The temporal structure of state self-esteem variability during parent–adolescent interactions: more than random fluctuations

Naomi de Ruiter


By: Naomi de Ruiter, Ruud den Hartigh, Ralf Cox, Paul van Geert, & Saskia Kunnen
University of Groningen, The Netherlands

Self-esteem is conceptualized as having both a trait element (characterized as relatively stable and predictable across time), as well as a state element (characterized by fluctuations from moment to moment and a high level of variability) (Donnellan, Kenny, Trzesniewski, Lucas, & Conger, 2012). While the number of theoretical and empirical studies focusing on state self-esteem is increasingly growing, these studies tend to focus on the magnitude of state self-esteem variability (e.g., Leary & Downs, 1995). To date, very little theoretical or empirical research has been done concerning the nature of the moment-to-moment fluctuations that occur in state self-esteem, which we refer to as the temporal structure of state self-esteem variability.
The common conceptualization of state self-esteem stems from the notion that state self-esteem is the “barometric” element of self-esteem, which is variable across time and contexts and fluctuates around the relatively stable “baseline” level of self-esteem (Rosenberg, 1986a). State self-esteem is therefore approached as the “error” around (and independent from) – what is thought to be – a more meaningful baseline level that is trait self-esteem, where the “error” is contextually-based error. Following this basic theory, state self-esteem represents a short-lived experience, which – given the absence of a new contextual cue – will return back to the baseline level (Alessandri & Caprara, 2012). Given this conceptualization, the variability of state self-esteem should resemble white noise (Diniz et al., 2011; Gilden, 2001; Stadnitski, 2012; Van Orden, Holden, & Turvey, 2003, 2005), which is temporally random variability that is created when there is no carry-over effect from one state to the next.
We question the assumption that state self-esteem variability is purely a function of exogenous events, as well as the assumption that the temporal structure of the resulting variability is random (i.e., white noise). We examine this in the developmental context of adolescence, as a pivotal period for self-esteem development. Alternatively, we posit that each state self-esteem event is in itself a process, and that this process interacts with neighboring (i.e., future) state self-esteem processes. These dynamics are defined as interaction-dominant dynamics, where the coordination of the process at large is a function of the internal dynamics, which occur within a context, but which are not a function of the context alone (Van Orden et al., 2003). From this conceptualization, state self-esteem exhibits both short-term and long-term carry-over effects. We suggest, therefore, that state self-esteem is a self-coordinating process, rather than a passively reactive (i.e., stimulus-response like) and random process. Many human processes that have recently been conceptualized as depending on interaction-dominant dynamics have been found to exhibit pink noise, which is structured variability characterized by correlated activity across many time scales (Van Orden et al., 2003; Wijnants, Cox, Hasselman, Bosman & Van Orden, 2012). Moreover, the presence of pink noise is indicative of normative developmental processes, as its presence is exclusively found in healthy and well-coordinated systems (Herman, Giladi, Gurevich, & Hausdorff, 2005; Wijnants, Hasselman, Cox, Bosman, & Van Orden, 2012).

Design and Research Questions
We adopted a qualitative phenomenological approach to state self-esteem across real time, where adolescents’ (N = 13) positive and negative emotional and behavioral self-experiences that are expressed during interaction with their parent are observed and coded from moment-to-moment. Adolescent-parent dyads were video-recorded in their home environment during a semi-naturalistic interaction, including a neutral-conflict-neutral discussion, consecutively.
The observational videos were subsequently coded. Adolescents’ emotional and behavioural expressions of state self-esteem were coded for every action or utterance. State self-esteem was calculated based on the self-esteem related emotions/behaviour for each second of the interaction, resulting in a state self-esteem time series for every adolescent. Furthermore, static and context-independent levels of autonomy were gathered by means of a questionnaire administered before the filmed interaction took place; and static and context-dependent measures of state self-esteem and trait self-esteem were gathered before and after the filmed-interaction took place.
Detrended fluctuation analysis (DFA; Peng, Havlin, Stanley, & Goldberger, 1995) was applied to each state self-esteem time series. The DFA reveals a relation between different window sizes of data and the average fluctuation of the windowed data. The DFA produces a DFA exponent, which indicates whether the structure of variability resembles white noise (i.e., a highly random structure) or pink noise (i.e., long-range correlations; Hasselman, 2013; Wijnants et al., 2012). The DFA exponents of the time series were then compared to randomized surrogate time series (i.e. shuffled within-individuals) that functioned as a control group.
We hypothesized that the temporal structure of state self-esteem variability would be structured, rather than random, thereby resulting in long-range correlations as indicated by the presence of pink noise (Hypothesis 1). We also hypothesized that the temporal structure of state self-esteem would be a distinct concept from the level of self-esteem, such that there are no significant correlations between the temporal structure of state self-esteem (i.e. the DFA exponent) and the static measures of self-esteem levels (hypothesis 2). Finally, we hypothesized that the presence of pink noise would correspond with an indicator of healthy adolescent development, i.e. static and context-independent autonomy levels (hypothesis 3).

We found that the variability of adolescents’ state self-esteem during a parent-child interaction can be characterized by ‘pink noise’, and that the structure of state self-esteem variability is significantly different from the structure of variability that would be exhibited if state self-esteem was characterized by random fluctuations with no carry-over effect from one moment to the next, i.e., white noise (hypothesis 1). In addition, we found that the temporal structure of state self-esteem is a distinct concept from the valence level of (state and trait) self-esteem (hypothesis 2), and that the closer that state self-esteem came to approaching pure pink noise, the higher the adolescents’ scores for autonomy (hypothesis 3).

These results show that the temporal structure of adolescent state self-esteem variability has been unnecessarily disregarded (as ‘random’) in empirical studies of state self-esteem. Our results bring the passive and random nature of state self-esteem into question, and provide evidence that state self-esteem, as a real-time process, might be better conceptualized as an intrinsically dynamic and active process. This is an important shift in the theoretical conceptualization of the nature of state self-esteem. Moreover, these results indicate that this type of state self-esteem structure is a signature of healthy, efficient, and well-coordinated behavior during parent-child interactions.


Link to paper > Contact > Website >
Back to Spotlights