COMM 207
Tutorial 2
Outline
1. Review of hypothesis testing
2. Discussion of p-values vs. critical values
3. Examples of full worked problems. These are based on examples from Bowerman, Drougas,
Duckworth, Froelich, Hummel, Moninger, and Schur (2019) Business Statistics and Analytics in
Practice, 9th Edition. McGraw-Hill Ryerson.
Review of hypothesis testing
Hypothesis testing is a form of statistical inference. It’s about using information gleaned from a
sample to say something about the population. In the examples we saw already, and in the ones
we’ll go through today, we use the sample mean (a statistic) to make a conclusion about the
population mean (a parameter).
We have five steps in hypothesis testing (so far…more next week!)
1. State the null and alternative hypotheses.
There are three choices here (again, so far):
H 0 : μ ≤ μ0 ; H a : μ> μ0
H 0 : μ ≥ μ0 ; H a : μ< μ0
H 0 : μ=μ 0 ; H a : μ ≠ μ 0
The one you choose depends on the problem at hand. Remember, the null hypothesis will always
contain the equality.
We refer to the first two as “one-sided hypothesis tests”, because they each use only one side of
the distribution of the test statistic. We refer to the last as a “two-sided hypothesis test”, because
it involves both sides of the distribution of the test statistic. More on this later.
2. Calculate the test statistic.
So far, we have only one possible test statistic:
x−μ0
z=
σ /√ n
If you know the population standard deviation (sigma, σ ¿ then use z. Otherwise, use t (next
week). To show your work, show which equation you choose, how you plug in values, and your
final answer.
3. Calculate the p-value, or find the critical value.
Critical value: the value on the horizontal axis that provides a threshold (tells you whether to
reject the null hypothesis). A critical value is NOT a probability.
If the test statistic is further to the appropriate tail than the critical value, reject the null
hypothesis.
P-value: the probability that you see a test statistic as extreme or more extreme than the one you
see, given the null hypothesis is true.
If the p-value is less than the significance level, reject the null hypothesis.
The examples today will always show both the critical value method and the p-value method;
however, you only need to use one. If the question doesn’t tell you which one you need to use,
then pick the one you like better.
4. Use either the p-value or the critical value to decide whether to reject the null hypothesis.
Compare the p-value to the significance level=alpha=α . Remember, by default, α =0.05 (unless
you are told otherwise in the question).
Compare the test statistic to the critical value.
See file: COMM207_Tutorial2_Picture.pdf for an image that might help.
If the p-value is greater than the significance level, do not reject the null hypothesis.
If the critical value is further to the appropriate tail than the test statistic, then do not reject the
null hypothesis.
5. Interpret your result in the context of the original research question.
What does “reject the null hypothesis” or “do not reject the null hypothesis” mean regarding the
question that was asked?
Example 1 (based on Bowerman et al., 2014 exercise 7.14, page 231):
Suppose a bank manager has developed a new system to reduce waiting time during peak hours.
The manager hopes the new system will reduce the waiting time to less than 6 minutes. The
manger takes a sample of 100 waiting times, and gets a sample mean of 5.46 minutes. She knows
the population standard deviation is 2.47 minutes. Conduct an appropriate hypothesis test.
Example 2 (based on exercise 7.12, page 231)
Suppose we want to test the null hypothesis that the population mean is equal to 40 against the
alternative that it is not equal to 40, and the population standard deviation is known to be 18. We
take a sample of 81 measurements and find a sample mean of 35. Conduct an appropriate
hypothesis test.
