Table 2.
True estimated bias for each of the 16 NSFG proportions, along with [MUB(0), MUB(1)] intervals based on the normal model, [MUBP(0), MUBP(1)] intervals based on the probit model and 95% credible intervals for MUBP based on the fully Bayesian approach†
| Binary NSFG variable (males or females) | Cross-validated biserial correlation (Y,X) | Population proportion | Smartphone proportion | True estimated bias (× 1000) | Results for normal model (MUB) | Results for probit model (MUBP) | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| [MUB(0), Cover MUB(1)] TEB? | [MUBP(0), MUBP(l)] and Bayesian credible intervals for selected limits | MML interval cover TEB? | 95% credible interval for MUBP | Bayesian interval cover TEB? | |||||||
| Never been married (males) | 0.817 | 0.566 | 0.555 | −11 | [−8, −21] | Yes | [−10, −14] | Yes | [−7, −16] | Yes | |
| Never been married (females) | 0.726 | 0.468 | 0.466 | −2 | [−1,−4] | Yes | [−2, −5] | Yes | [i, −7] | Yes | |
| Age = 30–44 years (males) | 0.654 | 0.435 | 0.433 | −2 | [16,47] | No | [16, 39] | No | [14, 38] | No | |
| [(13,19), (31,46)] | |||||||||||
| Age = 30–44 years (females) | 0.612 | 0.467 | 0.460 | −8 | [8,29] | No | [8, 24] | No | [7, 24] | No | |
| [(6,11), (17,31)] | |||||||||||
| Currently employed (males) | 0.603 | 0.689 | 0.729 | 40 | [16,58] | Yes | [16, 46] | Yes | [16, 45] | Yes | |
| Children present in HU (males) | 0.573 | 0.371 | 0.366 | −5 | [−2, −10] | Yes | [−2, −4] | Closet | [3, −10] | Yes | |
| [(−5,0), (−15,7)] | |||||||||||
| Currently employed (females) | 0.482 | 0.626 | 0.657 | 31 | [12,74] | Yes | [12, 50] | Yes | [11,47] | Yes | |
| Children present in HU (females) | 0.454 | 0.548 | 0.538 | −10 | [−10.5, −76] | No | [−10, −47] | Yes | [−10, −45] | Yes | |
| ‘Other’ race (females) | 0.451 | 0.553 | 0.562 | 9 | [11,85] | No | [11,54] | Close‡ | [9, 51] | Yes | |
| [(9,13), (42,65)] | |||||||||||
| ‘Other’ race (males) | 0.410 | 0.590 | 0.596 | 6 | [15, 135] | No | [14, 102] | No | [11,85] | No | |
| [(12,17), (78,129)] | |||||||||||
| Education: ‘some college’ (males) | 0.368 | 0.299 | 0.322 | 23 | [5,67] | Yes | [5, 17] | Close‡ | [3, 21] | Close‡ | |
| [(3,7), (5,29)] | |||||||||||
| Education: ‘some college’ (females) | 0.340 | 0.328 | 0.342 | 14 | [1,22] | Yes | [2, 16] | Yes | [0, 16] | Yes | |
| Region =‘south’ (females) | 0.274 | 0.438 | 0.445 | 7 | [−3, −65] | No | [−3, −36] | No | [−2, −34] | No | |
| [(−5,−2), (−52,−20)] | |||||||||||
| Region =‘south’ (males) | 0.253 | 0.418 | 0.431 | 13 | [−1, −31] | No | [−1, −20] | No | [4, −26] | No | |
| [(−3,1), (−46,10)] | |||||||||||
| Income: $20000–59999 (males) | 0.249 | 0.417 | 0.422 | 5 | [−3, −72] | No | [−3, −123] | No | [−1, −120] | No | |
| [(−5,−1), (−130,−38)] | |||||||||||
| Income: $20000–59999 (females) | 0.156 | 0.388 | 0.393 | 5 | [0, 34] | Yes | [1, 72] | Yes | [−2, 72] | Yes | |
†
Values multiplied by 1000.
‡
Close, allowing for uncertainty in the input estimates (see the Bayesian credible intervals for selected limits).