIS TESTING OF MEANS
HYPOTHESIS TESTING OF MEANS
• We know that the significance level (𝛼) is 0.05.
• This implies that each rejection region contains 25% of the area under
the standard normal curve.
• The acceptance region of area 95% has two shaded regions which
contains two equal areas of 47.5% each.
• From the normal table, we want a value of 𝑧𝛼 for which the area under
2
the normal curve between 0 and 𝑧𝛼 is 0.475. That is, we are looking
2
for 𝑧 such that 𝑃 0 < 𝑍 < 𝑧𝛼 = 0.475. Hence, 𝑧𝛼 = 1.96.
2 2
• Now, we would like to determine the limits of the acceptance region,
that is, the critical value for the test.
HYPOTHESIS TESTING OF MEANS
• Two tailed Hypothesis test at 0.05 significance level, in original scale
HYPOTHESIS TESTING OF MEANS
• We know that the sampling distribution of 𝑥 is normal or
approximately normal, with mean 𝜇 = 80,000 and standard deviation
𝜎𝑥 = 400. (Refer Slide 11 of the ‘Critical regions’ Session).
• As a result we may convert 𝑥 into the standard normal random
variable and obtain
𝑥−𝜇 𝑥𝐿 − 𝜇 𝑥𝐿 − 𝜇
𝑃 > =𝑃 𝑍> = 𝛼/2
400 400 400
𝛼
• By definition of z𝛼/2 , 𝑃 𝑍 > 𝑧𝛼/2 = and we found 𝑧𝛼 above.
2 2
• Since both statements above refer to the standard normal distribution,
we see that
𝑥𝑅 − 𝜇
= 𝑧𝛼 = 1.96
400 2
HYPOTHESIS TESTING OF MEANS
• Hence, we have the upper critical value as
𝑥𝑅 = 𝜇 + 1.96 (400)
= 80,000 + 784
= 80,784 𝑝𝑜𝑢𝑛𝑑𝑠 𝑝𝑒𝑟 𝑠𝑞𝑢𝑎𝑟𝑒 𝑖𝑛𝑐ℎ
• Similarly, we see at the lower critical value that
𝛼
𝑃 𝑍 < −𝑧𝛼/2 =
2
• Therefore,
𝑥𝐿 = 𝜇 − 1.96 400 = 80,00 − 784 = 79, 216 𝑝𝑜𝑢𝑛𝑑𝑠 𝑝𝑒𝑟 𝑠𝑞. 𝑖𝑛.
HYPOTHESIS TESTING OF MEAN
• Two-Tailed Hypothesis test at 0.05 significance level: Acceptance
region and Sample mean shown
HYPOTHESIS TESTING OF MEANS
• The sample mean lies within the acceptance region and the
manufacturer could find no evidence to reject the null hypothesis.
• The manufacturer should accept the production run as meeting the
stress requirements.
• In the above calculation we kept the original variable the sample mean
𝑥 and reduced all calculations to this variable. This is called
‘Hypothesis Testing in the Original or Raw Scale’.
• Another way to do the above is to convert to the standard normal
variable in the beginning. This is called ‘Hypothesis Testing in the
Standardized Scale’.
HYPOTHESIS TESTING OF MEANS
Hypothesis Testing Standardized Scale:
• We can keep the original sample variable 𝑥 or convert it to the
standard normal random variable 𝑧.
• We used the variable 𝑧 in the previous section in the middle of the
calculation, but changed variables back to 𝑥 quickly.
• We cannot keep both the variables at once, they are different scales of
measurement.
• In the previous section, we noted that the sample mean 𝑥 = 79,600
lies between the lower limit 79,216 and upper limit 80,784.
HYPOTHESIS TESTING OF MEANS
• Both these scales are shown on the following picture
HYPOTHESIS TESTING OF MEANS
• Instead of converting the critical 𝑧 −values to the original scale to get
numbers directly comparable to the sample mean 𝑥, we could have
converted the sample mean 𝑥 to the standardized scale to get an
observed 𝑧 −value. To test the Hypothesis, we compare this 𝑧 −value
to the critical 𝑧 −value.
• In the present example:
𝑥 −𝜇𝐻0 79,600 −80,000
𝑧= = = −1.00
𝜎𝑥 400
• The critical values in the standardized scale are just
𝑧𝛼 = 1.96, −𝑧𝛼 = −1.96
2 2
• Thus, since 𝑧 = −1, we cannot reject the null hypothesis.
HYPOTHESIS TESTING OF MEANS
We will always follow the following standard five-step procedure for testing
hypotheses:
1. Decide whether this is a two-tailed or one-tailed test. State your
hypotheses. Select a level of significance appropriate for your decision.
2. Decide which distribution (𝑡 or 𝑧) to use (table on next page) and find the
critical value(s) for the chosen level of significance from the appropriate
𝑡 𝑜𝑟 𝑧 table.
3. Calculate the standard error of the sample statistic. Use the standard error
to convert the observed value of the sample statistic to the standardized
scale. (use table on page after next)
4. Sketch the distribution and mark the position of the standardized sample
value and the critical values for the test.
5. Compare the value of the standardized sample statistic with the critical
values for this test and interpret the result.
HYPOTHESIS TESTING OF MEANS
HYPOTHESIS TESTING OF MEANS
• Summary of Hypothesis Tests about a population mean with 𝜎 known
LOWER (LEFT) UPPER TWO-TAILED
TAIL TEST (RIGHT)TAIL TEST TEST

Hypotheses 𝐻0 : 𝜇 ≥ 𝜇0 𝐻0 : 𝜇 ≤ 𝜇0 𝐻0 : 𝜇 = 𝜇0
𝐻𝑎 : 𝜇 < 𝜇0 𝐻𝑎 : 𝜇 > 𝜇0 𝐻𝑎 : 𝜇 ≠ 𝜇0

Test Statistic 𝑥 − 𝜇0 SAME AS LOWER SAME AS LOWER

𝑧= 𝜎 TAIL TEST TAIL TEST
𝑛
Rejection Rule: Reject 𝐻0 if Reject 𝐻0 if Reject 𝐻0 if
Critical Value 𝑧 ≤ −𝑧𝛼 𝑧 ≥ 𝑧𝛼 𝑧 ≤ −𝑧𝛼 or if
2
𝑧 ≥ 𝑧𝛼
2
 HYPOTHESIS TESTING OF MEANS
• As another Example, we perform a one-tailed test of means:
Example:
• A hospital uses large quantities of packaged doses of a particular drug. The
individual dose of this drug is 100 𝑐𝑐.
• The action of the drug is such that the body will harmlessly pass off excess
doses. On the other hand, insufficient doses are to be avoided, since they
interfere with medical treatment.
• From the manufacturer, the population standard deviation is 2 𝑐𝑐.
• The hospital wants to know if the dosage in a particular shipment are too
small. The hospital inspects 50 doses of this drug at random from a very
large shipment and finds the mean of these doses to be 99.75 𝑐𝑐.
HYPOTHESIS TESTING OF MEANS
• If the hospital sets a 0.01 significance level and asks us whether the dosages
in this shipment are too small how can we find the answer?
Answer: It does not matter if the shipments’ dosage is more than 100 𝑐𝑐.
Thus, this is a Left tailed test of Hypothesis (look at the previous two tables).
The data should be written as:
𝜇𝐻0 = 100 ← 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑧𝑒𝑑 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑃𝑜𝑝. 𝑀𝑒𝑎𝑛
𝜎 = 2 ← 𝑃𝑜𝑝. 𝑆𝑡𝑑. 𝐷𝑒𝑣.
𝑛 = 50 ← 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒
𝑥 = 99.75 ← 𝑆𝑎𝑚𝑝𝑙𝑒 𝑀𝑒𝑎𝑛
𝐻0 : 𝜇 = 100 ← 𝑁𝑢𝑙𝑙 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝐻1 : 𝜇 < 100 ← 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝛼 = 0.10 ← 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒
HYPOTHESIS TESTING OF MEANS
• We know the population standard distribution and the sample size 𝑛 =
30. Hence, we can use the normal distribution from the table above.
• This is a Lower tailed test and by the table above the null hypothesis is
rejected if
𝑧 ≤ −𝑧𝛼
where 𝛼 = 0.10 and 𝑧𝛼 is the value of 𝑧 such that 10% of the area under
the curve lies to the left of 𝑧𝛼 . The remaining region is the acceptance
region and is 90% of the area. Examining the standard normal curve, we
see that
40% 𝑜𝑓 𝑡ℎ𝑒 𝑎𝑟𝑒𝑎 𝑢𝑛𝑑𝑒𝑟 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝑚𝑢𝑠𝑡 𝑙𝑖𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑧𝛼 𝑎𝑛𝑑 0.
 HYPOTHESIS TESTING OF MEANS
• This implies that 𝑃 −𝑧𝛼 < 𝑧 < 0 = 0.4.
• Hence, 𝑃 0 < 𝑧 < 𝑧𝛼 = 0.4
• From the 𝑍 −table, 𝑧𝛼 = 1.28, 𝑠𝑜 − 𝑧𝛼 = −1.28
• We now calculate the standard error of the mean using the value of 𝜎
𝜎
𝜎𝑥 =
𝑛
2
=
50
2
=
7.07
= 0.2829 𝑐𝑐 ← 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐸𝑟𝑟. 𝑜𝑓 𝑚𝑒𝑎𝑛
HYPOTHESIS TESTING OF MEANS
• Area under Curve for Left-Tailed Test at 0.1 significance level
HYPOTHESIS TESTING OF MEANS
• We now standardize the sample mean 𝑥:
𝑥−𝜇𝐻0
𝑧=
𝜎𝑥
99.75 −100
= = −0.88
0.2829

• This is more than 𝑧𝛼 = −1.28 hence we cannot reject the null

hypothesis. Thus, the drug shipment cannot be returned on the basis of
this test.
HYPOTHESIS TESTING OF MEANS
• Standardized Sample Value for Left-Tailed Hypothesis Test at 0.1
Significance Level
 Practice Sums
Example: 3
Hinton Press hypothesizes that the average life of its largest web press is
14,500 hours. They know that the standard deviation of press life is 2,100
hours. From a sample of 25 presses, the company finds a sample mean of
13,000 hours. At a 0.01 significance level, should the company conclude
that the average life of the presses is less than the hypothesized 14,500
hours?
HYPOTHESIS TESTING OF MEANS
Answer:
This is a Left tailed test of Hypothesis (look at the previous two tables).
The data should be written as:
𝜇𝐻0 = 14,500 ← 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑧𝑒𝑑 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝑃𝑜𝑝. 𝑀𝑒𝑎𝑛
𝜎 = 2,100 ← 𝑃𝑜𝑝. 𝑆𝑡𝑑. 𝐷𝑒𝑣.
𝑛 = 25 ← 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒
𝑥 = 13,000 ← 𝑆𝑎𝑚𝑝𝑙𝑒 𝑀𝑒𝑎𝑛
𝐻0 : 𝜇 = 14,500 ← 𝑁𝑢𝑙𝑙 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝐻1 : 𝜇 < 14,500 ← 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠
𝛼 = 0.01 ← 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒
HYPOTHESIS TESTING OF MEANS
• We know the population standard distribution and the sample size 𝑛 =
25. Hence, we can use the normal distribution from the table above.
• This is a Lower tailed test and by the table above the null hypothesis is
rejected if
𝑧 ≤ −𝑧𝛼
where 𝛼 = 0.01 and 𝑧𝛼 is the value of 𝑧 such that 1% of the area under
the curve lies to the left of 𝑧𝛼 . The remaining region is the acceptance
region and is 99% of the area. Examining the standard normal curve, we
see that
49% 𝑜𝑓 𝑡ℎ𝑒 𝑎𝑟𝑒𝑎 𝑢𝑛𝑑𝑒𝑟 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝑚𝑢𝑠𝑡 𝑙𝑖𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑧𝛼 𝑎𝑛𝑑 0.
HYPOTHESIS TESTING OF MEANS
• This implies that 𝑃 −𝑧𝛼 < 𝑧 < 0 = 0.49.
• Hence, 𝑃 0 < 𝑧 < 𝑧𝛼 = 0.49
• From the 𝑍 −table, 𝑧𝛼 = 2.33, 𝑠𝑜 − 𝑧𝛼 = −2.33
• We now calculate the standard error of the mean using the value of 𝜎
𝜎
𝜎𝑥 =
𝑛
2100
=
25
2100
=
5
= 420 ← 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐸𝑟𝑟. 𝑜𝑓 𝑚𝑒𝑎𝑛
HYPOTHESIS TESTING OF MEANS
• We now standardize the sample mean 𝑥:
𝑥−𝜇𝐻0
𝑧=
𝜎𝑥
13,000 −14,500
= = −3.57
420

• This is less than 𝑧𝛼 = −2.33 hence we reject the null hypothesis.

Thus, the average life is significantly less than the hypothesized value.
 Practice Sum
Example: 4
American Theaters knows that a certain hit movie ran an average of 84 days
in each city, and the corresponding standard deviation was 10 days. The
manager of the southeastern district was interested in comparing the movie’s
popularity in his region with that in all of American’s other theaters. He
randomly chose 75 theaters in his region and found that they ran the movie
an average of 81.5 days.
(a) State appropriate hypotheses for testing whether there was a
significant difference in the length of the picture’s run between
theaters in the southeastern district and all of American’s other
theaters.
(b) At a 1 percent significance level, test these hypotheses.