# Four dot points to end the chapter

• • Utility theory is a theory of measurement.
• • We can use utility theory to convert the payoffs to terrorist attacks (media attention, grassroots supporters, recruits, territory, fatalities etc.) into utility numbers to provide a more nuanced measurement of the relative attractiveness of different terrorist actions to terrorist groups. Terrorist groups don’t simply choose the action with the highest absolute payoff.
• • Both expected utility theory and prospect theory can be used to work out preference rankings or orderings of risky prospects, including attack methods. We can see how different assumptions about the terrorist decision-maker change these rankings.
• • Expected utility theory and prospect theory can be used together. The latter provides us with more pattern predictions. Together, both theories provide us with a useful analytical framework with which to inspect human decisionmaking, including our own.

# Notes

• 1 Di Telia and Schargrodsky (2004).
• 2 For example, transitivity. If you prefer A to В and В to C then you must prefer A to C.
• 3 It should be noted that measurements of utility we refer to are strictly within the framework of von Neumann and Morgenstern’s (1047) expected utility theory.
• 4 Also, his wife was not released from prison until 1985.
• 5 The following example can be re-worked with any payoff, including media attention (see Pohl 2015).
• 6 We have chosen this period deliberately, before the ‘unarmed assaults’ of 2015 and 2016 discussed in a previous chapter.
• 7 These estimates do need to be recalculated every so often to incorporate innovations in terrorist behaviour and law enforcement. However, like batting averages, once a reasonable number of results have been recorded, the average is not very sensitive to new observations.
• 8 This gives the probability that some variable, in this case inflicted fatalities, is exactly equal to some number. A basic statistics program can be used to run the calculation.
• 9 Schoemaker (1982) also notes that expected utility theory can be used prescriptively.That is, as a model that prescribes the correct choice.Thaler (2016), for example, notes that he himself would use expected utility theory to help him make the correct choice. This is a strong endorsement from one of the most cited of all of the behavioural economists.
• 10 The alternative purposes of economic models were discussed by Schoemaker (1982) in one of our favourite papers.
• 11 To be precise, the views of which we speak concern the payoffs and probabilities that characterise the alternatives. If these change, both the PT and expected utility measurements will change. Both measurements can also change over time as gains and losses accumulate. The difference between the two theories is that PT and CPT measurements may change based on a changed perspective or reference point even if the underlying payoffs and probabilities and everything else remains the same.
• 12 Prospect theory does not address the possibility of‘mistakes’ and as such does not represent a benchmark against which decisions can be judged in terms of optimality or correctness. Within PT and CPT, the decision-maker fully understands the situation and deals with it in particular ways that lead to divergences from expected utility theory. It is not mistakes that lead to such decisions. By contrast, one can certainly use the expected utility measurements as a clear benchmark against which to assess decisions and judge divergences as ‘mistakes’ if one so wishes.
• 13 On the basis of average fatalities and standard deviation for the period 1970-2014.