Table 3.
Random-effects linear regression models of the adjusted association between education and BMI over survey period
| Females | Males | |
|---|---|---|
| β (95% Confidence Interval) | β (95% Confidence Interval) | |
|
| ||
| <Primary School | Referent | Referent |
| Primary School | −0.08 (−0.30, 0.15) | −0.22 (−0.45, 0.00) |
| Secondary School | −0.45 (−0.76, −0.14) | 0.08 (−0.20, 0.36) |
| >Secondary School | −0.36 (−0.74, 0.02) | −0.07 (−0.46, 0.32) |
| Time × Primary School | 0.02 (0.01, 0.04) | 0.04 (0.02, 0.06) |
| Time × Secondary School | 0.03 (0.00, 0.05) | 0.04 (0.01, 0.06) |
| Time × >Secondary School | −0.03 (−0.06, −0.01) | 0.07 (0.04, 0.10) |
| Time | 0.11 (0.10, 0.12) | 0.14 (0.11, 0.18) |
| Mean* Time | −0.09 (−0.11, −0.07) | - |
| Time Squared | - | 0.00 (0.00, 0.00) |
| Birth Cohort (1= before 1955) | −2.98 (−4.19, −1.77) | 0.35 (0.06, 0.65) |
| Birth Cohort × Time | - | −0.04 (−0.06, −0.02) |
| Birth Cohort × Primary School | −0.11 (−0.41, 0.18) | - |
| Birth Cohort × Secondary School | 0.42 (−0.32, 1.16) | - |
| Birth Cohort × >Secondary School | 0.27 (−0.27, 0.82) | - |
| Urbanicity1 | 0.01 (0.01, 0.02) | 0.01 (0.00, 0.02) |
| Mean Urbancity1,2 | −0.01 (−0.02, −0.01) | 0.00 (−0.01, 0.01) |
| Birth Cohort × Urbanicity1 | −0.01 (−0.02, 0.00) | −0.02 (−0.03, 0.00) |
| Birth Cohort × Mean Urbanicity1,2 | 0.04 (0.03, 0.06) | 0.04 (0.03, 0.06) |
| Person-specific Mean Age1,2 | 0.20 (0.10, 0.30) | 0.38 (0.33, 0.43) |
| Person-specific Mean Age Squared1,2 | 0.00 (0.00, 0.00) | 0.00 (−0.01, 0.00) |
| Birth Cohort × Mean Age1,2 | 0.38 (−0.09, 0.84) | - |
| Birth Cohort × Mean Age Squared1,2 | −0.01 (−0.01, 0.00) | - |
| Current Smoker | - | −0.29 (−0.39, −0.20) |
| Intercept | 22.47 (22.17, 22.78) | 21.13 (20.86, 21.41) |
|
| ||
| Number of Observations | 23,068 | 18,693 |
| N | 7,314 | 6,492 |
1
Variable means are the person-specific/group mean value for these variables over the survey period. The coefficient on the variable for the person-specific mean represents the difference between the between and within effects for each variable. These are included for time-varying variables for which the between and within coefficients were significantly different.
2
Urbanicity, mean urbanicity, mean age and mean age squared were mean centered before inclusion in the regression.