109 Idiographic personality networks 5 to other nodes in the network and is thereby indicative of the overall network structure, equivalent to how items with high loadings explain a lot of variance in factor analysis (Christensen & Golino, 2021; Hallquist et al., 2021). To obtain the variability between estimated idiographic network structures across the 30 windows, we calculated the standard deviation between these 30 node strengths. We present these per participant, per node and averaged across nodes. To exploit the intuitiveness of networks we decided to visualize within-person variability. To do so, we counted all within-window statistically significant edges and presented these counts in summarizing networks of idiographic homogeneity. However, this only indicates variability between idiographic networks, but it does not show how stable network variability is over time (i.e., (non)stationarity). To illustrate change over time, the temporal sequence of networks was plotted in an animated video through the graph.animate function (Epskamp et al., 2012). These videos were created separately for each participant. Visual inspection of dynamic network videos provided us with a first impression on network changes over time and how this occurred. That is, some variability of networks that is inconsistent over time would indicate non-stationarity. Lastly, Kernel change-point analysis, as implemented in kcpRS (Cabrieto et al., 2022), was used to examine changes in the original idiographic network’s correlation structure within overlapping 30-day windows statistically differed from those of 1000 permutations at p < 0.05. This was done for each individual, allowing us to evaluate if (and when) there was at least one statistically significant change-point in the estimated idiographic network structure. Such a sudden change from one stable network structure to another stable structure would provide more conclusive evidence of non-stationarity and indicate that idiographic network change was potentially meaningful. 3. Results 3.1 Sample description The sample (N = 26) consisted of adolescents and young adults with a mean age of 22.7 years (SD = 5.5; range 15–33). Their average IQ was 72.3 (SD = 10.4). There were slightly more women (n = 15, 58 %) than men. The case records of 21 participants (81 %) showed one or more DSM-5 based diagnoses comorbid to their intellectual disability. We counted 14 unique comorbidities, of which posttraumatic stress disorder (n = 7) and autism spectrum disorder (n = 4) were the most recurring. A personality profile was estimated for each participant, based on the relative difference of their scores on the SURPS (Woicik et al., 2009) compared
RkJQdWJsaXNoZXIy MTk4NDMw