“…Finally, we showed how the network comparison test can be used to compare network models across samples (Van Borkulo et al, ). This approach is similar to item response theory Differential Item Function (DIF) testing (Hambleton and Jones, ; Embretson and Yang, ) or invariance testing in structural equation modeling (Marsh, Morin, Parker, & Kaur, ; Chin, Mills, Steel, & Schwarz, ). We found, in these data, that the structure, but not global strength, varied across clinical versus nonclinical samples, meaning that there were different maintaining symptoms and interactions between symptoms based on the type of sample.…”