. 2015 Jul 22;15:82. doi: 10.1186/s12903-015-0066-y

Table 3.

Logistic regressions of ECC on socioeconomic variables

Socioeconomic background		Adjusted estimated prevalence	Likelihood ratio test p-value	Effect size for socioeconomic variable of interest¹	Effect size for the complete model¹
Parents’ level of education			< .001	3.14 %	10.70 %
	Incomplete compulsory education	0.61^a
	Compulsory education	0.25^b
	Apprenticeship (vocational school)	0.23^b
	Secondary education (high school diploma)	0.25^b
	Tertiary education (university)	0.13^c
Parents’ professional level			< .001	3.74 %	11.26 %
	Unemployed	0.42^a
	Lower occupations	0.23^b
	Intermediate occupations	0.12^c
	Higher occupations	0.10^c
	Self-employed	0.17^b,c
Family income			< .001	3.35 %	10.94 %
	Low (< CHF 4,000)	0.36^a
	Medium (CHF 4,000 to CHF 6,000)	0.17^b
	High (> CHF 6,000)	0.10^b
	Other	0.15^b
	Did not want to answer	0.18^b
Parents’ literacy			< .001	3.45 %	11.00 %
	Problems	0.37^a
	No problems	0.17^b
Parents’ immigration status			< .001	7.74 %	14.96 %
	Developing countries (HDI < 0.8)	0.40^a
	Developed countries (HDI > 0.8) and Swiss citizens	0.13^b

^{a, b, c}For significant pairwise comparisons, a same subscript letter within a column denotes that proportions did not differ; two different subscript letters denote that proportions differed at the 0.05 level

¹Effect size measured with the McFadden pseudo R-square. The whole model included all demographics and covariates