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

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

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

230 Comparison of several groups<br />

Table 8.2 Presence or absence of S. mansoni eggs <strong>in</strong> the stool.<br />

Age (years) 0± 10± 20± 30± 40± Total<br />

Positive 14 16 14 7 6 57<br />

Negative 87 33 66 34 11 231<br />

Total 101 49 80 41 17 288<br />

The expected number of positives for the age group 0± is<br />

…57†…101†=288 ˆ 19 99:<br />

The set of expected numbers for the 10 cells <strong>in</strong> the table is<br />

Total<br />

19 99 9 70 15 838 113 6 57<br />

81 01 393 0 64 17 32 89 13 64 231<br />

Total 101 49 80 41 17 288<br />

The fact that the expected numbers add to the same marg<strong>in</strong>al totals as those observed is a<br />

useful check.<br />

The contribution to X 2 from the first cell is<br />

and the set of contributions for the 10 cells is<br />

giv<strong>in</strong>g a total of<br />

…14 19 99† 2 =19 99 ˆ 1 79,<br />

1 79 4 09 0 21 0 15 2 07<br />

0 44 1 01 0 05 0 04 0 51<br />

X 2 ˆ 10 36:<br />

The degrees of freedom are k 1 ˆ 4, for which the 5% po<strong>in</strong>t is 9 49. The departures from<br />

the null hypothesis are thus significant at the 5% level …P ˆ 0 035†.<br />

In this example the column classification is based on a cont<strong>in</strong>uous variable, age, and it<br />

would be natural to ask whether the proportions of positives exhibit any smooth trend with<br />

p<br />

age. The estimated proportions, with their standard errors calculated as …piqi=ni† , are<br />

0 14 0 33 0 18 0 17 0 35,<br />

0 03 0 07 0 04 0 06 0 12<br />

the last be<strong>in</strong>g based on particularly small numbers. No clear trend emerges (a method for<br />

test<strong>in</strong>g for a trend is given <strong>in</strong> §15.2). About half the contribution to X 2 comes from the<br />

second age group (10±19 years) and there is some suggestion that the proportion of<br />

positives <strong>in</strong> this group is higher than <strong>in</strong> the neighbour<strong>in</strong>g age groups.<br />

To illustrate the use of (8.30), call the numbers of positives ri. Then<br />

X 2 ˆ…14 2 =101 ‡ ...‡ 6 2 =17 57 2 =288†=…0 1979†…0 8021†,

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

Saved successfully!

Ooh no, something went wrong!