Table 3.
Regression coefficients from multilevel models for expected BMI in kg/m2 and WHtR for women in the three cohorts of the West of Scotland Twenty-07 study, 1987–2008
BMI† | WHTR‡ | |||||
---|---|---|---|---|---|---|
Model 1 | Model 2 | Model 3 | Model 1 | Model 2 | Model 3 | |
Main effects | ||||||
Age | 0.215*** | 0.055+ | 0.048 | 0.493*** | 1.158*** | 1.147*** |
Age2 | −0.003*** | −0.007*** | −0.006*** | −0.001** | 0.016*** | 0.016*** |
Manual social class | 0.385+ | 0.388+ | 0.961*** | 0.998*** | 0.964** | 1.709*** |
1950s cohort | −3.721*** | −2.171*** | −7.827*** | −6.874*** | −12.984*** | −13.246*** |
1930s cohort | −6.336*** | −5.076*** | −5.364*** | −11.971*** | −10.704*** | −11.055*** |
Interactions | ||||||
Cohort×age | ||||||
1950s×age | 0.167*** | 0.160*** | −0.672*** | −0.682*** | ||
1950s×age2 | 0.003** | 0.003** | −0.002 | −0.002 | ||
1930s×age | 0.183*** | 0.175*** | −1.452*** | −1.464*** | ||
1930s×age2 | 0.003** | 0.004** | 0.001 | 0.001 | ||
Social class×age | ||||||
Manual×age | 0.023* | 0.03* | ||||
Manual×age2 | −0.001* | −0.000 | ||||
Constant | 29.928*** | 28.496*** | 28.314*** | 56.980*** | 62.605*** | 62.363*** |
Random effects | ||||||
Random intercept area | 0.413 | 0.417 | 0.400 | 0.955 | 0.930 | 0.927 |
Random intercept person | 5.274 | 5.274 | 5.271 | 6.828 | 6.886 | 6.885 |
Random slope age | 0.150 | 0.150 | 0.149 | 0.174 | 0.180 | 0.179 |
Correlation intercept slope | 0.732 | 0.730 | 0.732 | 0.810 | 0.802 | 0.803 |
Residual | 1.748 | 1.742 | 1.742 | 3.497 | 3.370 | 3.370 |
Likelihood | −21 066.572 | −21 053.371 | −21 045.915 | −25 399.418 | −25 219.131 | −25 214.27 |
χ2 (df) | 26.40 (4) | 14.91 (2) | 360.57 | 9.72 | ||
p Value forχ2 | <0.0001 | 0.0006 | <0.0001 | 0.0077 |
*p<0.05, **p<0.01, ***p<0.001.
†Measurement occasions N=8417, individuals N=2245, areas N=62.
‡Measurement occasions N=8347, individuals N=2244, areas N=62.
BMI, body mass index, WHtR, waist-to-height ratio.