Proposing network analysis for early life adversity: An application on life event data.

Area of Science:

• Psychology
• Epidemiology
• Network Science

Background:

• Traditional methods for modeling early life adversity (ELA) struggle to capture its complex nature.
• Existing approaches include sum-scores, latent class/trajectory models, single-adversity, and factor-analytical methods.

Purpose of the Study:

• To propose and apply network analysis as an alternative method for modeling ELA.
• To construct a network of fourteen adverse events (AEs) before age 16 using data from the TRacking Adolescents Individual Lives Survey (TRAILS).
• To compare network analysis findings with tetrachoric correlation analyses to understand associations between AEs.

Main Methods:

• Network analysis was used to model ELA.
• Fourteen adverse events (AEs) from the TRAILS dataset (N=1029) were included.
• Tetrachoric correlation analyses were performed for comparison.

Main Results:

• The ELA network revealed direct and complex indirect relationships between AEs, with fifteen edges emerging.
• While correlations suggested many AEs were associated, network analysis showed some associations were explained by interactions with other AEs (e.g., parental divorce mediating parental addiction and familial conflicts).

Conclusions:

• Network analysis effectively captures the complex, interconnected nature of ELA.
• This method provides valuable insights into the mechanisms underlying associations between AEs and potential life course outcomes.
• Future research should refine network model estimation, selection, and sample size requirements.