Entering variables in blocks.

noetsi

No cake for spunky
#1
I know a chi square difference test for nested models, but is that what is occurring here? They entered the predictors in two blocs.

For demographic and socioeconomic factors, the likelihood ratio test for model goodness-of-fit was significant, χ2 (14,N= 14,229) = 1,900.75, p < 0.001, indicating that demographic and socioeconomic characteristics contributed to the differentiation of employment status at application. The Nagelkerke R2 was 0.17. After adding the second block of service variables into the model, the likelihood ratio test remained significant, χ2 (38, N= 14,229) = 2,939.06, p < 0.001, indicating that the set of all explanatory variables in the model contributed to the differentiation of employment status at application. The Nagelkerke R2 was 0.26, indicating a large effect size for the overall model and indicating a 9% increase in the variance of employment status explained by adding service variables into the model. The change of likelihood ratio was significant, χ2 (24, N= 14,229) = 1038.31, p < 0.001, indicating that VR services were significantly associated with consumers’ employment status at application. The two groups of applicants (competitively employed vs. unemployed) were also differentiated, as we will see, based on the patterns of services received
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Yup - The first two are likely the model vs. an empty model and the last a direct comparison of them. As you know they all have to be nested versions of the comparison. Look at the degrees of freedom 14 + 24 = 38. But without reading the paper in it's entirety it is not fully known.