Comparison of PTSD Symptom Centrality in Two College Student Samples
Network analysis is becoming more widely used as a method of understanding the structure and potential causal and influential relationships between symptoms within and across psychiatric syndromes such as posttraumatic stress disorder. Because large samples are often needed to yield more stable findings in network analysis, researchers often struggle to overcome limitations associated with small sample size. So, researchers often combine samples to ensure appropriate statistical power for analyses. Little research has been done, however, to determine whether such strategies are appropriate. The present study evaluates the network structure and indices of two college student samples (N = 668 and 456) from mid-sized cities to examine similarities and differences that might inform whether similar samples can be combined for network analysis. The findings suggest that the overall network structures are not different based on a network comparison analysis; however, centrality stability coefficients for centrality indices across both networks were below recommended cut-offs, indicating that the networks were generally unstable. We discuss the implications for these findings in the paper, highlighting that network comparison alone is likely insufficient in determining whether or not to combine samples for network analysis.
Journal of Psychopathology and Behavioral Assessment
Eddinger, J., McDevitt-Murphy, M., Hu, J., Jobe-Shields, L., & Williams, J. (2020). Comparison of PTSD Symptom Centrality in Two College Student Samples. Journal of Psychopathology and Behavioral Assessment, 42 (2), 354-363. https://doi.org/10.1007/s10862-020-09792-w