Increasing Inequality in Social Exclusion Occurrence: The Case of Sweden During 1979–2003

Table 8 Regression model when interaction terms for three variables are included

Variable	Odds ratio	P-value	Odds ratio	P-value
Year 1979
Years 1986–1987	0.752	0.004	0.867	0.248
Years 1994–1995	0.974	0.780	1.233	0.231
Years 2002–2003	0.996	0.973	1.343	0.217
Women
Men	1.604	0.000	2.019	0.000
Time * Men			0.915	0.166
55–64 Years
25–34 Years	1.232	0.081	1.240	0.074
35–44 Years	1.210	0.107	1.227	0.086
45–54 Years	1.009	0.931	1.033	0.760
Couples without children
Couples with children	2.017	0.000	3.600	0.000
Singles without children	3.290	0.000	3.057	0.000
Singles with children	5.549	0.000	2.897	0.009
Time * Couples with children			0.788	0.005
Time * Singles without children			1.021	0.817
Time * Singles with children			1.269	0.078
Swedes
1st-Generation Swedes	1.412	0.055	1.264	0.621
Immigrants	3.864	0.000	2.705	0.000
Time * 1st-generation Swedes			1.050	0.748
Time * Immigrants			1.149	0.095
College
University	1.184	0.333	1.165	0.383
High school	1.510	0.010	1.468	0.017
Vocational school	1.566	0.002	1.549	0.002
Compulsory school	2.185	0.000	2.159	0.000
Not completed education	4.811	0.000	4.642	0.001
Higher white collars
Middle white collars	1.504	0.022	1.505	0.023
Lower white collars	1.589	0.018	1.590	0.019
Skilled blue collars	2.982	0.000	2.978	0.000
Unskilled blue collars	2.979	0.000	3.029	0.000
Self-employed	1.945	0.002	1.974	0.002
Farmers	2.575	0.003	2.581	0.003
Students	3.798	0.000	3.882	0.000
Others	8.707	0.000	9.492	0.000
Constant	exp(−5.395)		exp(−5.592)
Log-likelihood	−4,189.179		−4.175.300
Log-likelihood score			27.758	0.000

Note Significant parameters at the 5% level are in bold. The log-likelihood score is a test of whether the inclusion of more explanatory variables significantly improves the model fit. The log-likelihood score is a comparison with the first model that does not include any interaction terms