Table 3.
Estimated variance factors for the group variance models of current smoking status, and ln(cigarettes/day) for race/ethnicity, educational attainment, and household income (Equations 3–4). variance factor (95% CI uncorrected for multiple comparisons) Holm-adjusted p-valuesa
| current smoker status (N = 54024) | cigarettes per day (N = 10674) | ||||||
|---|---|---|---|---|---|---|---|
| state level terms | |||||||
| eσν,E | educational attainment | 1.017 | (1.017, 1.017) | ns | 1.009 | (1.009, 1.009) | ns |
| eσν,I | household income ($1000) | 1.012 | (1.012, 1.012) | nsc,1 | 1.004 | (1.004, 1.004) | nsc,3 |
| eσν,b | black | 1.12 | (1.117, 1.123) | ns | 1 | na | b |
| eσν,w | white | 1.072 | (1.072, 1.073) | ns | 1.011 | (1.011, 1.011) | ns c,4 |
| eσν,n | American Indian/Aleut, Eskimo | 1.617 | (1.434, 1.898) | ns c,2 | 1.130 | (1.128, 1.133) | ns c,2 |
| eσν,a | Asian/Pacific Islander | 1.415 | (1.378, 1.456) | ns | 1.065 | (1.065, 1.066) | ns c,4 |
| eσν,l | Latino | 1.034 | (1.034, 1.034) | ns c,2 | 1.114 | (1.113, 1.115) | ns |
| eσν,m | Non Latino | 1.026 | (1.026, 1.026) | ns c,2 | 1.032 | (1.032, 1.032) | ns c,1 |
| county level terms | |||||||
| eσμ,E | educational attainment | 1.041 | (1.041, 1.041) | 0.226 | 1.007 | (1.007, 1.007) | ns c,2 |
| eσμ,I | household income ($1000) | 1.030 | (1.030, 1.030) | 0.390 | 1.030 | (1.030, 1.030) | ns b |
| eσμ,b | black | 1.277 | (1.266, 1.289) | ns | 1.049 | (1.049, 1.050) | ns c,1 |
| eσμ,w | white | 1.190 | (1.187, 1.192) | 0.189 | 1.013 | (1.013, 1.013) | ns c,3 |
| eσμ,n | American Indian/Aleut, Eskimo | 1.927 | (1.700, 2.250) | 0.900 | 1.043 | (1.043, 1.044) | ns c,3 |
| eσμ,a | Asian/Pacific Islander | 1.137 | (1.133, 1.142) | ns | 1.056 | (1.055, 1.056) | ns c,3 |
| eσμ,l | Latino | 1.410 | (1.393, 1.428) | 0.061 | 1.100 | (1.099, 1.102) | ns c,1 |
| eσμ,m | Non Latino | 1.098 | (1.097, 1.099) | ns | 1.015 | (1.015, 1.015) | ns c,2 |
| individual level terms | |||||||
| eσε,E | educational attainment | 1.034 | (1.034, 1.034) | ns | |||
| eσε,I | household income ($1000) | 1.348 | (1.281, 1.433) | ns | |||
| eσε,b | black | 1.865 | (1.798, 1.939) | <0.001 | |||
| eσε,w | white | 1.796 | (1.68, 1.937) | 0.277 | |||
| eσε,n | American Indian/Aleut, Eskimo | 2.067 | (1.966, 2.182) | <0.001 | |||
| eσε,a | Asian/Pacific Islander | 1.925 | (1.802, 2.073) | 0.010 | |||
| eσε,l | Latino | 1.513 | (1.454, 1.58) | ns | |||
| eσε,m | Non Latino | 1 | na | b | |||
a
Holm adjustment is a sequential procedure with a stopping criterion; all parameter estimates with p-values larger than that of the first rejected p-value are rejected; such rejection is indicated by “ns”.
b
Zero terms indicate that there is no variance for the corresponding fixed effect at this level. (Duncan, Jones, & Moon, 1998; Twisk, 2006)
c
This term estimated zero variance for the indicated number of imputed data sets.