2.3. Network analysis

Typically, network researchers compute centrality indices of the network structure to index the (relative) importance of single variables (Newman, ²⁰⁰¹; Opsahl et al., ²⁰¹⁰). In the present study, we calculated the strength centrality index as well as one‐step EI. Strength quantifies how strongly a variable is directly associated to other variables (i.e., the sum of the absolute value of edge parameters; Epskamp et al., ²⁰¹⁷). Using the bootstrapped difference test (described in Epskamp et al., ²⁰¹⁸), we tested if the Healthy Adult and dysfunctional modes significantly differed in strength. This test estimates a 95% confidence interval (CI) of the difference in strength between two variables in a network. If this CI does not contain 0, then we reject the null hypothesis that two variables do not differ in strength.

Additionally, we calculated EI. EI is similar to the strength centrality index, but differs in that it takes into account negative associations between variables. It is calculated by summing all associations without first taking their absolute value (for a more in‐depth explanation and validation study, see Robinaugh et al., ²⁰¹⁶). This value can take positive and negative values, which indicate that the sum of associations between one schema mode and the others is positive or negative. As such, EI indexes how and how strongly a mode is associated to other modes.