How does cross-sectional sampling bias our understanding of adolescent romantic relationships?: An agent-based...

Area of Science:

• Developmental Psychology
• Computational Social Science
• Adolescent Studies

Background:

• Adolescent romantic relationships are crucial for development but are often brief and context-dependent.
• Individual differences and age-related changes complicate sampling in adolescent relationship research.
• Most adolescents experience multiple romantic relationships, raising questions about which to study.

Purpose of the Study:

• To investigate sampling biases in studying adolescent romantic relationships.
• To compare cross-sectional sample estimates with population parameters using a computational model.
• To understand how computational models can illuminate sampling issues in complex social phenomena.

Main Methods:

• An agent-based computational model simulated 1000 individuals over 60 months, including varied relationship durations.
• The model generated 1000 iterations to capture relationship formation and dissolution over 5 years.
• Two sets of 1000 cross-sectional samples were drawn to compare with population parameters.

Main Results:

• Cross-sectional samples systematically over-represented longer adolescent romantic relationships.
• The detection of individual differences in relationship duration and partner number varied significantly with time.
• Computational models revealed systematic distortions in understanding adolescent relationships due to sampling methods.

Conclusions:

• Cross-sectional studies may be time-sensitive, potentially distorting the understanding of adolescent romantic relationships.
• Oversampling longer relationships in cross-sectional research can skew findings.
• Computational modeling provides valuable insights into sampling biases and complex social dynamics.