Table 4.

Logistic regression models predicting low educational attainment using the number of household smokers in childhood, stratified by gender.

	Crude model		Adjusted models
			Model 1		Model 2
	OR	(95% CI)	OR	(95% CI)	OR	(95% CI)
Men (n = 1077)
Number of household smokers
≥ 3	5.36	(3.08–9.31)	5.29	(3.01–9.29)	3.97	(2.21–7.12)
2	2.67	(1.79–3.97)	2.64	(1.77–3.96)	2.22	(1.47–3.36)
1	1.34	(0.97–1.85)	1.31	(0.94–1.81)	1.24	(0.89–1.73)
0	1	reference	1	reference	1	reference
Women (n = 1021)
Number of household smokers
≥ 3	3.94	(2.18–7.10)	3.71	(2.03–6.78)	3.36	(1.83–6.20)
2	3.06	(2.01–4.65)	3.11	(2.03–4.78)	2.90	(1.88–4.46)
1	2.12	(1.47–3.06)	2.04	(1.41–2.96)	1.93	(1.33–2.81)
0	1	reference	1	reference	1	reference

OR: odds ratio; CI: confidence interval.

Notes: Model 1 adjusted for year of survey and geographical area; Model 2 adjusted for Model 1 + current smoking status.

Study conducted 2014–2017 among young adults (19–20 years) in Hidaka District, Wakayama Prefecture, Japan.