Table 1

The probability* of the true population value of GAM exceeding the 10% threshold for different counts of GAM cases and different design effects in a 33 × 6 (33 clusters of 6 children) survey.

Observed GAM count
GAM prevalence
Probability of GAM exceeding 10% threshold, %





DEFF** = 1
DEFF** = 1.1
DEFF** = 1.2
DEFF** = 1.3

14
7.1%
9.0%
10.0%
10.9%
11.9%

15
7.6%
13.2%
14.3%
15.4%
16.3%

16
8.1%
18.7%
19.9%
20.9%
21.8%

17
8.6%
25.6%
26.6%
27.5%
28.2%

18
9.1%
33.6%
34.4%
35.0%
35.5%

19
9.6%
42.5%
42.9%
43.2%
43.5%

20
10.1%
51.9%
51.8%
51.7%
51.6%

21
10.6%
61.0%
60.6%
60.2%
59.8%

22
11.1%
69.6%
68.9%
68.1%
67.5%

23
11.6%
77.1%
76.2%
75.3%
74.5%

* Based on the one-sided t-test with 32 degrees of freedom

** Design effect

Bilukha and Blanton Emerging Themes in Epidemiology 2008 5:25   doi:10.1186/1742-7622-5-25