Interaction between individuals and situations: using HLM procedures to estimate reciprocal relationships - A Special Issue: Focus on Hierarchical Linear Modeling

Journal of Management, Nov-Dec, 1997 by Mark A. Griffin

Individuals within a work environment not only react to the work situation as they perceive it, but also act to create that situation. For example, in a work group, an individual may choose to help others based on perceptions of group norms, the requirements of the task, and the expectation of future reciprocation. The action chosen by the individual, in turn, contributes to characteristics of the broader context. For example, choosing to help another person may strengthen group norms about helping behavior. In this way, there is an ongoing interaction between an individual's reaction to a work context and the creation and maintenance of that particular work context. In short, individuals and situations are reciprocally related.

The interaction between individuals and situations has been a critical concern in many areas of psychology. Bandura's theory of reciprocal determinism proposed a specific reciprocal relationship between individuals and situations (Bandura, 1978). More generally, the term "Interactional Psychology" represents efforts toward an integration of personality theory and situational theories as determinants of individual behavior (Pervin, 1989; Schneider, 1983). Terborg (1981) identified five conceptualizations of the term "interaction" within organizational literature. Of these five kinds of relationship, "reciprocal interaction" best describes the mutual influence of individuals and situations in the example described above.

A number of studies have emphasized the importance of reciprocal interaction between persons and situations. Drawing on the work of Schneider (1987), Chatman's (1989) study of value congruence between individuals and organizations noted that 'a truly interactive model would include the effects that people have on situations' (p.334). James, et al. (1978) described the relationship between individuals and situations as 'causally interactive' (p.797) in the construction of psychological climate. Woodman, Sawyer, and Griffin's (1993) recent theory of organizational innovation noted that 'individual factors both are influenced by and influence social and contextual factors' (p.301).

The above examples indicate that reciprocal interaction has been incorporated into a variety of organizational behavior theories. However, there is a frequent tone of pessimism that such an interaction can be captured by anything except the most complex research enterprises. Kozlowski and Doherty's (1989) integration of climate and leadership concepts included the comment that it was 'too early to specify and test reciprocal effects' (p.552). George and Brief's (1992) review of individual organizational spontaneity and group affect disclaimed 'available data are too sparse...but, on conceptual, grounds it probably could be argued that the relationship is reciprocal' (p.321). More broadly, the estimation of reciprocal effects has been methodologically problematic in organizational research (Williams & Podsakoff, 1989).

The current paper explores an example of interaction between two levels of analysis. The relationship between individual and their group context is explored over a number of time periods. The paper presents a simplified model of reciprocal interaction and proposes an analytic procedure for measuring the interaction. A group laboratory task study is presented and discussed in relation to the proposed procedure and the simulation study.

Context for Current Study

This paper uses Hierarchical Linear Models (HLMs, Bryk & Raudenbush, 1989, 1992) to investigate reciprocal interaction between individuals and situations. The estimation of upward influence - the impact of individuals on situations - -is a central problem for specifying reciprocal relationships. This paper proposes an extension of cross-level analyses that links levels of analysis over time periods to estimate reciprocal interaction between individuals and their group context. Figure 1 displays the theoretical model that underlies the HLM application used in the current study. The model depicts individual and group level measures at two time periods.

The model in Figure 1 indicates that group level cohesion influences the experience of individual affect within the group. Individual affect, in turn, is proposed to increase the production of behaviors that engage other group members such as spontaneous helping and task orientation (George & Brief, 1992). These behaviors contribute to the subsequent group context by increasing the experience of cohesion among group members. Therefore, over time, individual processes influence group outcomes. The above model provides a basic framework for linking individual and group processes over time in a reciprocal interaction. When variation occurs at both the individual and the group level of analysis, researchers have disagreed about the most appropriate means to partition this variance (e.g., George & James, 1993; Yammarino & Markham, 1992). The current study emphasizes the need to consider both levels of analysis. Particular emphasis is placed on estimating the path depicted by the bold arrow which represents the influence of individuals on group outcomes.