116 Chapter 5 between all possible overlapping 30-day windows along each participant’s 60-day timeline. Although the network structures of some individuals changed more drastically over time than for others (e.g., the videos and Table 1 show that there is more consistency over time for participant #24 than for participant #23), some non-stationarity was evident in all participant’s network structures. Nevertheless, Kernel change-point analyses revealed no statistically significant change-points in pairwise correlations between windows for all but one participant (except #28). This indicates that – although different time-windows for each person yield (sometimes vastly) different network structures (see dynamic network videos) – this withinperson heterogeneity cannot be reduced to a single time-point that marks a change from one stable phase to another.
RkJQdWJsaXNoZXIy MTk4NDMw