Table 2.
Multilevel logistic model for daily breakfast consumption outcomes, MCMCa estimates (Monte Carlo standard error)
| Fixed effects | Model 1b | Model 2c | Model 3d | Model 4e | |
|---|---|---|---|---|---|
| Cons | 0.38 (0.07) | 0.46 (0.05) | 0.53 (0.05) | 0.49 (0.06) | |
| Age | -0.21 (0.01) | -0.21 (0.01) | -0.10 (0.02) | -0.10 (0.02) | |
| Age squared | 0.04 (0.01) | 0.04 (0.01) | 0.04 (0.01) | 0.04 (0.01) | |
| Sex (ref: Male) | Female | -0.53 (0.03) | -0.53 (0.03) | -0.71 (0.05) | -0.71 (0.05) |
| Family structure (ref: Both parents) | Single mother | -0.35 (0.04) | -0.35 (0.04) | -0.35 (0.04) | -0.14 (0.07) |
| Single father | -0.44 (0.09) | -0.44 (0.09) | -0.44 (0.09) | -0.13 (0.18) | |
| Step family | -0.25 (0.04) | -0.24 (0.04) | -0.24 (0.04) | -0.21 (0.09) | |
| Other | -0.20 (0.12) | -0.20 (0.12) | -0.21 (0.12) | 0.04 (0.32) | |
| Year (ref: 1994) | 1998 | 0.22 (0.08) | |||
| 2002 | -0.03 (0.08) | ||||
| 2006 | 0.05 (0.08) | ||||
| 2010 | 0.10 (0.08) | ||||
| Yearcontf | 0.02 (0.01) | 0.01 (0.01) | 0.02 (0.01) | ||
| Age*Yearcontf | -0.01 (0.001) | -0.01 (0.001) | |||
| Sex*Yearcontf (ref: Male*Yearcontf) | Female*Yearcontf | 0.02 (0.004) | 0.02 (0.005) | ||
| Family structure * Yearcontf interaction (ref: Both parents*Yearcontf) |
Single mother*Yearcontf | -0.02 (0.01) | |||
| Single father*Yearcontf | -0.03 (0.02) | ||||
| Step family*Yearcontf | -0.004 (0.01) | ||||
| Other*Yearcontf | -0.02 (0.02) | ||||
| Random effects | |||||
| Level 1 (child) varianceg | 1 | 1 | 1 | 1 | |
| Level 2 (school) variance | 0.056 (0.011) | 0.056 (0.011) | 0.041 (0.010) | 0.042 (0.010) | |
| Level 3 (region) | 0.028 (0.010) | 0.026 (0.009) | 0.026 (0.009) | 0.026 (0.009) | |
| h | 34815.5 | 34818.2 | 34791.4 | 34776.9 | |
| PDi | 255.2 | 254.7 | 214.2 | 219.9 | |
| DICj | 35070.7 | 35072.9 | 35005.6 | 34996.8 | |
avia Markov chain Monte Carlo (MCMC); estimates are based on a chain of length of 50,000 following a burn-in of 5,000
bModel 1 adjusts for age, age2, sex, grade, family structure and year as a categorical variable
cModel 2 adjusts for age, age2, sex, grade, family structure, year as a continuous variable, and year marker following 2002 when the question on breakfast consumption changed
dModel 3 adjusts for age, age2, sex, grade, family structure, year as a continuous variable, year marker for years following 2002, and interaction terms between year and age and year and sex
eModel 4 adjusts for age, age2, sex, grade, family structure, year as a continuous variable, year marker for years following 2002, interaction terms between year and age and year and sex, and an interaction term between year and family structure. Interaction between family structure and year marker for years following 2002 is not significant
fYearcont refers to continuous Year variable, ranging between 1 (1994) and 17 (2010)
gVariance at the child level is constrained to 1
h is the expectation of the deviance and is a measure of how well the model fits the data
iPD is the effective number of parameters
jDIC is the Deviance Information Criterion; the larger this is, the worse the model fit