Professional Documents
Culture Documents
We proceed with our test, drawing the hypothesized model and finding the z value
4
for the data.
We form an alternative hypothesis that is either one- or two-tailed based on what we
2
want to learn.
We form a null hypothesis specifying the parameter of a model we’ll test using our
1
data.
If the data are far out of line with the null hypothesis model, we will reject the null
5 hypothesis and if the data are consistent with the null hypothesis model, we will not
reject the null hypothesis.
3 We check the appropriate assumptions and conditions.
religious services at least once a week is higher than 30%. The variable is whether or not a
person attends services. The parameter of interest is the proportion of young adults in the area
Ho: p ≠ 0.30 Ha: p > 0.30 CORRECTION: Ho: p = 0.30 Ha: p > 0.30
surveyed are independent of one another. It’s possible a few of those selected may be
friends or family who influence each other, but independence is still roughly true.
• 10% condition: I assume the 132 in this sample are less than 10% of the population of
• Success/Failure condition: In the sample, there are 48 people who attend religious
CORRECTION: Should check
services and 84 who do not. Both are greater than 10. np=39.6 and nq=92.4.
Because the conditions are satisfied I’ll use a Normal model for the sampling distribution
The null model is a Normal distribution with a mean of 48/132 = 0.364 and a standard
(0.3)(1 − 0.3)
deviation of SD( pˆ ) = = 0.0399.
132
If the true proportion of young adults attending weekly religious services was 0.3, then an
observed value of 48/132 is not more than 2 standard deviations above the mean.
I reject Ho, I have evidence that the true proportion of young adults in this area who attend
CORRECTION: Since z is less than 2, I fail to reject Ho. There is no evidence that the true
proportion of young adults in this area who attend weekly religious services is different than the
national 30%.
4. A medical study claims the results of testing the effectiveness on a new medication are
statistically significant. What does this mean?
The observed gain in effectiveness is too large to be attributed just to variability or random chance.
5. Describe in context the Type I and Type II errors that are possible in question 4
Type I: I conclude the new medication is more effective when actually it is not.
Type II: I conclude the new medication is not more effective when actually it is more effective.
6. The government of Monroe County, New York, reports to a local newspaper an overall
conviction rate for DWI cases of 73 percent, one of the highest in New York. A reporter
suspects the rate is lower and obtains a random sample of 211 DWI cases of which 140 ended
in convictions. He performs a hypothesis test and rejects his null hypothesis. Circle the true
statements.
a. It was found that the probability of obtaining a statistic as least this extreme is very small,
assuming there has been no change in the DWI conviction rate.
b. There is statistically significant evidence that the conviction rate is 73%.
c. The true conviction rate is actually lower than 73%
d. This is strong evidence that the true conviction rate is less than 73%
e. The sample statistic is likely to occur due to sampling variation.
f. No samples of 211 DWI cases will have 140 convictions.
140
g. p = 0.73 and pˆ = .
211
Null hypothesis: 53% of dogs belonging to owners who exercise regularly are
Think
overweight.
Alternative hypothesis: Less than 53% of dogs belonging to owners who
Hypothesis
exercise regularly are overweight.
In symbols: H 0 : p = 0.53, H A : p < 0.53 .
proportion z-test.
n = 291 , x = 136 ,
pˆ = 136
291
≈ 0.4674
pˆ − p
z=
pq
Mechanics n
0.4674 − 0.53
z=
(0.53)(0.47)
P = P( pˆ < 0.4605)
291
= P( z < −2.14)
z = −2.14
= 0.016
The P-value is quite low (or z < -2), so we reject the null hypothesis. There is
Tell
Conclusion evidence to suggest that less than 53% of dogs whose owners exercise regularly
are overweight.
The conditions are satisfied, so I can use the Normal model to perform a one proportion z-test.
Mechanics.
n = 100 , x = 73 , pˆ = 100
73
= 0.73
pˆ − p
z=
pq
n
P = P ( pˆ < 0.73)
0.73 − 0.8
z= = P ( z < −1.75)
(0.8)(0.2)
= 0.04
100
z = −1.75
Conclusion.
The P-value is quite low, so reject the null hypothesis. There is evidence to suggest that the proportion of
satisfied customers at this insurance company is lower than the company’s claimed 80%.
Alternatively if using the Rule of 2, there is not enough evidence to reject the null hypothesis that the
proportion of satisfied customers at this insurance company is lower than the company’s claimed 80%.