Table 2
|
Estimates of smoking prevalence for four possible samples from a population with 50% males, ω = 0.5, when the true smoking prevalence is π = 0.2 and male and female smoking prevalence are πM = 0.25 and πF = 0.15, respectively (πD = 0.10) |
||||||
| Possible samples |
||||||
|
|
||||||
| 1 |
2 |
3 |
4 |
|||
|
|
||||||
| Prop. Male |
 |
0.6 |
0.6 |
0.4 |
0.4 |
|
|
|
||||||
| Gender-specific prevalence estimator |
Male |
 |
0.2 |
0.3 |
0.15 |
0.3 |
| Female |
 |
0.25 |
0.1 |
0.2 |
0.1 |
|
| Difference |
 |
-0.05 |
0.2 |
-0.05 |
0.2 |
|
|
|
||||||
| Prevalence estimator |
Crude |
 |
0.2200 |
0.2200 |
0.1800 |
0.1800 |
| Adjusted (using πD) |
 |
0.2100 |
0.2100 |
0.1900 |
0.1900 |
|
Direct adjusted (using  ) |
 |
0.2250 |
0.2000 |
0.1750 |
0.2000 |
|
| Covariance adjusted |
 |
0.2248 |
0.2008 |
0.1752 |
0.1994 |
|
Li et al. Epidemiologic Perspectives & Innovations 2008 5:2 doi:10.1186/1742-5573-5-2 |
||||||