The Balance of Power:

The Alley Game:

An Evaluation of Power Differentials and Personal Attributes

on the Decision to Confront, Reward or Retreat from Aggression

Richard B. Tanksley, University of Idaho

Dr. Jack E. Vincent, Borah Professor of Political Science, University of Idaho

ABSTRACT

International relations theory has tried for a long time to determine probable responses to aggression in situations involving "balanced" vs. "imbalanced" power capabilities between competing actors. In spite of scholars giving "power differential" a great deal of attention, both in theory and research, there have, in fact, been disagreements as to the most likely response of competing actors in this regard. Our research approaches this critical issue by using a "survey" to determine the most likely response of respondents to a hypothetical conflict situation called the "Alley Game." This is done, in part, by evaluating various power combinations of the actors involved in the game. Respondents are asked what they would do, if, in walking down an alley, they are confronted by an individual who demands money before they can pass. He also pushes the respondent in the chest to signal his intention to use force if necessary. Various power combinations (of those in conflict in the alley) are evaluated as to the respondent's choice to "confront, reward or retreat." These power combinations include being armed with a knife (or a gun) as well as variations in the physical size of the aggressor. Our primary research objective is to test Organski's view of the likely effects of balance of power. That is, we wish to determine whether there are more "confrontational responses" when power is "balanced" than when power is "imbalanced." In addition, personal attributes are also analyzed. These included age, gender, military service and political orientation as well as other selected personal or emotional characteristics. The survey responses were coded into variables and using standard statistical tests the data was analyzed. The statistical tests show significant differences for many of the variables analyzed. They indicate that both power differential (of the type that supports Organski's theory) and selected personal attributes appear to be important factors in assessing the likely response of the respondents in the hypothetical Alley Game confrontation.

INTRODUCTION:

The goal of this research is to help identify variables that may be associated with decision-making during confrontation. Both individual characteristics and power differentials are examined through survey research depicting hypothetical situations. The survey itself is called the alley game. This game or survey features scenarios where the respondent has an important meeting to attend and to arrive on time they will have to pass through an alley. [1] In the alley, a man pushes the victim (survey respondent) and demands twenty dollars before he will allow them to pass. The initial choices for the victim are: to push back, pay, or retreat. The scenario is repeated with varying degrees of power differentials as to the man’s size, or what type of weapon the victim possesses. For those that initially choose to push back and possess a knife or gun, a second scenario is offered. In this latter scenario, the aggressor pulls a large knife, and the victim is given a second set of choices: to either pay, retreat or pull out their own weapon. By comparing the means of the key variables, the size of the man, and what weapons are present, various power differentials are analyzed. Additional respondent variables are also scrutinized to reveal what characteristics might influence the respondent’s choice to push back, pay, or retreat. Thus, the alley game becomes a framework to investigate the balance of power and other auxiliary theories. Questions addressed include: what is the relationship of different balance of power conditions to possible confrontational behavior? If power is balanced, what is the more probable outcome, conflict, acquiescence or retreat? What role does gender, or personal attitudes play in respondent’s choices to push back, pay or retreat? Obviously, these are important questions that influence decision making in regards to how a state or individual might respond to changes in the power matrix, issues of deterrence and the prospects for peace. This question is also important since theory and research by others has led to dissimilar conclusions. Organski asserts that the international system is less likely to achieve peace when there is a balance of power (Organski, 1968, p. 27). Morgenthau & Thompson, reach the opposite conclusion stating that “ the balance of power and politics aiming at its preservation are not only inevitable but are an essential stabilizing factor in a society of sovereign nations” (1985, p. 167). Other determinants of decision-making about conflict will also be discussed. With this background information, predictions to a survey will be presented along with its data, methodology, limitations and conclusions.

Concepts and Literature Review

Before going any further into this inquiry, it is worthwhile to discuss several of the concepts that surround this issue. The term conflict is subject to a variety of definitions. To most it may refer to groups or individuals competing over scarce resources or disparate goals. Lewis A. Coser defines conflict as “a struggle over values and claims to scarce status, power and resources in which the aims of the opponents are to neutralize, injure, or eliminate their rival” (Dougherty, & Pfaltzgraff, 1999, p. 187). Conflict involves the interaction of people, with at least one party initiating a demand or forcing the other party to make less desirable choices than it would make when left alone.

The definitions of power and balance of power are more elusive. While power is a central concept within international relations, its usage has been subject to “definitional confusion” (Rothgeb, 1993, p. 19). Often power is utilized to portray a state’s resources or the capabilities of a state. This type of power is somewhat easy to measure as it consists of economic, military and manpower that is available to a state. Still, even with the measurement of these resources, it is difficult to make an accurate and realistic assignment of the relative importance of each asset. How the resources convert into the ability to have one’s way is not so easy to evaluate (Rothgeb, p. 44). Morgenthau & Thompson also make this point about how difficult it is to assess the power calculations of a nation [2] (1985, pp. 223-235). Rothgeb’s second conceptualization identifies power as a goal. It is not enough to just want power without really defining what the actor desires. Lastly, power can be seen as a means to control others or control the environment in which one is found (Rothgeb, 1993, p. 23). This definition seems the most consistent with other scholars. Morgenthau & Thompson speak of power as man’s control over the minds and actions of other men (1985, p. 32). In other words, can the party obtain what they desire? If so, that is power. This is the power definition that is used for this study. One person, the aggressor, seeks to control the victim by demanding money. The victim, on the other hand, must react with power, acquiesce or retreat.

The notion of a balance of power or equilibrium has been around for thousands of years (Morgenthau & Thompson, 1985, p. 4). As a term though, the balance of power has been used indiscriminately to convey a multitude of situations. Sheehan [3] examines several lists of meanings concerning the balance of power. Examples of such alternatives include the existing distribution of power. Conveying the balance in this way could essentially represent all variants of the distribution of power (Sheehan, 1996, p. 16). This distribution can be at or near equilibrium or there can be an imbalance that would suggest that one state or alliance has a superior advantage. In order to avoid confusion, the balance of power used in this study is determined to be an equal distribution of power. This is consistent with Morgenthau’s “metaphor of two scales kept in balance by an equal distribution of weights on either side, providing the mechanism for the maintenance of stability and order on the international scene” (p. 223). This type of definition is echoed by Organski as the balance of power rests on the premise that “the various nations group themselves together in such a way that no single nation or group of nations is strong enough to overwhelm the others, for its power is balanced by that of some opposing group” (Organski, 1968, p. 274). Often such a balance is seen as structural or a “system” where a shift away from equilibrium leads to countershifts. Such balancing then occurs much like the equilibrium is found in free markets where the quantity supplied will shift in response the quantity demanded (Spanner & Wendzel, 1996, p. 109)

The survey depicts only a simple set of circumstances and limits the victim’s choices should they seek retribution. Still, the victim can push back and brandish a weapon when the option is available. It is noted that such a response is fully justified whether an individual or state is involved. “Aggression justifies two kinds of response: a war of self defense by the victim and a war of law enforcement by the victim and any member of international society” (Walzer, 1988, p. 161). Furthermore, “any use of force or imminent threat of force by one state against the sovereignty or territorial integrity of another constitutes aggression” (Walzer, 1988, p. 161). Walzer also suggests that aggression or war in the international context also has a domestic analogy such as murder or robbery. Thus, the word aggressor seems appropriate in this domestic context, with the victim being justified to respond in self-defense. While it can be said that the victim who confronts the aggressor is seeking justice, the word retribution is used more specifically to indicate that the victim is seeking some type of punishment. “The retributive urge also serves a purpose in present day society, which is to give people a motive to ward off transgression” (Shavell, 2000). It is exactly this urge or response to conflict that is scrutinized in this domestic hypothetical situation. Of course, there may also be implications at the state level as well.

Just as economic equilibrium is obtained differently under different market structures, the structure of how any conflict is framed can lead to different conclusions. For instance, later work by Organski and Kugler looked at not only the balance of power model, but also models of collective security and power transition (1980, pp. 13-19). In addition to other models or structures, specific determinants such as nuclear capability, superpower status, geography, and size, are said to contribute to how conflict is started and resolved (Organski & Kugler, 1980). Offensive and defensive capabilities are also discussed along with a multitude of other variables that have been aggregated by research programmes such as the University of Michigan’s correlates of war project (Mouritzen, 1988, p. 229). While this research recognizes the importance of such variants, the schema for this study is to specifically address qualities that may predispose one to seek retribution or not. Namely, are there differences between a balanced and unbalanced scenario? In addition to this simplification, this research does not address the controversial question of how many powers are considered optimal for any such balance. Instead, this research attempts to provide clues as to the characteristics of a situation involving only two actors.

A survey of theoretical literature shows that the expected outcome from a situation involving a balance of power is far from consensus. While Organski provides an insight into the definition of the balance of power, he dissents concerning it’s usefulness, going so far as to label the theory as having “erroneous conclusions” (1968, p. 288). To Organski and others, conflict is more likely when the powers are closer to being equal. “In both domestic and international affairs, as long as one group or one state has the preponderance of power there may be peace, but it will be an imposed peace, with the terms defined by holders of power” (Singer, 1972 p. 85).

Inversely, Morgenthau and others hold that the balance of power does help to subdue conflict and prevents domination. Morgenthau and Thompson (1985) use several examples of how the balance of power works, and deviations from equilibrium result in conflict. Historical examples that are cited, among others, include: Belgium, (p. 196) Korea, (p. 198) and Austria (p. 229). It must be noted that these examples involve competition more than a direct confrontation.

Several studies have tried to empirically demonstrate whether the balance of power favored peace or aggression. Tindell and Vincent (1970) specifically tried to assess balanced versus unbalanced power conditions on human behavior. Their experimental study simulated direct confrontation, as subjects were pitted one against the other. Various conditions were then applied where each subject was allocated a number of shocks that they could apply to their opponent. The results indicated that the game conditions did effect how the shocks were distributed. When each side had an equal number of shocks, fewer shocks were utilized and more cooperation resulted. Inversely, when one person had more shocks, the number of shocks that were distributed increased. The number of shocks were set at the beginning of this experiment. Thus, this experiment could be considered static. Vincent and Schwerin (1971) in their “Ratios of Force and Escalation in a Game Situation” created a game with a dynamic process which allowed the purchasing of shocks. Interestingly, in this game, those in a position where there was inequality were found to be more cooperative and submissive versus those that started the game under a balanced condition. Thus, the static game conditions of Tindell and Vincent produced more of a Morgenthau interpretation, while the Vincent and Schwerin dynamic game conditions produced more of an Organski interpretation.

It is important to note that the Tindell/Vincent and Vincent/Schwerin studies allowed for more repeated confrontations. Shocks were available to be distributed over a number of times. This was not the case for this survey. Respondents were not led to believe that another confrontation could occur in the future by the same aggressor. It is also important to note that the confrontation depicted in this survey is limited to two powers. While research on several powers would be more desirable, the entire balance of power issue has some application within a two power context. As to the pattern of conflict, Morgenthau sees the struggle for power (conflict) as originating from two patterns: competition and direct opposition (1985, p. 192). This survey assumes a pattern of competition that is in direct opposition with no third party involvement.

Two extremes are offered concerning the level of analysis. Organski notes: “that it could be argued of course, that nations should not be our main concern, that it really is individual human beings with whom we are dealing. It would be an even greater mistake, however, to treat national leaders as individuals representing no one but themselves” (1968, pp. 15-16). While the balance falls somewhere in between, there is a strong case that much can be learned through individuals. In a roundtable discussion, Fink asserts: “when one is studying crises, one of the most profitable ways to go about it, at least as far as I’ve seen, is to deal with the decision-maker as an individual in that time-stress situation” (Givens & Nettleship, 1973, pp. 70-71). Even simple game models such as “chicken” or the “prisoner’s dilemma” look specifically at the behavior of individuals. With this in mind, if certain attributes in individuals do lean toward aggression or retreat, they are worthy of study.

Dougherty and Pfaltzgraff suggest that social science is divided into two groups (1999, p. 189). The micro group looks specifically at human beings and the origins of conflict. Meanwhile, the macro group looks at institutions. Neither is wrong, and more recent attempts at merging the two approaches should be applauded. Even works such as Kenneth Waltz’s, Man, the State and War, and other theories regarding international relations make assumptions about the nature of man (Dougherty & Pfaltzgraff, 1999, p. 273). Thus, research such as this study remains important and may be applicable, if not critical, in the study of international relations.

The survey also provided a means to examine various personal traits and their relation to conflict. While all of the traits were anticipated to suggest more or less conflict, one in particular warrants further discussion. Within the literature of international relations is a debate about whether there is a gender difference between men and women. Specifically, will women will be more passive and less prone to violence, with all things being equal, as opposed to their male counterparts (Caprioli, 2000, p. 51). Generally it is suggested that as women are given more equality in making foreign policy decisions, the results will be more pacifist. Caprioli particularly looks at whether gender equality corresponds to less militarism and concludes that this is the case. Other statistics have shown that men engage in more risky behavior (Mitchell, 2001, p. 8). Dr. Mitchell (M.D.) writes this off as more testosterone which: leads men to “react more aggressively when provoked.”

Coding Rules

In an attempt to predict what respondents would do in an alley confrontation, a hypothetical questionnaire was written by Vincent et al. A total sample of 83 was taken among students, working adults and retirees from both Idaho and Tennessee. This survey provided the respondent with nine hypothetical situations that are labeled as conditions in Appendix A. All the situations started with size being a key independent variable. The sizes of the aggressor included:

1. Small, thin

2. Just your size

3. Much larger than you and very muscular

You are walking through an alley to avoid a long walk in a large city, since you will be late to an important meeting if you take the long way around. Suddenly a small, thin man jumps in your way and demands $20 to pass. Before you can speak he pushes you backward with considerable force. Between the three options listed, which, would you chose?

The dependent variable was the victim’s response. Initially there were three choices:

1.	Push him out of the way and proceed forward.	(coded 1)
2.	Pay him the $20 so you can pass.	(coded 2)
3.	Retreat and take the long way around.	(coded 3)

A second method of coding was also utilized. (The only difference is that pay or retreat are coded equally).

1.	Push him out of the way and proceed forward.	(coded 1)
2.	Pay him the $20 so you can pass.	(coded 2)
3.	Retreat and take the long way around.	(coded 2)

In addition to the aggressor’s size, another independent variable introduced was the type of weapon the victim had. These included:

1. none

2. a large knife

3. a large gun

When the victim had a knife or gun, (conditions one through six) those that pushed back were confronted with these choices:

After you push him, he pulls a large knife from his belt.

Between the three options listed, which, would you chose? First method Second method

_____Pull a large gun from your belt.	(coded 1)	(coded 1)
_____Pay him the $20 so you can pass.	(coded 2)	(coded 2)
_____Retreat and take the long way around.	(coded 3)	(coded 2)

Again, both coding methods were proposed in order to see if there was a difference in the pay or retreat options. Initially it was thought that retreating was the least aggressive option. In reality, some respondents likely thought that paying was the least aggressive option. Due to this possible inconsistency, it was very difficult to separate out the pay and retreat options. Therefore, most of the survey is reported with retreat and pay coded equally. This not only adds parsimony but the least aggressive option (pay or retreat) cannot be confused with the most aggressive option (push back). This method of coding also creates a dichotomous variable with the victim either seeking retribution (pushing back) or not seeking retribution (paying or retreating).

Altogether, this survey was intended to answer some of the theoretical questions concerning the balance of power by using independent variables that approximated different levels of potential power. In addition to variables that might measure the balance of power, other characteristics were also included to test whether personality traits and environmental influences might affect the responses. More ancillary theory is identified in the assumptions for each of these variables.

ASSUMPTIONS:

The assumptions used included the following:

1. Using the Organski interpretation of the balance of power, the situation where the victim is most likely to push back (no weapon) would be when the aggressor is of the same or lessor size.

2. In the case where the victim is smaller, (no weapon) Organski would predict that the chances of peace would be greater, thus the victim in this case would either pay or retreat

3. When a knife is present, the Organski interpretation would predict that the victim would first push back since the weapon the victim possesses would seemingly give them the edge. The victim will most likely pull his knife on the aggressor (who has pulled a knife out already) when the victim is the same size or larger than the aggressor. When the victim is of lesser size and both have knives the victim is expected to retreat or pay.

4. When the victim has a large gun, an Organski interpretation would predict that the victim would push back. When the aggressor produces a knife, it is then expected that the victim would produce a gun rather than pay or retreat.

5. It is expected that individuals who support the United Nations will also be more likely to react like the U.N. So what kind of reaction might we expect this to be? To the well informed, the U.N has a number of different organs, functions, and bodies. However, as Falk admits: “it is the peace and security agenda that serves as the prism through which the U.N. is judged by the media and the public” (Falk, 1995). It is predicted that this peaceful agenda of the U.N. is emphasized by its supporters. Article 2.3, of the U.N. Charter notes: “All Members shall settle their international disputes by peaceful means in such a manner that international peace and security, and justice, are not endangered.” (Simons, 1994, p. 269). While the U.N. stands for justice, protection from any conflict is unlikely to be immediate, is always multi-lateral and occurs only after all other measures have failed. Therefore, it is assumed that supporters of the U.N. will be more likely to pay or retreat.

6. Liberals, or those that believe in affirmative action, should also be more likely to pay or retreat. Conservatives are expected to seek retribution and they dislike affirmative action. While these two measures are coded separately, it should be noted that some multicollinearity exists between the affirmative action and liberalism. Jackman and Muha for instance, suggest that “support for affirmative action policy is often attributed to the more ‘liberal’ members of society” (Niemann & Dovidio, 1998, p. 783). This relationship is shown in Appendix B. This graph demonstrates that the strong conservatives are least likely to support affirmative action. The survey did not provide formal definitions of what constitutes a liberal or conservative. Ideally, tests could have determined this with each respondent. Researchers have identified various attitudes for what constitutes a liberal or conservative (Knight, 1999, p. 61). Items that identify conservatives include: preferring stricter law and order, greater punishment for offenders, less equality, and even racial segregation (Knight, 1999, pp. 86-142). Others see liberals as “more inclined to give than to save” and more “generous-spirited” (McWilliams, 1995, p. 177 & 203). Liberals are more likely to see the aggressor as a victim. In discussing disorderly street conduct: “commentators with a hyper-equalitarian outlook single-mindedly aim at redistributing wealth, status, and opportunity to the poorest at hand-in this context, to a street person whose misbehavior has annoyed other pedestrians” (Ellickson, 1996).

7. Those that have a strong sense of standing up for what is right, should view the aggressor as performing an unjust act and respond more aggressively. This assumption is based on Deutsche’s comments: “If each side in a conflict tends to perceive its own motives and behavior as the more benevolent and legitimate, it is evident that the conflict will intensify” (Duetsch, 1988, p. 137).

8. Those who avoid conflict or who do not perceive that violence is an acceptable solution to solving conflict should consistently pay or retreat. Pacifists would fall into this category. Fitting in this group are those that have a “non-resistance to violence by principle” [4] (Aron, 1973, p. 367).

9. Respondents that believe people are trustworthy should be more inclined to pay. A similar question on trust is surveyed in the National Opinion Research Center's General Social Survey (Gleaser, et al., 2000). Researchers have found results are “difficult to interpret” due to varying beliefs in what trust is or what constitutes “most people.” Glaeser, applied trust to a game situation where a payout was involved and found that those who self reported that they trusted strangers also gave higher payouts than those who were less likely to trust strangers (The Economist, 2001 p. 8). In the Alley Game, the aggressor pushes the victim and demands twenty dollars. Those less likely to trust may have difficulty in taking the aggressor at his word.

10. Respondents that reported themselves as more submissive should have a higher propensity to pay, or retreat. As mentioned in the introduction, Morgenthau wrote of power as: “man’s control over the minds and actions of other men” (1985, p. 32). Sympathy is likely to dilute man’s ability to control others.

11. Those that are more disposed to commit a felony should also be more aggressive and push back.

12. It is expected that older respondents will be more likely to pay or retreat. Interaction with other independent variables could provide even further explanations as an elderly person may see themselves as potential targets and avoid confrontation for a lack of trust.

13. Based on gender research in international relations it is anticipated that women are less likely to push back. It is noted however though that the survey is gender biased since the hypothetical aggressor is always male. While at first glance this seems to be a flaw, in the international arena most women leaders will be confronted by a predominately male leadership circle on their staff, in opposing governments, and within their own military. However, it is undetermined whether females would respond differently than males when the aggressor was also a female.

PREDICTIONS

Treatment Conditions

Independent Variable	Dependent Variable	Direction As	Dependent Variable	Direction as
	(initial response)	Predicted	(After pushing back the	Predicted
			aggressor pulls a knife)

Victim is smaller (no weapon)	pay or retreat	Yes	not an option	n/a
Victim is the same size (no weapon)	push back	Yes	not an option	n/a
Victim is larger ( no weapon)	push back	Yes	not an option	n/a

Victim is smaller (with knife)	push back	Yes	pay or retreat	No
Victim is the same size (with knife)	push back	Yes	pull knife	No
Victim is larger (with knife)		Yes	pull knife	No

Victim is smaller (with gun)	push back	Yes	pull gun	Mixed
Victim is the same size (with gun)	push back	Yes	pull gun	Mixed
Victim is larger (with gun)	push back	Yes	pull gun	No

Knife present versus no weapon [5]	push back	Mixed	pull knife	Mixed
Gun present versus no weapon	push back	Yes	pull gun	Yes
Gun present versus knife	push back	Yes	pull gun	Yes

Respondent Characteristics & Results

Independent Variable	Dependent Variable	Dependent Variable	Direction as
			Predicted
	(Seek Retribution)	(Pay or Retreat)

Age	younger	older	Mixed
Gender	male	female	Yes**
military background	yes	none	Yes**
I am liberal/conservative	conservative	liberal	Yes*
Do you agree with United Nations policies?	disagree	agree	Yes**
I stand up for what I think is right	agree	less agreeable	Mixed
I think violence is an acceptable solution	highest	lowest	Yes**
I will take great measures to avoid being in a fight	lowest	highest	Yes**
Most people are trustworthy	lowest	highest	Yes*
My level of athletic ability is	highest	lowest	Yes**
I think affirmative action is a good policy	disagree	agree	Yes**
How do you rate your level of sympathy to others	lower level	higher level	Yes**
Personality in dealing with others?	more aggressive	more submissive	Yes**
Felony without penalty	more likely	less likely	Yes**

** significant with one tailed test at the .01 level.

* significant with one tailed test at the .05 level.

Findings

With pushing back, coded as a 1, and retreat or pay coded as a 2, the following means were obtained for the treatment conditions. (These means are only for the initial decision with the aggressor) A lower mean indicates more respondents chose to push back.

	No weapon	Victim has knife	Victim has large gun
Smaller aggressor	1.47	1.51	1.40
Same size aggressor	1.57	1.59	1.43
Larger aggressor	1.80	1.78	1.56

As reported, when the aggressor increased in size, the number of victims that pushed back declined. However, there was little difference between those that were willing to push back with no weapon versus those that would push back when they possessed a knife. This simply indicates that fewer people were willing to change their responses when they possessed a knife. Conversions of respondent answers occurred more often when the large gun was introduced to the victim. Most respondents were in fact willing to initially push back when they possessed a large gun. Thus, the relative means are consistent with our predictions and are basically congruent with an Organski interpretation. The only caveat might be that the weapon had to clearly be perceived as providing an advantage. For those that originally chose to push back the following means were obtained from their second option, whether to pull a large knife or large gun. (A lower mean indicates that more victims made the choice to brandish their weapon).

	No weapon	Victim has knife	Victim has large gun
Smaller aggressor	n/a	1.49	1.14
Same size aggressor	n/a	1.47	1.15
Larger aggressor	n/a	1.45	1.12 [6]

For these options the results are mixed. The means of the variables representing the size of the aggressor are nearly equal. However, these options do indicate that the balance or unbalanced conditions of power were more clearly represented with what type of weapon was present. Those victims possessing a large gun would seek retribution from a person wielding only a knife.

All treatment conditions combined for a non-parametric test

(Kendall=s Tau b)

With pay or retreat coded as equals, the direction predicted versus the direction found (between various means) was observed to be:

Tau B Correlation coefficient .481

Significance .000**

Classification Table

Predicted

1 2 Percent Correct

Observed

1 20 7 74.07%

2 7 20 74.07%

Overall 74.07%

Interpretation: 27 means were projected to be lower, 20 were found to be lower. Inversely, 27 means were projected to be higher and 20 were found to be higher.

With pay coded as a 2 and retreat coded as a 3, the direction predicted versus the direction found (between various means) was observed to be:

Tau B Correlation coefficient .556

Significance .000*

Classification Table

Predicted

1 2 Percent Correct

Observed

1 21 6 77.77%

2 6 21 77.77%

Overall 77.77%

Interpretation: 27 means were projected to be lower, 21 were found to be lower. Inversely, 27 means were projected to be higher and 21 were found to be higher.

Results in both coding situations were obtained by comparing the direction of the means between the various sizes of the aggressor, and between the same size aggressor and variations in weapons. Regardless of how pay and retreat are coded, there was predictive success in whether a victim would seek retribution. Such predictions would lend credence to an interpretation that when victims possess similar power or greater power, they will also be more likely to seek retribution. The Tau b coefficients throughout this study can be interpreted as suggesting that there is an n % more likelihood that the mean tested would be in the same direction as the one predicted (Siegal & Castellan, 1990). The predictive rate of success for the direction of the means is significant for the treatment conditions. However, the difference in the magnitude that any individual or group will seek retribution can not be determined.

Kendall’s Tau b tests for each respondent variable

Next, the personal questions were analyzed individually to answer what personal characteristics might help determine whether an

aggressor would be confronted or not. The dependent variables consisted of either push, (1) or pay/retreat (2). The independent variables listed below were coded with continuous numbers for age, dichotomous variables for military background and gender, and scaled numbers for all other questions. Means were analyzed using Kendalls’s Tau b tests to examine whether the direction predicted was also the direction found. Several of the variables were highly significant as to the direction of their means.

Age:

Correlation Coefficient .200 Significance .141

Overall this variable was not found to be significant. However, if the conditions where the man produced a knife are excluded, the expectation that a younger person would push back more did occur in all of the original nine conditions. For these conditions, the average age of the group pushing back was approximately three to five years younger.

Gender: Male =1, female = 2

Correlation Coefficient 1.000 Significance .000**

As predicted, the male gender was more likely to push back. This occurred under all conditions including the six options where weapons were available. For instance, with no weapon and the smaller aggressor, (condition 1) 64% of men would push back, versus only 20% of women. Means for this same condition were 1.36 for those that pushed back, 1.77 for those that paid, and 1.70 for those that retreated.

Military: yes = 1, no = 2

Correlation Coefficient .733 Significance .000**

A pattern emerges here with those of military status seeking retribution at a greater rate than those without under the initial confrontation. However, when weapons are presented, results are mixed even when the categories of pay and retreat are combined. This seems to suggest that when confronted by weapons, some military personnel may seek to pay or retreat when given the option. Perhaps the explanation of this seeming anomaly is that this conflict is not worth the trouble to bring about involvement?

I am:

Liberal 1__________2__________3__________4__________5__________ Conservative

Correlation Coefficient .333 Significance .036*

This variable asking respondents to scale whether they were conservative or liberal was marginally predictive. What seems to be occurring with this variable is that professed liberals pushed back the most under the conditions where they clearly had an advantage or at times when they were evenly matched. Inversely, the more conservative respondents chose to push back against the larger man. Once a weapon is produced though, the more liberal respondents were less responsive to pulling out a knife or gun regardless of the size of the aggressor. Meanwhile the conservative group was more likely to seek retribution when they possessed any weapon, the size of the man was much less important.

Do you agree with the United Nations polices ?

Low 1__________2__________3__________4__________5_________ High

Correlation Coefficient .467 Significance .006**

Those that self-reported as being more supportive of the United Nations were the least likely to push back. Inversely, those that chose to use weapons were less agreeable with U.N. policies. Thus, U.N. support is indicative of a more pacifistic group.

I stand up for what I think is right:

Agree 1__________2__________3__________4__________5__________ Disagree

Correlation Coefficient .200 Significance .141

It was predicted that the group that would stand up for what is right would also be the most willing to push back or pull a weapon. These results were mixed and not significant.

I think violence is an acceptable solution to solving human conflict and differences:

Agree 1___________2_________3__________4__________5__________ Disagree

Correlation Coefficient .600 Significance .001**

As predicted, the group more likely to push back also scored the lowest on this scale.

I will take great measures to avoid being in a fight:

Agree 1___________2_________3__________4__________5__________ Disagree

Correlation Coefficient .867 Significance .000**

The group with the lower means for this question was also more likely to pay or retreat. This was exactly as projected.

Most people are trustworthy:

Agree 1___________2_________3__________4__________5__________ Disagree

Correlation Coefficient .333 Significance .036*

It was predicted that those who agreed with this statement would also be the respondents who would be more likely to pay. With pay and retreat coded separately this was not always the case. With retreat and pay coded together, the results were significant: the groups more trusting did tend to pay or retreat more often.

My level of athletic ability is:

Low 1___________2_________3__________4__________5__________ High

Correlation Coefficient .867 Significance .000**

Those who are more athletically inclined also were more willing to seek retribution. The only exception occurred in an optional treatment condition where the larger aggressor pulled a knife and the smaller aggressor still decided to pull their knife as well.

I think affirmative action is a good policy:

Agree 1___________2_________3__________4__________5__________ Disagree

Correlation Coefficient 1.000 Significance .000**

This variable under all treatment conditions indicated that those who agreed with affirmative action also were the least likely to push back. This is interesting given that the self reported liberal/conservative scale was not as predictive.

How do you rate your level of sympathy to others:

Low 1___________2_________3__________4__________5__________ High

Correlation Coefficient 1.000 Significance .000**

This variable’s means were just as predicted under all treatment conditions. The group most likely to report themselves as having greater sympathy to others was also the least likely to push back. Average means typically ranged from 3.2 for the group that choose to push to 4.4 for those that would prefer to pay or retreat.

How do you view your personality in dealing with others?

Submissive 1__________2_________3__________4__________5__________ Aggressive

Correlation Coefficient 1.000 Significance .000**

As predicted, the group whose means suggested they were the most submissive also preferred to pay or retreat.

How likely are you to commit a felony if you know you cannot be penalized:

Unlikely 1__________2_________3__________4__________5__________ Very likely

Correlation Coefficient 1.000 Significance .000**

All means compared for this question ranged from 1.5 to around 2. Even with this close proximity, those more apt to commit a felony without penalty were also more likely to push under all scenarios.

All respondent variables combined (Kendall=s Tau b test)

Correlation coefficient .686 Significance .000**

Next, all respondent variables were combined for one larger Tau B test. This was done by combining all the respondent variables into one Tau B test. This provides an overall indication of how the respondent variables in their entirety can predict the direction of the means. The test confirms that there is a 68.6% chance that the direction predicted would also be the direction found. The direction is highly significant, though the exact magnitude of the direction can not be determined.

Classification table

Predicted

1 2 Percent Correct

Observed

1 177 33 84.29%

2 33 177 84.29%

Overall 84.29%

Conclusions

To this point, the survey and means analysis accomplished the goals that were set out for this research. Using an Organski approach when the balance of power was dominant or equal, more pushing back occurred. Even as an investigational research design, the Organski interpretation best explained the differences in the behavior of individuals when confronted by an aggressor. When the victim was weaker, as measured by size or weaponry, the more likely response was to retreat. These first option conditions, fit the data perfectly on a relative basis. This suggests that randomness is unlikely. Altogether, the means for pushing back, pay, or retreat do not suggest that conflict will occur in a balanced situation, but certainly it does seem to indicate that conflict is more probable.

The second options of pulling a large knife or gun when confronted with a knife have more opaque results. Perhaps the goal of getting to a meeting on time was not enough to warrant the brandishing of a large knife or large gun. Future testing could enforce a greater loss to see if more dire circumstances would change responses. This inconclusiveness of these six conditions represents the need for more specific testing and should not invalidate the results of the other nine conditions.

Many of the attributes of the other independent variables were found to be important in this study. The direction of the means was exactly as predicted for variables concerning gender, affirmative action, likelihood of committing a felony, personality in dealing with others, and the level of sympathy toward others. Other variables such as military service, liberal/conservative scale, United Nation’s policy, violence is an acceptable solution to conflict, taking measures to avoiding a fight, level of trust and athletic ability were also significant and could explain some of the variation in whether the respondents would seek retribution or avoid conflict. Altogether, twelve respondent questions were significant while only two were not. The Kendall’s Tau b tests for both treatment conditions and respondent variables combined were also found to be significant.

Altogether, conditions of size and weaponry and personal characteristics do help determine one=s predisposition to seek, reward or ignore injustice. The outcome in this study seems to counter Morgenthau’s claim that the balance of power thwarts conflict. Instead, the Organski approach seems vindicated. More retribution may be sought, resulting in conflict, when there is a balance of power. It is wise to be cautious about taking these individual simulation results and applying them wholeheartedly to the international nation-state system. Nonetheless, this survey does add to the body of knowledge concerning the power relations between states.

Lastly, this survey and analysis has various implications for future research. Difficulty remains in trying to measure anything more than a balanced versus unbalanced condition. Some of the respondent variables are likely measuring similar traits and could be consolidated. Alliances and conflict possibilities where more than two powers are present also warrant further study. Future research designs could incorporate techniques such as logit or probit to reveal the significance of the independent variables while at the same time controlling for the others. More attention should also be given to hypothetical situations such as this research incorporated. While these are complete simulations, the coding is objective and not subject to any coders specific interpretation. Simulations such as this survey should not be considered replacements for real situations but can offer supplemental evidence to support real case studies or experiments. Lastly, future surveys could expand to determine if such attributes to seek retribution or not are the same for all people groups and states. Research could also include more variables relating to educational background, income, race, religion and other cultural attributes. Only by including these types of attributes can such research truly become more generalized toward all people.

REFERENCES:

A Matter of Trust; Economic Focus. (2001, February 17). The Economist, (US). Retrieved December 5, 2001, from Expanded Academic Index.

Aron, R. (1973). Peace and War, A Theory of International Relations. R. Howard & A. Baker-Fox (Trans.) Garden City, N.Y.: Anchor Press.

Caprioli, M. (2000). Gendered Conflict, Journal of Peace Research, 37, 51-68.

Dougherty, J.E. & Pfaltzgraff, R.L. (1999). Contending Theories of International Relations. New York: Harper & Row.

Deutsch, M. (1983). The Malignant (Spiral) Process of Hostile Interaction. Political Psychology 4, (no. 1), n.p.

Ellickson, R.C. (1996).  Controlling chronic misconduct in city spaces:  Of panhandlers, skid rows, and public-space zoning.  The Yale Law

                Journal. 105, (no.5). 1165-1248.

Falk, R.A. (1995). Appraising the U.N. at 50: the looming challenge, Journal of International Affairs, 48, 625-646.

Glauser, E. L., Laibson, D., Scheinkman, J. & Soutter, C. (2000, August). Measuring Trust. Quarterly Journal of Economics, 115, Retrieved

November 16, 2001 from Expanded Academic Index.

Givens, R.D. & Nettleship, M.A. (Eds.). (1973). Individual Aggression as a Cause of War, (a roundtable discussion ). In Discussions of War

and Human Aggression (pp. 69-86). The Hague: Mouton Publishers.

Knight, K. (1999). Liberalism and Conservatism. In J. Robinson & L. Wrightsman (Eds.), Measures of Political Attitudes (pp. 59-150).

San Diego: Academic Press.

McWilliams, W.C. (1995). Ambiguities and Ironies: Conservatism and Liberalism in the American Political Tradition. In W.L. Taitte, (Ed.),

Moral Values in Liberalism and Conservatism (pp. 175-212). Dallas: University of Texas at Dallas.

Mitchell, T. (2001). Hey, Superman You Won=t Be Invincible Forever. USA Weekend, The Daily Times 6/8/01. pp. 6-9.

Morgenthau, H..J. & Thompson, K. (1985). Politics Among Nations. 6^th ed. N.Y.: Alfred A. Knoph.

Mouritzen, H. (1988). Finlandization, Towards a General Theory of Adaptive Politics. Avebury:Aldershot.

Nachmais, C.H. & Nachmais, D. (1996). Research Methods in the Social Sciences. New York: St. Martin=s Press.

Niemann, Y. & Dovidio J. (1998). Tenure, race/ethnicity and attitudes toward affirmative action: a matter of self-interest? Sociological Perspectives, 41, Retrieved December 5, 2001, Expanded Academic Index.

Organski, A.F.K. (1968). World Politics. (2^nd ed.) New York: Alfred A. Knopf.

Organski, A.F.K. & Kugler, J. (1980). The War Ledger. Chicago: University of Chicago Press.

Rothgeb, J.M. (1993). Influence of Force in the Contemporary International System. New York: St. Martin=s Press.

Shavell, S. (2000). Other Functions of Sanctions: Incapacitation, Rehabilitation and Retribution. In Principles. Retrieved November 11, 2001 from Boston University, Economics Web site: http://econ.bu.edu/Weiss/Ec337/Shavell/bg231e.pdf.

Siegal, S. & Castellan, N.J. (1990). Nonparametric Statistics, New York: McGraw-Hill Companies.

Simons, G. (1994). The United Nations Chronology of Conflict. Houndmills, U.K.: Macmillan Press.

Singer, M.R. (1972). Weak States in a World of Powers. New York: The Free Press.

Spainel, J. & Wendzel, R.L. (1996). Games Nations Play. (9^thed.) Washington, D.C.: CQ Press.

Tindell, J. & Vincent, J.E. (1970). The Effects of Different Ratios of Force on Aggression: An Experiment, Quarterly Journal of the Florida Academy of Sciences 33, 294-320.

Vincent, J. E. (2001). Alley Game Orientation. Unpublished.

Vincent, J. E. & Schwerin, E. (1971). Ratios of Force and Escalation in a Game Situation, Journal of Conflict Resolution 5, 489-511.

Waltz, K. (1990). Man, the State and War. New York: Columbia University Press.

Walzer, M. (1988). The Theory of Aggression. In S. Luper-Foy (Ed.), Problems with International Justice (pp. 151-162). Boulder: Westview Press. (Reprinted from Just and Unjust Wars: A Moral Argument with Historical Illustrations, 1977, New York: Basic Books).

APPENDIX A

Alley Game Survey

Do NOT place your name anywhere on this paper

Place an X on where you think you fall: on a scale of 1-5

Example: My level of athletic ability is:

Low 1________2________3____X___4________5________ High

------------------------------------------------------------------------------ --------------------------------------

Do you agree with the United Nations polices?

Low 1________2________3________4________5________ High

I am:

Liberal 1________2________3________4________5________ Conservative

I stand up for what I think is right:

Agree 1________2________3________4________5________ Disagree

I think violence is an acceptable solution to solving human conflict and differences:

Agree 1________2________3________4________5________ Disagree

I will take great measures to avoid being in a fight:

Agree 1________2________3________4________5________ Disagree

Most people are trustworthy:

Agree 1________2________3________4________5________ Disagree

My level of athletic ability is:

Low 1________2________3________4________5________ High

I think affirmative action is a good policy:

Agree 1________2________3________4________5________ Disagree

How do you rate your level of sympathy to others?

Low 1________2________3________4________5________ High

How do you view your personality in dealing with others?

Submissive 1________2________3________4________5________ Aggressive

How likely are you to commit a felony if you know you cannot be penalized?

Unlikely 1________2________3________4________5________ Very likely

Personal characteristics of the respondent:

Gender: M F

Age:

Military Background:

THE ALLEY GAME

Condition 1

You are walking through an alley to avoid a long walk in a large city, since you will be late to an important meeting if you take the long way around. You always carry a large knife. Suddenly a small thin man jumps in your way and demands $20 to pass. Before you can speak he pushes you backward with considerable force. Between the three options listed, which, would you chose?

______ Push him out of the way to proceed forward. (If you pick this option, go to the spot marked *****)