Daan Hulsmans

119 Idiographic personality networks 5 if this was the case, we would have expected to see change-points from one stable network to another stable network over time, which our change-point analyses did not reveal. We therefore deem it unlikely that interventioninduced learning was responsible for the variability in the idiographic networks over time. The second explanation, being that unobserved situational changes underlie non-stationarity, is more probable. Dynamic systems approaches to personality all stress the inseparability between the internal personality system and situational features (Cervone, 2005; Danvers et al., 2020; DeYoung, 2015; Fajkowska, 2015; Mischel and Shoda, 1995; Nowak et al., 2005; Read et al., 2017; Sosnowska et al., 2019). However, current idiographic personality network studies (Beck and Jackson, 2020; Costantini et al., 2019; Lazarus et al., 2020) – our study included – did not model the situations. Previously, the situational if-then signatures (e.g., if John is excited at a party with friends, then he tends to blurt things out, but if he is excited at work then he is restrained; Mischel & Shoda, 1995) have even been projected on a temporal lag-1 personality network structure (if John is excited now then he blurts things out later; Beck & Jackson, 2020), without including variable coding for situational features. Notably, empirical studies into situational if-then signatures (e.g., Shoda et al., 1994) relied on psychological perceptions of situations, which of course also differs between persons (what is exciting for John may not be exciting for someone else). The CAPS, KAPA and other dynamic systems theories posit that the internal personality system – in continuous interaction with the situations – produces behavioral patterns which are variable across different situations but relatively stable within (similar) situations (Cervone, 2005; Mischel and Shoda, 1995). As such, the non-stationarity we found most likely shows that an idiographic network does not capture (changes in) the underlying personality system, but rather reflects unobserved changes in situational features from day to day. This does not mean that networks in principle cannot capture the underlying (relatively stable) personality system. What we need is theory to inform research about three crucial elements. Firstly, we need to know the timeframe within which a stable personality network can be found. With our once-per-day measurement frequency and duration of two months, we found high variability of estimated idiographic network structures over time. Had we measured longer, would that have resulted in a stationary pattern? The evidence for long-term trait-stability is not entirely clear-cut. Life events (e.g., graduation, marriage, parenthood) have been associated with withinperson changes at the trait-level (Bleidorn et al., 2018; Wrzus & Roberts, 2017). Thus, over a lifespan, the mean-level of personality traits change within-

RkJQdWJsaXNoZXIy MTk4NDMw