12.01.2013 Views

Confidence Intervals and Hypothesis Tests: Two Samples - Florida ...

Confidence Intervals and Hypothesis Tests: Two Samples - Florida ...

Confidence Intervals and Hypothesis Tests: Two Samples - Florida ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

STATSprofessor.com<br />

Chapter 9<br />

Example 144 In 1995, the American Cancer Society r<strong>and</strong>omly sampled 1500 adults of which 555 smoked.<br />

In 2005, they surveyed 1750 adults <strong>and</strong> found that 578 of them smoked. At the 5% level of significance<br />

test the claim that the proportion of smokers in the population decreased over the ten year period.<br />

What is the p-value for the test?<br />

Let’s now look at the confidence interval for the difference between two proportions:<br />

<strong>Confidence</strong> Interval for( p − p ) :<br />

1 2<br />

p q p q pˆ qˆ pˆ qˆ<br />

p − p ± z + ≈ p − p ± z +<br />

( ˆ ˆ ) ( ˆ ˆ )<br />

1 1 2 2 1 1 2 2<br />

1 2 α / 2<br />

n1 n2 1 2 α / 2<br />

n1 n2<br />

Steps for constructing the <strong>Confidence</strong> Interval for the True Difference between the Population<br />

Proportions (large, independent samples):<br />

Step 1 Gather Data from Problem, 1 1 1 2 2 2<br />

Step 2 Find Zα / 2<br />

n , pˆ , qˆ , n , pˆ , qˆ , <strong>and</strong> the <strong>Confidence</strong> level.<br />

Step 3 Use the results from steps 2 <strong>and</strong> 1 to get the margin of error, E =<br />

Step 4 Form ( ˆ ˆ ) , ( ˆ ˆ )<br />

⎡ p − p − E p − p + E⎤<br />

⎣ 1 2 1 2 ⎦<br />

z<br />

pˆ qˆ pˆ qˆ<br />

1 1 2 2<br />

α /2 +<br />

n1 n2<br />

21

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!