Composition and decomposition in nonmarital fertility.

Area of Science:

• Sociology
• Demography
• Family Studies

Background:

• Previous research by Gray, Stockard, and Stone (2006) proposed a nonlinear model for nonmarital fertility.
• Ermisch (2009) critiqued this model, suggesting an incorrect testing methodology.

Purpose of the Study:

• To identify limitations in the generalizability of a nonmarital fertility model.
• To demonstrate that a model fitting a population may not fit its subgroups or aggregations.

Main Methods:

• Theoretical analysis of statistical model generalizability.
• Examination of hierarchical data structures in demographic research.

Main Results:

• The study identifies a fundamental flaw in the application of the nonmarital fertility model across different population levels.
• A model validated for a general population may not accurately represent its constituent subgroups.
• Conversely, a model fitting subgroups may not be applicable to the aggregated population.

Conclusions:

• The generalizability of nonmarital fertility models is contingent on population structure.
• Researchers must consider subgroup and aggregation effects when applying demographic models.
• Further research is needed to develop models that account for varying population levels.