. 2015 Dec 16;15:1244. doi: 10.1186/s12889-015-2608-2

Table 3.

Covariate adjusted prevalence of obesity (BMI ≥ 30)

	Men		Women
	2006	2012	2006	2012
Education level
No education	24.0^a# ± 3.2	26.6^a# ± 2.7	37.3^a ± 2.1	36.8^a ± 1.9
Elementary school	28.6^a# ± 2.6	27.9^a# ± 2.0	38.4^a ± 1.2	41.6^b* ± 1.1
Middle school	27.4^a# ± 3.1	29.9^a# ± 2.1	35.6^a ± 1.4	37.2^a ± 1.1
High school	27.2^a ± 3.0	29.8^a ± 2.4	30.1^b ± 1.6	34.3^a* ± 1.5
College/University	25.1^a ± 3.1	29.6^a ± 2.4	26.4^b ± 1.9	30.0^c ± 1.8
Marital status
Single	21.9^a# ± 2.5	26.7^a# ± 2.4	32.0^a ± 1.7	35.6^a ± 1.4
Married/cohabitating	28.0^b# ± 2.7	29.8^a# ± 1.9	35.4^a ± 0.9	37.8^a* ± 0.8
Widowed/separated	31.1^ab ± 5.2	26.8^a# ± 2.8	35.3^a ± 1.8	37.9^a ± 1.3
Occupational status
Other^†	28.6^a ± 1.9	28.7^a# ± 1.7	32.9^a ± 2.3	37.0^a ± 2.8
Student	28.5^a ± 5.5	24.6^a ± 4.0	32.3^a ± 5.2	35.4^a ± 4.5
Housekeeper	31.5^a ± 7.5	32.7^a ± 5.5	34.6^a ± 0.8	38.3^a* ± 0.8
Paid worker	24.6^a# ± 0.9	27.0^a#* ± 0.7	35.3^a ± 1.2	37.1^a ± 1.0
Overall prevalence	27.3 ± 2.5	28.9 ± 1.8	34.8 ± 0.7	37.4* ± 0.6

All estimates are covariate-adjusted prevalence ± standard errors obtained through predictive margins from a survey design-based multiple logistic regression model. Country region, area of residence, a wealth index and its squared term, age and age squared were included in the model along with the other covariates

^†Includes the retired, the disabled and workers in a family business without pay

Different letters indicate significant differences (P < 0.05) between covariate categories in the same sex and survey year

*P < 0.05 change from 2006 to 2012 within sex

^# P < 0.05 men vs. women at the same survey year