Table 2.

Odds ratios (95% CI) and P values of reducing the food insecurity score by duration of participation in the program, paternal employment status, maternal country of birth, region, type of household, number of children in the family and paternal educational level

Variable	Category	OR (95% CI)	P values
Duration of participation in the program (months)		1.063 (1.016–1.112)	0.008
Paternal employment Reference level: Never unemployed	Unemployed at the pre- and post-intervention period	0.890 (0.686–1.154)	0.379
	Unemployed at the pre- but not in the post-intervention period	1.759 (1.179–2.623)	0.006
	Unemployed at the post- but not in the pre-intervention period	1.017 (0.663–1.560)	0.938
Maternal country of birth Ref. level: Greece	Other than Greece	1.234 (1.009–1.509)	0.040
Region Ref. level: Attica	Thessaloniki	0.693 (0.538–0.892)	0.004
	Rest of Greece	0.760 (0.572–1.010)	0.058
Type of household Ref. level: Living with either or both of its parents	Living with another family without its parents	2.584 (1.083–6.168)	0.032
Number of children in the familya Ref. level: <4	≥4	1.520 (1.108–2.084)	0.009
Paternal educationb		0.889 (0.808–0.979)	0.016

No. of observations (excluding missing values for all explanatory variables) 2080; percentage correctly classified 61.3%. Statistically significant variables were identified using forward-stepwise (Wald) model selection. Explanatory variables considered: duration of participation in the program (months); region; child’s gender, age and country of birth; no. of children in the household and type of household; paternal and maternal education, employment and country of birth. Bold: statistical significant at 5% level.

a: Number of children in the family taken as continuous variable was not statistically significant, so different thresholds were tested.

b: Paternal education is taken as a continuous variable, higher values signalling higher educational level (taken as a categorical variable specific subcategories were not statistically significant).