Sie sind auf Seite 1von 29

Hypothesis Tests for Population

Proportions
Hypothesis Tests for Population Proportions
The four step process for hypothesis testing is the same when dealing with population
proportions as it is when dealing with population means; the only difference is the
method for calculating the standard deviation
The standard deviation of a distribution of sample proportions is:

p (1 p )
n
The z-score is then found by comparing the sample statistic ( p ) to the null hypothesis (p)
and dividing by the standard deviation:

p p
z
p(1 p) / n
types of test

Two-tailed test
H0 : k
Ha : k
One-tailed test
Right-tailed test Left-tailed test

H 0 : k H0 : k
H a : k Ha : k
One-tailed two tailed Critical value

80% .10 .20 1.28


90% .05 .10 1.65
95% .025 .05 1.96
98% .01 .02 2.33
99% .005 .01 2.58
Hypothesis Tests for Population Proportions

Example 1:
Suppose the national unemployment rate is 3.5%.
In a survey of n = 450 people in a rural
Wisconsin county, 22 people are found to be
unemployed. County officials apply for state aid
based on the claim that the local unemployment
rate is higher than the national average. Test
this claim at the .05 significance level.
Example 2 ~
A random sample of n = 750 people is
selected, of whom 92 are left-handed.
Use these sample data to test the
claim that 10% of the population is
left handed.
Z-test for proportion
A hypothesis-test involving a proportion can
be considered as a binomial experiment
where there are only two outcomes and the
probability of a success does not change
from a trial to trial that the =np and
standard deviation is =npq for the
binomial distribution.
Formula for the z-test proportions

z= p - p
pq/n
where: X
n

X
p = (sample proportion)
n

p= population proportion

n=sample size
the formula is derived form the normal
approximation to the binomial follows the general
formula

test Value= (observable value)-(expected value)


standard error
One-tailed two tailed Critical value

80% .10 .20 1.28


90% .05 .10 1.65
95% .025 .05 1.96
98% .01 .02 2.33
99% .005 .01 2.58
5 steps in solving z-test:

step 1. state the hypothesis and identify the claim


step 2. find the critical values
step 3. compute the test value
step 4. make decision
step 5. summarize the result
A dietitian claims that 60% of people are trying
to avoid trans fat 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 = 0.05, is there enough
evidence to reject the dietitian's claim?
A statistician read that at least 77% of the
population oppose replacing $1 bills with $1
coins. to see if the claim is valid, the statistician
selected a sample of 80 people and found that
55 were oppose to replacing the $1 bills. at
=0.01, test the claim that at least 77% of the
population are opposed to the change.
Testing the difference between
Proportions
The z test with some modifications can be used to test the
equality of to proportions.

(p hat) is the sample proportion used to estimate the


population proportion, denoted by p.

p =X Where

n X= number of units that possess the characteritic


of interest
n= sample size
When you are testing the difference between two
population proportions p 1 and p2, the hypothesis can be
stated thus, if no difference between the proportions in
hypothesized.

H 0 : P1 P2 H 0 : P1 P2 0
OR
H1 : P1 P2 H1 : P1 P 2 0
Similar statements using < or > in the alternative hypothesis can be
formed for one-tailed tests. X2
For two proportions, p 1 X 1 / n1 is used to estimate p1and p 2 = n2
is used to estimate p 2. The standard error of the difference is

p1q1 p2 q2
p1 p 2
2
p1
2

p2
n1 n2
where and p 2are the variance of the proportions,
2 2
p2

q1 1 p1 , q2 1 p2 and n1andn2 are respective sample sizes.

since p 1 and p 2 are unkown, a weighted estimate of p


can be computed by using the formula

n1 p 1 n2 p 2
p
n1 n1
and q 1 p . this weighted estimate is based on the
hypothesis that p1 p2 . hence, p is a better estimate than either
, sincepit1orisp 2a combined average using both 1andp
p 2

x1 X2
p p 2
since n1
and
n2 , p can be simplified to

x1 x2
p
n1 n2
finally, the Standard error of the difference in
terms of the weighted esimate is

1 1
p p pq
1
n n2
1 2
Formula for the Z test for comparing two proportions

z
2 p1 p2
1 p
p
1 1
pq
n1 n2
where
x1 x2 1
X1
p p
n1 n2 n1
X2
q 1 p p2 =
n2
Assumptions for the z test for two proportions

1. The samples must be ramdom samples.


2. The sample data are independent of one
another.
3. for both samples np 5 and nq 5.
In the nursing home study mentioned in the chapter
opening statistics today, the researcher found that 12
out of 34 small nursing homes has a resident
vaccination rate of less than 80%, while 17 out of 24
large nursing homes had vaccintion rate of less than
80%. at =.05, test the claim that there is no
difference in the proportions of the small and large
nursing home with a resident vaccination rate of less
than 80%.
Confidence Interval for the difference
between two proportions
Confidence interval
is a specific interval estimate of a parameter determined by
isung data obtained from a sample and by using the specific
confidence level of the estimate.

90%
95%
99%
formula

p1q1 p 2 q2 p1q1 p 2 q2
p1 p 2 z p1 p2 p1 p 2 z
2 n1 n2 2 n1 n2
In the nursing home study mentioned in the chapter
opening statistics today, the researcher found that 12 out of
34 small nursing homes has a resident vaccination rate of
less than 80%, while 17 out of 24 large nursing homes had
vaccintion rate of less than 80%. at =.05, test the claim
that there is no difference in the proportions of the small
and large nursing home with a resident vaccination rate of
less than 80%.

find the 95% confidence interval for the difference of


proportions for the data .

Das könnte Ihnen auch gefallen