Cognitive-Affective States Predict Military and Political Aggression and Risk Taking.

A CONTENT ANALYSIS OF CHURCHILL, HITLER, ROOSEVELT, AND STALIN

This study of world leaders uses three cognitive-affective state measures and their interactions to predict aggressive and risk-taking behavior, These measures could consistently determine how information is perceived, encoded, and used, subsequently affecting aggression and risk taking (Contrada, Leventhal, and O'Leary 1990; Mischel 1973; Mischel and Shoda 1995; Pervin 1994). World leaders were chosen to provide outcome measures of cross-disciplinary interest and because historical documents provide ample verbatim materials for content analysis. My intent is to use theories of personality, cognition, and information processing to find reliable predictors of military aggression and risk taking. My goals are threefold: (1) to demonstrate the effects of cognitive-affective mediators on nontrivial real-world behavior, (2) to illustrate how empirically validated psychology and personality measures can be used to objectively study historical figures and events, and (3) to encourage the exploration of behavioral predictors without the traditional reliance on personality traits or unidimensional models.

EXPLANATORY STYLE AND BEHAVIOR

The theory of explanatory style first emerged in the context of the reformulated learned helplessness model (Abramson, Seligman, and Teasdale 1978). Explanatory style, or an individual's habits of explaining why an event occurred, has been shown to predict behavior in the face of aversive events and to influence action and decision making, particularly in crisis situations (Satterfield and Seligman 1994; Tennen and Herzberger 1985). An individual with a pessimistic explanatory style (the causes of negative events are internal, stable, and global) tends to give up in the face of failure or extreme challenge and becomes passive, indecisive, depressed, and poor at problem solving and coping (e.g., Peterson and Seligman 1984; Peterson, Seligman, and Vaillant 1988; Satterfield 1995; Seligman 1975, 1991; Seligman et al. 1979; Sweeney, Anderson, and Bailey 1986; Tennen and Affleck 1987).

Within the realm of politics, content analysis-derived explanatory style scores have been linked to electoral success, aggressive campaigning, resilience to stress, and military aggression and risk taking (Satterfield 1995; Satterfield and Seligman 1994; Zullow et al. 1988; Zullow and Seligman 1990). According to the learned helplessness theory, a world leader with an optimistic explanatory style would be more likely to take aggressive, bold, and risky actions than would that same leader in the same situation with a more pessimistic explanatory style. The pessimistic leader would be more likely to be frozen with rumination-inspired indecision and doubt, overly concerned with policy costs and risks, less confident in the face of challenges, and less motivated to overcome the inertia of maintaining inadequate status quo policies.

Problematically, the current models of explanatory style are unidimensional--a subject falls somewhere on a single continuum between optimism and pessimism. It seems plausible that optimism or pessimism can be grounded in either rationality or illusions. Optimism may be helpful when faced with a surmountable challenge but hurtful when careful, conservative action or reflection is indicated (Baumeister 1989; Seligman 1991). By looking at the interactions between explanatory style and other cognitive state variables, one might be able to make more specific behavioral predictions.

INTEGRATIVE COMPLEXITY

Individuals often rely on heuristics or other thinking shortcuts to make decisions when faced with stress, uncertainty, or limited resources (Einhorn and Hogarth 1981; George 1986; Holsti 1972; Hockey and Hamilton 1983; Nisbett and Ross 1980; Slovic, Fischoff, and Lichtenstein 1977; Tversky and Kahneman 1973, 1974; Kahneman and Tversky 1972, 1973, 1979, 1984; Tversky 1967). Some evidence suggests that a heuristically driven or otherwise poor decision-making process is correlated with a failure to meet policy objectives and escalating international conflict (Herek, Janis, and Huth 1987; Janis 1986; Kanwisher 1989). This line of research suggests that logic and decision-making complexity should be considered the a priori normative model for decision making, even though some have conceded that such a thing as too much complexity does exist (Janis and Mann 1991). Conversely, the cognitive manager model (Suedfeld 1992) suggests that thinking need only be as complex as is called for by the demands of the situation. In some cases, heuristically driven or noncomplex thinking could be indicated and best suited to reach a particular goal. In this study, I will conceptualize integratively noncomplex thinking as a risk factor (but not a guarantee) for cognitive errors or suboptimal thinking.

Given its close ties to encodings, beliefs, goals and values, and competencies, I chose to use integrative complexity as a state measure of thought quality. The integrative complexity (IntComp) coding system was originally used to score semiprojective tests measuring individual differences in the quality of information processing (Schroder, Driver, and Streufert 1967). Put simply, integrative complexity roughly measures the breadth and depth of the search inference (i.e., decision-making) process around a circumscribed issue by rating the degree of differentiation (i.e., looking at alternative or different viewpoints) and integration (i.e., drawing conceptual connections or weighing pros and cons of alternatives). The differentiation score most likely captures an individual's search for alternatives, evaluative dimensions or values, and survey of goals. Integration measures how the collected or stated information (e.g., goals, alternatives, dimensions) is assessed and combined.

The differentiation and integration dimensions of integrative complexity do not directly capture a leader's decision-making process, nor are they able to assess the gamut of possible decision-making strategies. However, archival analyses of historical documents have shown integrative complexity to be useful in understanding military conflicts and outcomes, political roles and careers, civil unrest, political ideologies, resilience, and the impact of psychosocial stressors on cognitive processes (Raphael 1982; Satterfield 1995; Suedfeld and Bluck 1993; Suedfeld, Corteen, and McCormick 1986; Suedfeld 1985, 1993; Suedfeld and Rank 1976; Suedfeld and Tetlock 1977; Suedfeld, Tetlock, and Streufert 1992; Tetlock 1983). Other studies have also indicated that drops in integrative complexity might signal subsequent surprise attacks or other "hawkish" behaviors in military leaders (Wallace, Suedfeld, and Thachuk 1993). Tetlock's (1985) and Tetlock and Boettger's (1989) analysis of Russian and American rhetorics lead them to conclude that high integrative complexity might be related to cooperation and compromise rather than aggression.

Two primary hypotheses have been offered to explain IntComp's link to aggression or cooperation--the intrapsychic cognitive model and impression management. In the intrapsychic model, shifts in a leader's thought processes leak into his or her public pronouncements, reflecting the leader's search for alternatives and evidence, as well as his or her ability to view different perspectives and carefully integrate and weigh options. Shifts in IntComp are thought to have causal significance. With the impression management hypothesis, shifts in IntComp are merely reflections of a leader's attempt to sway public opinion toward a previously decided course of action. Disentangling these two hypothesis has proven especially difficult (Tetlock 1985; Tetlock and Manstead 1985).

RUMINATION

For the purposes of this study, rumination is broadly defined as the tendency to think nonproductively about negative affects, situations, and attributions. In terms of cognitive-affective mediating units, pessimistic rumination indicates what aspects of a situation (i.e., positive or negative) the individual strongly attends to and how much he or she processes or ruminates about them. Content analysis scoring of pessimistic rumination provides a measure of the frequency and intensity of recorded pessimistic attributions, emotions, and events, making it a potentially important adjunct to explanatory style and other cognitive variables. Kuhl's (1981, 1984) critique of the learned helplessness reformulation argued that rumination about a "negative state orientation" and not just pessimism should be an indicator for depression. In other words, the presence, frequency, and intensity of pessimistic attributions should all be considered. Empirical support for this hypothesis was found by Zullow (1984), who showed that depressive symptoms were best predicted by a pessimistic explanatory style in conjunction with high pessimistic rumination. Other studies have offered further support for the relationship between rumination and depression (Kammer 1983; Hollon and Kendall 1980; Nolen-Hoeksema, Parker, and Larson 1994). It then follows that a leader who often ruminates pessimistically could more likely exhibit symptoms of depression such as passivity, self-doubt, and negative cognitive distortions.

HYPOTHESES

I hypothesize that explanatory style, integrative complexity, pessimistic rumination, and their interactions have utility in predicting the aggressive and risky behaviors of world leaders by affecting information processing, decision making, and motivation. An optimistic explanatory style is thought to predict aggression and risk taking in military actions, whereas a pessimistic style should predict the inverse--passivity and caution. High pessimistic rumination is also believed to be associated with passivity and caution. A leader who ruminates will be less likely to employ effective coping strategies or search for alternative solutions to difficult problems. Integrative complexity is hypothesized to have an inverse relationship with aggression and risk taking and a positive relationship with caution and passivity. Decreasing levels of integrative complexity prior to a military action should predict aggression and risk taking. An integratively noncomplex leader sees no alternative solutions, ignores counterarguments against action, and might not assess the full range of possible repercussions. However, there is no assumption that passivity and caution are more desirable than aggression and risk taking.

Furthermore, I expect the interactions between the cognitive state variables to be most useful. The interaction between explanatory style and integrative complexity provides a measure of the quality (complexity) and direction (optimism vs. pessimism) of attributions. The interaction between explanatory style and pessimistic rumination highlights the direction, frequency, and intensity of attributions. Passivity and caution should be most strongly predicted by Pessimistic Explanatory Style X High Integrative Complexity and by Pessimistic Explanatory Style X High Pessimistic Rumination. This decision-making state would encourage the search for alternatives and the integration of options, but negative evaluative dimensions, negative consequences/risks, and pessimistic predictions of failures would …