01.06.2013 Views

Statistical Methods in Medical Research 4ed

Statistical Methods in Medical Research 4ed

Statistical Methods in Medical Research 4ed

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

116 Analys<strong>in</strong>g means and proportions<br />

Probability<br />

0 . 3<br />

0 . 2<br />

0 . 1<br />

0<br />

5 1015 20<br />

Observed<br />

Fig. 4.8 B<strong>in</strong>omial distributions illustrat<strong>in</strong>g the 95% confidence limits for the parameter p based on a<br />

sample with five <strong>in</strong>dividuals of a certa<strong>in</strong> type out of 20. For p ˆ pL ˆ 0 09 the probability of 5 or<br />

more is 0 025; for p ˆ pU ˆ 0 49 the probability of 5 or less is 0 025. ( ) pL ˆ 0 09; (ÐÐÐÐ)<br />

pU ˆ 0 49.<br />

3 Replace pL…1 pL) and pU…1 pU† by p…1 p†. This is not too drastic a<br />

step, as p…1 p† changes rather slowly with changes <strong>in</strong> p, particularly for<br />

values of p near 1<br />

2 . Ignor<strong>in</strong>g the cont<strong>in</strong>uity correction, as <strong>in</strong> 2, we have the<br />

most frequently used approximation to the 95% confidence limits:<br />

p<br />

p<br />

1 96 …pq=n† ,<br />

where, as usual, q ˆ 1 p. The simplification here is due to the replacement<br />

of the standard error of p, which <strong>in</strong>volves the unknown value p, by the<br />

p<br />

approximate form …pq=n† , which can be calculated entirely from known<br />

quantities.

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

Saved successfully!

Ooh no, something went wrong!