“Ultimatum Bargaining Behavior: A Survey and Comparison of Experimental Results,” by Werner Guth and Reinhard Tietz, Journal of Economic Psychology, 1990, Vol 11. 417-449, http://ideas.repec.org/a/eee/joepsy/v11y1990i3p417-449.html.
“Genesis of Suicide Terrorism,” by Scott Atran, Science, ,7 March 2003, Vol. 299 1534-1539, http://www.sciencemag.org/cgi/reprint/299/5612/1534.pdf (also at Fascinet and Juanma1998).
“Journal: Reducing the Need to Conenct,” by John Robb, Global Guerrillas, 2 May 2004, http://globalguerrillas.typepad.com/globalguerrillas/2004/05/journal_reducin.html.
One of this week’s articles made something of a splash in the blogosphere, with two livejournalers — and John Robb — previously commenting on the findings. John concludes:
Humiliation is driving the growth of terrorist networks in the Middle East by increasing the susceptibility of people to terrorist recruitment (see Atran, “The Genesis of Suicide Terrorism” PDF). This is an excellent example of why the occupation of Iraq is working at cross-purposes to the GWOT (global war on terrorism). It is also a good demonstration of what military thinkers mean when they advocate a policy of de-escalation in Iraq and the GWOT. De-escalation is the systematic reduction of factors that drive people to terrorism and not a sign of weakness.
Yet the article itself is mostly institutional (arguing that the effective method is destroying suicide-recruiter-organizations), and ambiguous on the role of “humiliation”
Although humiliation and despair may help account for susceptibility to martyrdom in some situations, this is neither a complete explanation nor one applicable to other circumstances.
Ending occupation or reducing perceived humiliation may help, but not if the population believes this to be a victory inspired by terror (e.g., Israelâ€™s apparently forced withdrawal from Lebanon). (Atran 2005 4,5)
A better interpretation of the article focuses on Wary Cooperator model of human behavior. Humans are not rational, economic-man style machines, as has been long demonstrated through laboratory experiments. Who gets what may matter, but not the loser idea of “humiliation”
One reason why ultimatum bargaining become a widely known experimental paradigm is that experimentally observed ultimatum bargaining behavior clearly contradicts the most obvious rationality requiremetns of game theory and also of economic theory.
Experimentally observed ultimatum bargaining behavior reveals how considerations of distributive justice seriously destroy the prospects of exploiting strategic power.
(Guth and Tietz 1990 446)
Man may be naturally rational on a group level, but is naturally irrational on an individual level.
Rather, humans wish to cooperation to irrational levels but then punish to irrational levels when they are treated “unfairly.” This is what we see in places like Iraq, for example. With the fall of Saddam, the Sunni Arabs’ apartheid state (which allowed 15% of the population to control the other 85%) collapsed. Humans are naturally generous to their group and naturally vicious against those who treated them unfairly. Thus “altruistic super-punishers” emerge, greatly punishing a “cheater” (the progressive American and Iraqi forces) to benefit their group at the cost of their lives.
The rest of my notes are below the fold.
Thus, the kamikaze (â€œdivine windâ€) first used in the battle of the Philippines (November 1944) were young, fairly well educated pilots who understood that pursuing conventional warfare would likely end in defeat. (Atran 2003 1535)
In the Battle of Okinawa (April 1945) some 2000 kamikaze rammed fully fueled fighter planes into more than 300 ships, killing 5000 Americans in the most costly naval battle in U.S. history. Because of such losses, there was support for using the atomic bomb to end World War II (7). (Atran 2003 1535)
Through its multifaceted association with regional groups (by way of finance, personnel, and logistics), Al-Qaida aims to realize flexibly its global ambition of destroying Western dominance through local initiatives to expel Western influences (11). According to Janeâ€™s Intelligence Review: â€œAll the suicide terrorist groups have support infrastructures in Europe and North America.â€ (12). (Atran 2003 1535)
Recent treatments of Homeland Security research concentrate on how to spend billions to protect sensitive installations from attack (14, 15). But this last line of defense is probably easiest to breach because of the multitude of vulnerable and likely targets (including discotheques, restaurants, and malls), the abundance of would-be attackers (needing little supervision once embarked on a mission), the relatively low costs of attack (hardware store ingredients, no escape needs), the difficulty of detection (little use of electronics), and the unlikelihood that attackers would divulge sensitive information (being unaware of connections beyond their operational cells). (Atran 2003 1535)
A first line of defense is to prevent people from becoming terrorists. (Atran 2003 1535)
What research there is, however, indicates that suicide terrorists have no appreciable psychopathology and are at least as educated and economically well off as their surrounding populations. (Atran 2003 1535)
Psychologist Stanley Milgram found that ordinary Americans also readily obey destructive orders under the right circumstances (18). When told by a â€œteacherâ€ to administer potentially life-threatening electric shocks to â€œlearnersâ€ who fail to memorize word pairs, most comply. (Atran 2003 1536)
Suicide terrorists generally are not lacking in legitimate life opportunities relative to their general population. As the Arab press emphasizes, if martyrs had nothing to lose, sacrifice would be senseless (24): â€œHe who commits suicide kills himself for his own benefit, he who commits martyrdom sacrifices himself for the sake of his religion and his nation. . . . The Mujahed is full of hopeâ€ (25). (Atran 2003 1536)
Research by Krueger and Maleckova suggests that education may be uncorrelated, or even positively correlated, with supporting terrorism (26). (Atran 2003 1536)
Underemployment also seems to be a factor among those recruited to Al-Qaida and its allies from the Arabian peninsula (27). (Atran 2003 1537)
Suicide terrorists apparently span their populationâ€™s normal distribution in terms of education, socioeconomic status, and personality type (introvert vs. extrovert). Mean age for bombers was early twenties. Almost all were unmarried and expressed religious belief before recruitment (but no more than did the general population). (Atran 2003 1537)
Except for being young, unattached males, suicide bombers differ from members of violent racist organizations with whom they are often compared (29). Overall, suicide terrorists exhibit no socially dysfunctional attributes (fatherless, friendless, or jobless) or suicidal symptoms. They do not vent fear of enemies or express â€œhopelessnessâ€ or a sense of â€œnothing to loseâ€ for lack of life alternatives that would be consistent with economic rationality. (Atran 2003 1537)
Other research suggests that most people have more moderate views than what they consider their group norm to be. Inciting and empowering moderates from within to confront inadequacies and inconsistencies in their own knowledge (of others as evil), values (respect for life), and behavior (support for killing), and other members of their group (42), can produce emotional dissatisfaction leading to lasting change and influence on the part of these individuals (43). (Atran 2003 1538)
In a simple reward allocation experiment (see, for instance, Mikula 1973; and Kahneman et al. 1986a,b) a subject has to allocate the total reward c (> 0) among two individuals who both contributed to obtain the total reward c.
(Guth and Tietz 1990 419)
The ‘auction winners’ (data) in table 1 are the plays observed by Guth and Tietz (1985, 1986) who auctioned the position of ultimatum bargainers before letting the auction winners play the ultimatum bargaining game… Once subjects were familiar with bidding in second highest bigd-price auctions they were told that they were then going to bid for strategic positions. They were informed how to play ultimatum bargaining games with amounts c of 15, 55, and 100 German marks and then asked to bid either for the position of player 1 or for the one of player 2, i.e., we conducted an independent auction for both ultimatum bargaining positions and every single game. Afterwards the auction winners were determined and privately informed about the price (the next highest bid for the same position) which they had to pay for their strategic position.
(Guth and Tietz 1990 422)
A previous loss of efficiency by failing to reach an agreement on the bigger cake seems to increase the willingness to punish dramatically… An attempt to explain the relative demand behavior of player 1 globally by a linear regression function of the amount c and the experience parameter ‘game’ fails.’”
(Guth and Tietz 1990 425)
There is no significant difference in the proportion of equal splits of the Psych/Psych and the Psych/Com group (alpha = .37), but the average share of both these groups is significantly greater than the corresponding value of the Com/Psych group (alpha < .05). Thus, psychology students tend to propose equal splits more often than students of business administration.”
(Guth and Tietz 1990 431)
Although this shows that subjects hardly ever apply game-theoretic reasoning to determine their behavior but that the extent of paying attention to strategic aspects can be strongly affected by the experimental environment whenever the monetary incentives are rather low.
(Guth and Tietz 1990 431-432)
Table 4 contains the experimental results of Binmore et al. (1985: fig. 1). The level of aggregation in table 4 is determined by Binmore et al. who did not list individual decision data. Whereas the main tendency (30 out of 82 observations) in the 1st game is to propose an equal split, the strong tendency (50 out of 81 observations) in the 2nd game is to behave like a game theorist (Binmore et al. 1985: 1779).
(Guth and Tietz 1990 436)
In our view, this has important implications for the interpretation of many game-theoretic results as, for instance, the Folk Theorems of game theory (see Aumann 1981; Fudenberg and Maskin, 1986; and also the critical discussion of Guth et al. 1988).
(Guth and Tietz 1990 447)