9.

3 Hypothesis Testing for

Proportions

Objectives:

Test proportions using the z test

 Step 0 – Assumptions:
The assumptions for the hypothesis test for a proportion is similar to
the hypothesis test for a mean with one slight difference.
1) Identify the variable and ensure that the level of measurement is
at the nominal. (This is different! – tests of a mean have to be interval-ratio
level –but in proportions it will be a proportion of a nominal variable.
2) Identify the method of sampling. If the specific method is not
clearly stated – verify that the subjects or data were randomly
selected.
3) Identify the shape of the population distribution. If there is no
indication of normally distributed data – identify the sample size. If
and the distribution is made up of sample means then the Central
Limit Theorem guarantees an approximately normal distribution. If
sample size is smaller than 30 check to see that the following
properties are true in order to be able to use this test: and .
5-Step Process/Z-Test for a Proportion
We will use the same 5-Step process but with a couple of
differences.

• Step 0 – Assumptions:
• Step 1 - State the Hypothesis (Slightly Different)
instead of μ we use p.
• Step 2 - Critical Value(s) (The Same)
• Step 3 - Test Value (Displayed on next slide)
• Step 4 - Make the decision (The Same)
• Step 5 – Summarize the results (The Same)
 Test Value for a Z-Test for a Proportion
The formula for the z test
for a proportion is
Or, use:
pˆ  p
z
pq n
(read p “hat”)

= (sample proportion) = is a percentage

p = population proportion =

𝑇h𝑒𝑐𝑙𝑎𝑖𝑚𝑒𝑑 𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒(𝑢𝑠𝑢𝑠𝑢𝑎𝑙𝑙𝑦𝑢𝑠𝑒𝑑𝑖𝑛 𝑡h𝑒h𝑦𝑝𝑜𝑡h𝑒𝑠𝑖𝑠)

q= 1–p

n = sample size
Example: Avoiding Trans Fats
A dietician claims that 60% of people are trying to avoid trans fats in their diets. She
randomly selected 200 people and found that 128 people stated that they were trying
to avoid trans fats in their diets.
At α = 0.05, is there enough evidence to reject the dietitian’s claim?

Step 0 – Assumptions:

Step 1: State the hypotheses and identify the claim.

H0: p = 0.60 (claim) and H1: p  0.60
Step 2: Find the critical value.
α = 0.05. This test is a two-tailed test. Since this is a two-tailed test,
the critical value is z = ±1.96. we need
α/2 = .025
Example: Avoiding Trans Fats
A dietician claims that 60% of people are trying to avoid trans fats in their
diets. She randomly selected 200 people and found that 128 people stated
that they were trying to avoid trans fats in their diets. At α = 0.05, is there
enough evidence to reject the dietitian’s claim?
Step 3: Compute the test value by hand
(you need to have seen this at least once).

= = .64

p = population proportion
.60
q= 1–p
1 - .60 = .40
n = sample size
200

pˆ  p 0.64  0.60
z 
pq n  0.60 0.40  200  1.15
Example: Avoiding Trans Fats
A dietician claims that 60% of people are trying to avoid trans fats in their
diets. She randomly selected 200 people and found that 128 people stated
that they were trying to avoid trans fats in their diets. At α = 0.05, is there
enough evidence to reject the dietitian’s claim?
Step 3: Compute the test value (using GeoGebra).
 Example: Avoiding Trans Fats
Step 4: Make the decision.
Fail to reject the null hypothesis since the test value
does not fall inside the critical region.
 Example: Avoiding Trans Fats
Step 5: Summarize the results.

There is not enough evidence to reject the claim that

60% of people are trying to avoid trans fats in their diets.
 Example: Family and Medical Leave Act
The Family and Medical Leave Act provides job protection and
unpaid time off from work for a serious illness or birth of a
child.
In 2000, 60% of the respondents of a survey stated that it was
very easy to get time off for these circumstances. A researcher
wishes to see if this percentage has changed.

A sample of 100 people who used the leave said that 53%
found it easy to do so. At α = 0.01, has the percentage
changed?
 Example: Family/Medical Leave Act
In 2000, 60% of the respondents of a survey stated that it was very easy to get time off for these
circumstances. A researcher wishes to see if the percentage who said that it was very easy to get
time off has changed.
A sample of 100 people who used the leave said that 53% found it easy to use the leave. At α = 0.01,
has the percentage changed?

Step 0 – Assumptions:
Step 1 State the hypotheses and identify the claim.

Step 2 Find the critical value(s). you must use

Since α = 0.01 and this test is two-tailed,
The critical values are ±2.58.
 Example: Family/Medical Leave Act
In 2000, 60% of the respondents of a survey stated that it was very easy to
get time off for these circumstances. A researcher wishes to see if the
percentage who said that it was very easy to get time off has changed.
A sample of 100 people who used the leave said that 53% found it easy to
use the leave. At α = 0.01, has the percentage changed?

Step 3 Compute the test value. In order to use GeoGebra we need x

(the number of successes) and n (the sample size)
pˆ  p
z
pq n

Z = -1.43
 Example: Family/Medical Leave Act
Step 4 Make the decision.
Because the test value of z = -1.43 is greater than
the critical value of z = -2.58, do not reject the null.

Step 5 Summarize the results.

There is/is not enough evidence to support/reject the

claim that the percentage of those who found it was easy
to get time off under the medical leave act has changed.