Table 2 Components of design effect associated with sample design
From: Developing the design of a continuous national health survey for New Zealand
Component of design effect | All ethnicities | Māori | Pacific | Asian |
|---|---|---|---|---|
Varying weights due to disproportionate allocation to strata (a) | 1.12 | 1.13 | 1.06 | 1.06 |
Over- or undersampling due to disproportionate allocation to strata (b) | 1.00 | 0.90 | 1.13 | 1.19 |
Varying weights due to 1/household sampling (c) | 1.29 | 1.33 | 1.37 | 1.30 |
Over- or undersampling due to 1/household sampling (d) | 1.00 | 1.00 | 1.25 | 1.15 |
Varying weights due to unequal probabilities of selection of households within strata (e) | 1.17 | 1.14 | 1.08 | 1.06 |
Over- or undersampling due to unequal probabilities of selection of households within strata (f) | 1.00 | 0.71 | 0.68 | 0.91 |
Unequal probability of selection of households within strata (product of e and f) | 1.17 | 0.81 | 0.74 | 0.96 |