114 Chapter 5 strengths = 3.92). For this participant, in particular nervousness was extremely variable (SD node strength = 12.79). The most time-invariant network structures in our sample were those of participant #24 and #29 (average node strengths SD of 0.17 and 0.18, respectively). There was no correlation between idiographic network variability (i.e., average SD of all node strengths) with any of the four trait levels (anxiety sensitivity, negative thinking, impulsivity, sensation seeking as measured with the SURPS). What does this idiographic network variability over time look like? So far, all idiographic networks (e.g., the two examples in Figure 1) summarize partial correlations for the entire 60-day timeline. Now, we estimate partial correlations between the six variables, per individual, within each 30-day epoch on that 60-day timeline, and visualize these as networks. Figure 4 shows the idiographic network structures for day 1–30, day 16–45, and day 31–60 for the same two participants as depicted in Figure 1. Based on the SURPS, participant 1’s personality profile (i.e., the highest dimension z-score) was anxiety sensitive. Associations with nervousness – an item derived from the anxiety sensitivity scale in the SURPS – changed throughout her 60-day timeline. On average, between day 1 and day 30, she reported higher levels of nervousness on days with higher levels of restlessness evidenced by the blue edge between those nodes. When estimating the network in other window (day 16–45 or 31–60), the association with restlessness had disappeared. In the last time window (day 31–60), the positive association between nervousness and worrying appears. From all five bivariate associations of this participant in Figure 1, none occur in each of the 30 windows. That is, the rightmost panel in Figure 4 shows no count of statistically significant edges across all windows that adds up to 30. Scores on the SURPS of Participant 2 resulted in personality profile impulsivity. Figure 4 demonstrates that, in three example 30-day windows, he had a positive association between the two impulsivity items (doing things without thinking and doing things that were later regretted). In fact, these were significant across all 30 windows (see rightmost panel Figure 4). The overall structure of the network is very similar between the windows 1–30 and 16–45. However, the connectivity between all six variables substantially increased in the network that reflects day 31 to day 60. The structure of his networks changed over time, albeit to a lesser extent than for Participant 1. Nevertheless, we reach the same conclusion: the network that summarizes his 60 days (Figure 1) differed from that of the 30-day epochs within his entire timeline.
RkJQdWJsaXNoZXIy MTk4NDMw