Professional Documents
Culture Documents
Unit 3 Hypothesis Testing
Unit 3 Hypothesis Testing
Unit 3 Hypothesis Testing
1. A Retailers Association believes that the average amount spent on grocery by shoppers on each visit to
a supermarket is N$ 1800.
2. A mobile phone service provider claims that it has at least 25% of the mobile phone market.
3. A quality inspector believes that the variability of the fills of all 500 ml hand sanitisers produced is at
most 5 ml.
▪ Hypothesis testing begins with a value being assumed for a given population
parameter.
▪ The hypothesised value is derived from a claim/assumption or a research question.
▪ Two statistical hypotheses, the null 𝐻0 and alternative 𝐻𝑎 hypotheses, are formulated
based on the hypothesised population parameter value.
▪ The null hypothesis usually represents the status quo while the alternative hypothesis is
the opposite of the null hypothesis.
▪ It is not always obvious how the null 𝐻0 and alternative 𝐻𝑎 hypotheses should be
formulated.
▪ The hypotheses must be structured appropriately so that the hypothesis testing
conclusion provides the information the researcher or decision maker wants.
▪ Important to understand the context of the situation when formulating the
hypotheses.
▪ In some cases it is easier to identify the alternative hypothesis first. In other cases the
null hypothesis is easier
ABA 3692 Summer Term 2023
Step 1: Formulating the Hypotheses
▪ Three possible forms of hypotheses:
Note:
• The equality (=, ≤ 𝑜𝑟 ≥)always appears in the null hypothesis.
• We might begin with a belief or assumption that a statement about the value of a
population parameter is true.
• We then use a hypothesis test to challenge the assumption and determine if there is
statistical evidence to conclude that the assumption is incorrect.
Example: The label on a soft drink bottle states that it contains 330 ml.
Hypotheses to be tested:
𝐻𝑜 : 𝑡ℎ𝑒 𝑙𝑎𝑏𝑒𝑙 𝑖𝑠 𝑐𝑜𝑟𝑟𝑒𝑐𝑡.
𝐻𝑎 : 𝑡ℎ𝑒 𝑙𝑎𝑏𝑒𝑙 𝑖𝑠 𝑖𝑛𝑐𝑜𝑟𝑟𝑒𝑐𝑡.
• In such cases, it is often best to begin with the alternative hypothesis and make it the
conclusion that the researcher hopes to support.
• The conclusion that the research hypothesis is true is made if the sample data provide
sufficient evidence to show that the null hypothesis can be rejected.
Example : A new sales force bonus plan is developed in an attempt to increase sales.
Hypotheses to be tested:
▪ The probability of making a Type I error when the null hypothesis is true is called the level of
significance denoted by 𝛼.
▪ Applications of hypothesis testing that only control the Type I error are often called significance
tests.
Note: Common choices for 𝛼 are 0.01, 0.05 and 0.1
• The level of significance is the area/probability in the tails of a sampling distribution and this is
where the rejection region is located.
• In a two 𝛼tailed test, 𝛼 is split between the lower and upper regions of rejection and each tail
contains 2 .
• For a lower tailed test, 𝛼 is contained in the lower tail of the sampling distribution while for an
upper tailed distribution, 𝛼 is contained in the upper tail of the sampling distribution.
ABA 3692 Summer Term 2023
Step 3: Calculate the Sample Test Statistic
• Sample data is used to calculate a sample statistic that provides the evidence to test the
validity of the null hypothesis.
• The sample statistic must be expressed in standardised unit of measure as the critical limits of
the region of non-rejection.
• Formulas to calculate sample test statistics are available depending on the type of
hypotheses test.
• The critical limits/values of the region of non-rejection (of the null hypothesis) depends on the
sampling distribution and the level of significance defined in step 2.
• The sample statistic either lies within the region of non-rejection or it lies in the rejection region
of the null hypothesis.
Statistical conclusion:
• Depending on the outcome of the comparison in Step 4, the null hypothesis is either rejected or not rejected at
the given level of significance. This conclusion is based on the sample evidence and level of significance for the
test.
Practical conclusion:
• This considers the claim, assumption or research question to be answered. A practical interpretation of the
statistical conclusion above should be given i.e. if the null hypothesis is rejected, what does it mean in terms of
the claim, assumption or research question given the available sample evidence ?
• p-value approach : making use of probabilities to determine whether the null hypothesis can
be rejected or not (STEP 4).
• A p-value is the probability of observing the sample statistic or a more extreme value if the
hypotheised parameter value is assumed to be true.
• A small p-value i.e. close to zero indicates a low probability of observing the sample statistic if the null hypothesis
were true and this provides strong evidence to reject 𝐻0 .
• A large p-value i.e. close to one indicates a high probability of observing the sample statistic if the null
hypothesis were true and this provides strong evidence not to reject 𝐻0 .
• For a lower tailed test, the p-value is the area/probability in the lower tail of a sampling
distribution from the sample statistic.
• For an upper tailed test, the p-value is the area/probability in the upper tail of a sampling
distribution from the sample statistic.
• For a two tailed test, the p-value is the combined area/probability in the lower and upper
tails of the sampling distribution.
Based on the survey, the average value of grocery purchase was found to be N$ 1750. Assume
that the population of grocery purchase values is normally distributed with a variance of 4624
(N$ squared).
Required:
What can the retailers association conclude about the average amount spent by retailers on
grocery shoppers on each visit to a supermarket? Conduct a hypothesis test at a 1% level of
significance and use the critical value approach.
• Rejection rule:
Lower tail: Reject 𝐻0 𝑖𝑓 𝑝 − 𝑣𝑎𝑙𝑢𝑒 ≤ 𝛼
Upper tail: Reject 𝐻0 𝑖𝑓 𝑝 − 𝑣𝑎𝑙𝑢𝑒 ≤ 𝛼
Two-tailed: Reject 𝐻0 𝑖𝑓 𝑝 − 𝑣𝑎𝑙𝑢𝑒 ≤ 𝛼
Based on the survey, the average value of grocery purchase was found to be N$ 1750. Assume
that the population of grocery purchase values is normally distributed with a variance of 4624
(N$ squared).
Required:
What can the retailers association conclude about the average amount spent by retailers on
grocery shoppers on each visit to a supermarket? Conduct a hypothesis test at a 1% level of
significance and use the p-value approach.
Required:
a) State the null and alternative hypotheses and determine the critical value(s) at the 5%
and 10% levels of significance.
To test this claim, eight delivery times are randomly selected and their delivery times (in minutes)
are given below:
42 50 42 40 45 51 46 44
Required:
What can be concluded about the average delivery times of the restaurant Conduct a
hypothesis test at a 5% level of significance and use the critical value approach.
• For a lower tail test: the standard normal probability distribution table can be used to find
the critical value −𝑧𝛼 , a z –value with an area of 𝛼 in the lower tail.
• For an upper tail test: the standard normal probability distribution table can be used to find
the critical value 𝑧𝛼 , a z –value with an area of 𝛼 in the upper tail.
• For a two tailed test: the standard normal probability distribution table can be used to find
the critical values −𝑧𝛼/2 𝑎𝑛𝑑𝑧𝛼/2 , z –values with an area of 𝛼/2 in the lower tail and upper tail,
respectively.
• Rejection rule:
Lower tail: Reject 𝐻0 𝑖𝑓 𝑧 𝑡𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 ≤ − 𝑧𝛼
• Rejection rule:
Lower tail: Reject 𝐻0 𝑖𝑓 𝑝 − 𝑣𝑎𝑙𝑢𝑒 ≤ 𝛼
A random sample of 100 mobile users is selected and it if found that 24 users subscribe to XYZ
Mobile as a service provider.
Required:
a) What can be concluded about XYZ Mobile’s share of the mobile phone market? Conduct
the hypothesis test at a 1% level of significance and use the critical value approach.
Required:
a) State the null and alternative hypotheses and determine the critical value(s) at the 5%
level of significance.
• For a lower tail test: the chi-square distribution tables can be used to find the critical value 𝜒(1−𝛼)
2
,a
chi-square value with an area of 𝛼 in the lower tail.
• For an upper tail test: the chi-square distribution tables can be used to find the critical value 𝜒𝛼2 , a
chi-square value with an area of 𝛼 in the upper tail.
• For a two tailed test: the chi-square distribution tables can be used to find the critical
𝛼
values 𝜒(1−𝛼/2)
2 2
𝑎𝑛𝑑 𝜒(𝛼/2) , chi-square values with areas of 2 in the lower and upper tails.
• Rejection rule:
Lower tail: Reject 𝐻0 𝑖𝑓𝜒 2 𝑡𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 ≤ 𝜒(1−𝛼)
2
To test this requirement, a random sample of 20 filled bottles is selected and their volumes
measured. The standard deviation is found to be 6.18 ml.
Required:
a) Conduct the hypothesis test at a 5% level of significance and determine whether the filling
machine is operating within the quality standards.
70 80 60 90 75
Required:
At 90% confidence, conduct a hypothesis test to determine if the variance of the population
is less than or equal to 50. Use the critical value approach.