Network analysis is quickly gaining popularity in psychopathology research as a method that aims to reveal causal relationships among individual symptoms. To date, four main types of psychopathology networks have been proposed: (1) association networks, (2) regularized concentration networks, (3) relative importance networks, and (4) directed acyclic graphs.We examined the replicability of these analyses based on symptoms of major depression and generalized anxiety between and within two highly similar epidemiological samples (i.e., the National Comorbidity Survey -Replication [n = 9282] and the National Survey of Mental Health and Wellbeing [n = 8841]). While association networks were stable, the three other types of network analysis (i.e., the conditional independence networks) had poor replicability between and within methods and samples. The detailed aspects of the models-such as the estimation of specific edges and the centrality of individual nodes-were particularly unstable. For example, 44% of the symptoms were estimated as the "most influential" on at least one centrality index across the six conditional independence networks in the full samples, and only 13-21% of the edges were consistently estimated across these networks.One of the likely reasons for the instability of the networks is the predominance of measurement error in the assessment of individual symptoms. We discuss the implications of these findings for the growing field of psychopathology network research, and conclude that novel results originating from psychopathology networks should be held to higher standards of evidence before they are ready for dissemination or implementation in the field.Key words: Network analysis; psychopathology; causal inference; psychopathology networks; replication crisis.
REPLICABILITY OF PSYCHOPATHOLOGY NETWORKS 3General Scientific Summary: A statistical method called network analysis is quickly gaining popularity for analyzing the relationships between symptoms of mental disorders.This study found that popular network analysis methods produce unreliable results, particularly for the symptom-level aspects of the models. We highlight the need to be particularly cautious in interpreting, disseminating, or implementing results that arise from psychopathology networks.
REPLICABILITY OF PSYCHOPATHOLOGY NETWORKS 4
Evidence that Psychopathology Symptom Networks do not ReplicateThe popularity of network analysis is spreading quickly in the study of psychopathology. In particular, a growing number of studies using cross-sectional analyses of networks of psychopathology symptoms have appeared in the literature since Cramer et al.