Rationality and choice



Rational choice theory

Since the 1970s, there has been the development of rational choice' models, with the common assumption that most criminal acts are based on a form of means-end deliberation (that is, the criminal act is a means to a particular end or goal). It does not mean that the decisions are rational in an objective sense. In other words, there is a reasoning criminal' - a rational decision-maker (Hollin, 1989).

The concept of environmental criminology' assumes that criminals choose when to commit a crime based on environmental opportunities and situational constraints. Cohen and Felson (1979) analysed patterns of property crime in the 1960s in America and found they reflected changes in society. For example, the increasing number of married women who worked meant that houses were empty during the day, and thus there was an increase in daytime burglary. Similarly, there has been a lot of interest in opportunist crime' - crimes committed without planning. An example of this might be when keys are left in the ignition of a parked car.

An alternative view is known as the ‘deterrence hypothesis’. Crime is due to a lack of general deterrence (that is, the threat of punishment) or specific deterrence (that is, the actual punishment of future behaviour). So a calculation is made by the criminal of subjective benefits against the costs of deterrence before a crime is committed. This simple equation is probably more applicable to instrumental crimes' (with material benefits, like robbery), than ‘expressive crimes' (like sexual offences which seek non-material needs) (Blackburn, 1993).

Thus studies have looked at the perception of risk. In a hypothetical experiment with students, Rettig and Rawson (1963) found that the preference for unethical choices was influenced by many factors, but the severity of the punishment was most important. Meanwhile, other research found that for both offenders and non-offenders monetary gain was much more important than punishment. Only one-third of the participants concentrated on the severity of the penalty.


Theories from economics about rationality have been applied with limited success to crime. The expected utility maximization model, for example, assumes that individuals collect all the relevant information beforehand and calculate the objective probabilities of their actions. This is unrealistic. A better approach is to see decisions as based on subjective probabilities (like the available information in the memory).   Tversky and Kahneman (1981) have found that different choices will be made depending on how the problem is ‘framed' (or described).


In contrast to determinist theories of crime which presuppose offenders are passively driven to commit offences, rational choice theories of crime focus on the cognitive aspects of offending. Rational choice theories adopt a perspective whereby offenders are seen to think about their actions and exercise a judgement as to whether the risk of detection is outweighed by the potential reward before they choose to offend (Cornish & Clarke, 1986). This can be seen as an information processing approach, which assumes that offenders make rational decisions about the benefits of criminal behaviour after collecting and evaluating all the relevant information. Van Den Haag (1982) expresses the concept of rational choice very clearly:

I do not see any relevant difference between dentistry and prostitution or car theft, except that the latter do not require a license ... The frequency of rape, or of mugging, is essentially determined by the expected comparative net advantage, just as is the rate of dentistry and burglary. The comparative net advantage consists in the satisfaction (produced by the money or by the violative act itself) expected from the crime, less the expected cost of achieving it, compared to the net satisfaction expected from other activities in which the offender has the opportunity to engage. Cost in the main equals the expected penalty divided by the risk of suffering it. (1982)

There are some inherent difficulties with rational choice theory.  It does not explain the impulsive crimes that many young men commit (Crawford, 1998). It could be argued that rationality is also impaired in some circumstances, for instance when an offender is mentally disturbed or under the influence of alcohol. Moreover, when offenders are completing the true cost—benefit analysis they may not be in full possession of all the facts or the true likelihood of apprehension, so their decisions in these circumstances cannot necessarily be viewed as rational.

A major difficulty, however, is the finding that offenders at different stages of their criminal career will tend to judge crime opportunities differently, and may also take into account previous experiences of punishment. This weakens any general theory of crime, which attempts to explain offending in terms of its onset or beginning, its perseverance or continuation, and its end. Thus Carroll and Weaver (1986) found that when they analysed offenders’ verbalised thoughts during crime simulations, experienced shoplifters were adept at evaluating opportunities for offending but used the information selectively, while novice shoplifters focused on the major question of whether to offend at all. Their cognitive processes were thus utilised quite differently, and the concepts of rationality and choice were shown to be problematic.

Yochelson and Samenow (1976) also focus on cognitive processes but emphasise the role of cognitive dysfunction in criminal behaviour. The concept of rationality therefore disappears. They suggest that criminals have quite distinct and erroneous thinking patterns’ which differentiate them from non-criminals. From their interviews with 240 male offenders they conclude that criminals may be less intelligent than non-criminals but they are essentially in control of their lives and their criminality is the result of choices made from an early age. Yochelson and Samenow emphasise cognitive processes, which lead to a distorted self-image and result not only in criminal choices but a denial of responsibility. They identify 40 thinking errors’ made by criminals, the errors falling into three categories:


      Criminal thinking patterns which are characterised simultaneously by fear and a need for power and control. Other features include a search for perfection, lying, and inconsistencies or fragmentation of thinking.

      Automatic thinking errors, which include a lack of empathy and trust, a failure to accept obligations, and a secretive communication style.

      Crime-related thinking errors, which include optimistic fantasising about specific criminal acts with no regard for deterrent factors.


Yochelson and Samenow are suggesting therefore that criminals are not necessarily impulsive, that they will have planned and fantasised about their actions, and it is these thinking patterns which need to be confronted in treatment. Wulach (1988) however, has criticised their approach, pointing out that Yochelson and Samenow are simply describing psychopaths and their theory cannot therefore be regarded as a general theory of crime.


Cook (1980) focuses on the potential offender as someone who operates rationally. He asserts that potential criminals weigh the possible con­sequences of their actions, both positive and negative, and take advantage of a criminal opportunity only if it is in their interest to do so. In setting out this assertion, Cook looks back to Jeremy Bentham (1748—1832) and to one of his basic propositions: “The profit of the crime is the force which urges a man to delinquency. The pain of the punishment is the force employed to restrain him from it. If the first of these forces is the greater, the crime will be committed; if the second, the crime will not be committed” (quoted by Zimnng and Hawkins, 1973, p. 75). Cook accepts that people will respond differently to equivalent criminal opportunities because they differ: in their willing­ness to take risks; in their “preference for honesty;” in their evaluation of “profit” from a crime — most of all in crimes in which the pay-off is something other than money (rape, vandalism, etc.); and in their objective circumstances (income, the value they place on their time, their skills in successfully committing crimes and in evading capture, their reputations in the community). Both profit and pain are evaluated differently by different persons, but it is true for people in general that changes in either the probability or in the average severity of pain will cause some people to change their minds and, hence, their behaviour. Small changes in either probability or severity will affect only those close to the “point of indifference.”


The theory in outline


Clarke and Cornish have elaborated this rapidly developing approach into the rational choice theory of crime (RCT), which makes several assumptions, set out by Cornish and Clarke (1987):

1.      Offenders seek to benefit themselves by criminal behaviours:

2.      Doing so involves making decisions and choices, however rudi­mentary these might be.

3.      The decision-making process is constrained by the time available (many criminal opportunities have a limited life-span), by the avail­ability of relevant information (frequently this will be incomplete) and by the offender’s own cognitive abilities (related, presumably, to verbal IQ). It follows that rationality will be limited, rather than complete.

4.      Both the decision-making process and the factors taken into account by offenders vary greatly at different stages of decision-making and between different crimes (and presumably also between different offenders within crimes; there are marked differences in success-rates, with planning ahead being a key feature of the more successful offenders).

Cornish and Clarke (1987) argue the need to be (a) crime-specific when analysing criminal choices and (b) to treat decisions as relating to varying stages of the involvement of an offender in a particular crime. Thus, they distinguish between initial involvement, the event, con­tinuation, and desistance. The general sequence of involvement should be analysed separately from factors such as target selection which are related to the crime itself. (Essentially, this group of factors all concern the performance phase of involvement.)

The rational choice perspective asserts that specific crimes are chosen by offenders and are committed for specific reasons. We need to understand the factors which are taken into account by offenders when they perform a rudimentary cost—benefit analysis of a range of factors, including the incentive, or anticipated pay-off, the risk involved, and the skills needed, all in relation to their goals, motives, experience, abilities, expertise, and preferences. All of these variations combine to make criminal opportunities differentially attractive to particular individuals and groups. They are termed by Cornish and Clarke (1987) their choice structuring properties.

The following is a list of choice structuring properties for crimes involving cash (i.e., money rather than goods, from bank robbery to computer fraud, Cornish and Clarke 1987).

1.      Availability (number of targets, accessibility)

2.      Awareness of method (i.e., technical know how).

3.     Likely cash yield per crime.
4.     Expertise needed.
5.     Planning necessary.
6.     Resources required.
7.     Solo versus assistance required.   

8.     Time required to commit.
9.     Cool nerves required.                   -

10.    Risk of apprehension.

11.    Severity of punishment (if caught).

12.    Instrumental violence required.

13.    Confrontation with victim.

14.    Identifiable victim.

15.    Social cachet [in the criminal world] (safebreaking versus mugging).

16.    Fencing accessories (getting rid of any goods stolen).

17.    Moral evaluation.

The above were derived by Cornish and Clarke (1987) on an a priori basis. Detailed research might result in their modification, and they are certainly not all taken into account by every offender.

Clarke and Cornish (1985) base RCT very explicitly on economic models of criminal decision making [which] “effectively demystify and routinise criminal activity. Crime is assumed … to involve rational calculation and as an economic transaction or as a question of occupational choice” (Clarke and Cornish 1985, p. 156). Economic models of crime are not confined to those motivated by financial gain, but are extended also to crimes of violence.

Another source of RCT is to be found in information-process­ing models and strategies in relation to real-life decision-making (Kahneman, Slovic, and Twersky 1982), which are then applied to criminal decision-making (e.g., Carroll 1982).


Research findings


The most important evidence concerning the rational choice theory of crime comes from recent research into specific offences carried out by professional property offenders, and is exemplified below. But, the scene was set by several earlier researchers, who presented respon­dents with hypothetical criminal situations to ascertain both the extent to which they could combine information concerning rewards and costs, and which are the more important of such variables. In a series of studies, Rettig and colleagues (e.g., Rettig 1966, Rettig and Turroff 1967) gave students a hypothetical criminal opportunity in which several potential determinants of a decision to steal were systematically and simultaneously varied. The greatest effect was exerted by the amount of punishment involved — this exceeded both the probability of detection and the incentive present.

Feldman (unpublished data, described in detail in Feldman, 1977) adapted Rettig’s approach to younger and academically much less able populations of adolescent institutional offenders and to school pupils matched for age and intelligence. The detailed results make clear the importance of the exact combination of rewards and costs to a decision whether or not to offend — in this case breaking in to a house. Throughout, the level of cost involved exerted a more powerful effect when rewards were high than when they were low. For both groups of respondents the order of importance of rewards and costs was the same: within rewards, the level of monetary incentive was followed by the degree of boredom experienced by the potential offender and then the certainty of the house containing money; the most important cost (deterrent) was the presence of a policeman on a nearby beat, followed by the level of punishment if caught, the certainty of detection and the difficulty of breaking into the house. The vast majority of “yes” answers from both groups were “rational” — boys rarely answered yes to a low temptation pairing (low reward, high cost) unless they also did so to the medium and high temptation versions of the same pair of determinants. The fact that this degree of rationality was displayed by boys most of whom were below the average level of intelligence supports the rational choice theory of crime. However, only two vari­ables were presented at a time, one a temptation and one a deterrent, making the task cognitively more simple, but also much less like the real world in which several variables have to be considered simul­taneously, often under time pressure. Two other early simulation studies (Piliavin, Hardyck, and Vadum 1969, Piliavin, Vadum, and Hardyck 1969) also gave some support to the rational choice approach to crime.

More recent studies of subjectively perceived costs and benefits (e.g., Bridges and Stone 1986) have shown the importance of prior experi­ence of punishment and the nature of that experience for the perceived risks of being caught for a subsequent offence. Such results indicate the importance for decision making of the individual’s previous criminal history (rewards as well as costs) in addition to the current situational factors — exactly as would be expected by a broadly behavioural view of criminal behaviour. The more information available on both, the more accurately can one predict an individual’s response to a particular criminal opportunity (Cimler and Beach 1981)


A specific property offense: burglary


Clarke and Cornish (1985) have provided a detailed illustration of RCT, applied to residential burglary, specifically in a middle-class suburb. They point out that the offenders involved are generally older and more experienced than those operating in housing pro­jects, but less sophisticated than those specializing in much wealthier residences.

Initial involvement: Boxes 1—3 refer, very briefly, to background, cognitive, and learning factors. Next, the budding burglar (who has probably already carried out more minor offences) passes through a series of decision points, of which two, boxes 7 and 8, are considered by Clarke and Cornish to be of particular importance. “...Box 7 is the individual’s recognition of his ‘readiness’ to commit this particular offence to satisfy certain of his needs for money, goods or excitement... [he] has decided that under the right circumstances he would commit this offence” (Clarke and Cornish

1985, p. 167).

“The second decision (box 8)... is precipitated by some chance event. . . the. . . need for money.., drinking with associates who suggest committing a burglary. . . he may perceive an easy oppor­tunity. . . during the course of his routine activities” (Clarke and Cornish 1985, p. 167).

Next comes a consideration of the variables associated with the criminal event, and concern successive yes/no decisions (this area and that house within the area). Both decisions emphasize the relatively greater ease of access and escape and the greater affluence of the final choice. Clarke and Cornish point out that this is an idealized model and that full rationality may often not be attainable: “. . . the decision process may be telescoped, plan­ning may be rudimentary, and there be may last minute (and perhaps ill-judged) changes of mind.., alcohol may cloud judgement” (Clarke and Cornish 1985, p. 170).

In many cases, burglars commit hundreds of offences over a period of years. This process of continuance focuses on a number of important consequences of a long-term criminal career. Such consequences will include criminal convictions as well as “skin of the teeth” escapes from capture. Convictions only accelerate the move away from the straight world, ..... as opportunities to obtain legitimate work decrease and as ties to family and relations are weakened” (Clarke and Cornish 1985, p. 172).

The final stage is that of desistance. Clarke and Cornish point that there is less information about this stage than the others, but the sequence set out rings true (the move to less strenuous forms of crime, to non-crime, or to some mixture of the two, is likely to start before middle-age).

Clarke and Cornish have provided us with a well worked-out model of the rational decision approach to residential burglary. Their work is supplemented and illustrated by a report by Walsh (1980), based on interviews with 45 men in British prisons who had been convicted of burglary. Their ideal target was a business firm rather than a private house (more to be stolen) and, while half used information, the other half burgled on impulse (presumably, the latter were more likely to be caught and hence were more available for interview). There was much fear concerning being interrupted during a burglary, with the con­sequent possibility of violence and, hence, of a more severe sentence if arrested. They found it easier to justify to themselves a business than a household target (particularly if the latter were “ordinary”), adding to their preference for the former. They saw themselves as desisting with increasing age (the risks of capture increased, sentences stiffened and the risk/return balance generally less favourable).


Tobacco sales to minors

Informed by rational choice theory, this paper identifies the factors that predict illegal sales of tobacco to under age youth. Data for this study are based on 439 merchant compliance checks in Ontario. Analysis reveals that enforcement activity has a significant negative effect on sales to minors. This effect, however, is slightly weakened with the introduction of background factors (i.e., rural/urban distinction, type of business operation), and all but eliminated once event factors are taken into consideration. The event factors that have the strongest independent effects on sales are time of day (evening), age and gender composition of the youth attempting to purchase, and the legal compliance behaviour of tobacco merchants (asking for age and ID). These findings help to support a rational choice model of offending, and suggest that the situational context of the criminal event relates strongly to decisions to offend. O'Grady,-William; Asbridge,-Mark; Abernathy,-Tom  Canadian-Journal-of-Criminology. 2000 Jan; Vol 42(1): 1-20


Different framing of problems

Research has shown that people respond differently to the same problem depending on how it is framed.

Problem 1

Imagine that there is an outbreak of a rare disease, which is expected to kill 600 people. Two different programmes to combat the disease have been proposed. Assume that the exact scientific estimate of the consequences of the programme is as follows.

1.      · If programme A is adopted, 200 people will be saved.

2.      · If programme B is adopted there is a one-third probability that 600 people will be saved, and a two third probability that no people will be saved.

Which programme would you choose?

Using students, Tversky and Kahneman found that 72% chose A and 28% B.

Problem 2

Use the same scenario as Problem 1, but with these options.

1.      · If programme C is adopted, 400 people will die.

2.      · If programme D is adopted, there is a one-third probability that nobody will die, and a two-third probability that 600 people will die.

Which programme would you choose?

Tversky and Kahneman found that 22% choose C and 78% programme D.

If looked at carefully, all the programmes are offering the same outcome, but they are framed (worded) differently, and this influences the decision made.


Kevin Brewer (2000), Psychology and Crime, Heinemann, ISBN 0-435-80653-X.

Philip Feldman (1993), The Psychology of Crime, Cambridge, ISBN 0-521-33732-1

Julie Harrower, (2001) Psychology in Practice, Crime, Pages 31-33, Hodder & Stoughton, ISBN 0-340-84497-3.