Using Multilevel Systems Theory to Integrate Dynamic Person-in-Context Systems

From a multilevel systems theory (MST) perspective (cf. Anderson, 1998; Cacioppo & Bernstein, 1992; Campbell, 1990; Derryberry & Tucker, 1991; Pattee, 1973; Peck, 2007; Salthe, 1985; Schneirla, 1949; Sheldon, 2004), the person-situation debate has forestalled complacency and hastened an increasingly detailed analysis of persons, situations, and their dynamic interactions. Despite the resulting movement from general to more differentiated views of person-in-context systems (e.g., Shoda, LeeTiernan, & Mischel, 2002), theoretical progress relevant for understanding and integrating dynamic person-in-context systems is slowed by (a) over-reliance on general linear models (e.g., ignoring heterogeneity and non-linear relations) and (b) the lack of well-developed multilevel theories (e.g., assuming all aspects of person-in-context systems exist within or across a single unidimensional levels system). For example, conceptualizing behaviors, cognitions, and emotions as distinct “units” – existing either at different “levels of analysis” or within a single level – “slices personality in a wholly fictitious way” (Allport, 1961, p. 259) by ignoring the multilevel organization of the brain (Lewis & Todd, 2007; Peck, 2007).

Accounting simultaneously for both person-by-context interactions and the parallel distributed processes characterizing the intraindividual dynamics that produce the behavior that influences those interactions appears to require several different multilevel systems (cf. Sheldon, 2004; Peck, 2007). For example, the person-by-context interface reflects the classic (Type 1) subject-object distinction and conforms to a unidimensional series of levels of organization ranging from the sub-atomic to the astrophysical. In contrast, intraindividual dynamics reflect the phenomenological (Type 2) subject-object distinction and implicate no less than three additional unidimensional levels systems (each of which is related obliquely to the other three): levels of representation corresponding to our evolutionary history, levels of integration corresponding to our developmental history, and levels of consciousness corresponding to our volitional history (Peck, 2007; Roeser & Peck, in press).

Integrated person-in-context models will require careful attention to the nature of, and relations among, operating characteristics specific to each of these multilevel systems (cf. Magnusson, 2003). For example, consistent with Allport's (1961) conceptualization of traits as “neuropsychic structures” (p. 347) that influence (but are not defined by) behavior, MST suggests that personality traits and “cognitive-affective processing units” (Shoda et al., 2002, p. 317) will remain relatively nebulous concepts until they are defined in terms of the multilevel, networked architecture of the brain. Rather than reducing psychological concepts to biological concepts, MST suggests that biological and psychological concepts are typically alternative verbal descriptions of the same underlying parts and processes. That is, given that the biological complexity of brain has been organized across evolutionary time in a series of increasingly sophisticated regulatory systems (e.g., brain stem, limbic system, cortex) (Derryberry & Tucker, 1991; MacLean, 1990; Schneirla, 1949); and given that these regions of the neuroaxis correspond closely to psychological systems described in terms of temperament (Buss & Plomin, 1984), sensory-affective-motor (iconic) schemas (Bowlby, 1988; Case, 1991), and verbal-symbolic beliefs (Schultheiss, 2001); the differences between biological and psychological definitions of personality appear to be more ideological than substantive. MST facilitates the integration of these perspectives by highlighting their common referents.