The purpose of punishment, then, is nothing other than to dissuade the criminal from doing fresh harm to his compatriots and to keep other people from doing the same.
—Cesare Beccaria, 1764
If the quote above seems familiar, it is probably because Beccaria’s words reveal some deeply held beliefs about the role of criminal punishment in contemporary American society. In his treatise on the role of legal sanctions, On Crimes and Punishments, Beccaria advanced a theory of punishment which has come to be known as the theory of criminal deterrence.
Deterrence theory assumes that people are innately rational in the sense that they are able to comprehend and compare the costs and benefits associated with various courses of conduct. Actions which, on balance, yield more costs than benefits will be avoided, while actions providing more benefits than costs will be pursued. According to this theory, legal punishments should be formulated for crimes so that they impose costs for offenders that just exceed the benefits derived from their wrongdoing. (Overly harsh punishments are to be avoided to prevent the commission of more crimes designed to help an offender escape harsh punishment.)
This brief overview of deterrence theory is relevant because Beccaria’s ideas continue to play a central role in criminal justice policy and practice in Alaska and elsewhere. Although some discussions of criminal law give priority to sanctions as a form of retribution for an offense, and still others place primary emphasis on the notion of criminal incapacitation, the idea of deterrence is an element in all contemporary conversations justifying criminal punishments. Sometimes punishments are said to be aimed specifically at the person responsible for a crime (specific deterrence); sometimes they are said to be a warning to all who might contemplate committing the same or a similar offense as the convicted offender (general deterrence); and sometimes punishments are discussed as being directed at both actual and potential offenders. In all instances, however, the idea of preventing future crimes through instilling fear of punishment prevails.
This article examines one half of the cost-benefit equation proposed first by Beccaria and subsequently incorporated into American criminal jurisprudence: legal sanctions. More specifically, I examine legal sanctions for gun crime in terms of how such sanctions are perceived by the general public. The perception of criminal sanctions is integral to deterrence, because if an actual or potential offender is not conscious of their existence or likelihood of enactment, she or he cannot include them in a cost-benefit calculation, and therefore cannot be deterred by them.
The data discussed here came from a random telephone survey of Anchorage households conducted by the Justice Center in spring 2004. In all, 585 individuals in 551 households were interviewed.
Dependent Variables: Deterrence
Since Beccaria’s time, theorists have asserted that there are three facets of criminal sanctions that influence their effectiveness: certainty, celerity, and severity. The certainty, or likelihood, of a particular sanction being applied is important, because if a potential offender is aware of the punishment for a particular crime, but does not think that it is likely to be applied, she or he will not be discouraged from engaging in prohibited conduct. Punishment celerity is the amount of time between the commission of an offense and the infliction of a punishment. It is thought that a small amount of time between crime and punishment helps cement the two together—crime and punishment—in the minds of actual and potential offenders. The third dimension of effective deterrence—severity—reflects the idea that the costs incurred must exceed the benefits derived from prohibited behavior if future crimes are to be prevented.
Four items were included in the survey to measure Anchorage residents’ perceptions of sanction certainty spanning three stages of the criminal justice process: detection/apprehension, charging/prosecution, and conviction. Respondents were asked to estimate the likelihood (very likely; somewhat likely; neither; somewhat unlikely; very unlikely) of the following events:
- A person carrying a gun illegally would be caught by authorities.
- A person caught committing a gun crime would be prosecuted for that offense.
- A person charged with a gun crime would be convicted in state court.
- A person charged with a gun crime would be convicted in federal court.
Sanction celerity was measured using a single item which asked respondents to estimate how quickly they thought a criminal prosecution would take place for a person who was caught committing a gun crime. As with the certainty measures, respondents were asked how likely the following was to occur:
- A person caught committing a crime with a gun would be prosecuted swiftly.
Finally, sanction severity was measured using two questionnaire items which asked respondents to estimate the likelihood of an offender’s sanction being reduced via a plea agreement and of an offender receiving a severe punishment:
- A prosecutor would allow a gun offender to plead guilty to a less serious offense in order to secure a conviction.
- A person convicted of a gun crime would receive a long prison sentence.
Independent Variables: Demographics
Included in the survey as independent variables (that is, variables that might influence how a person responds to a survey item) were several questions tapping respondents’ demographic characteristics: age, race/ethnicity, gender, level of formal education, current individual income and current work status. Respondent answers to the each of the deterrence measures are compared across each of these factors in order to identify differences in response patterns.
Description of Sample
Table 1 contains the distributions for the demographic variables, with data from the 2000 Census provided for comparison. The survey sample was generally representative of the Anchorage population. The largest observed differences were found in the three social variables included in the survey: education, individual income, and work status. In terms of education, the survey under-sampled those on the lower end of the spectrum—those who have a 12th grade education or less. The survey also under-represented residents with annual incomes of less than $40,000. In addition, the survey did not capture an equal proportion of people who were unemployed. However, none of the observed percentage differences presented in Table 1 was found to be statistically significant. Nevertheless, because there is evidence to suggest that certain groups of eligible participants may have been excluded from the survey, sample weighting was used in the analysis.
Descriptive statistics for all of the deterrence measures are presented in Table 2. Interviewers read each survey item to respondents, and then asked them how likely they thought each one was to occur: very likely; somewhat likely; neither likely nor unlikely; somewhat unlikely; or very unlikely. Each response was given a numeric score ranging from 2 (very likely) to -2 (very unlikely). With this scoring system, a positive mean score indicates that, on average, Anchorage residents thought a particular event was more likely than unlikely to occur, whereas a negative mean is interpreted as more unlikely. Mean scores close to zero are indicative of a generalized ambivalence on a particular item, while scores approaching 2 or -2 demonstrate strong views in either direction. The number of respondents who answered each item, the median (i.e., middle score), mean (i.e., average score), and standard deviation are provided for each deterrence measure.
Overall, it appears that Anchorage residents perceive there to be a better than even chance that persons who commit gun crimes will be legally sanctioned for their misdeeds, although this generalized confidence varies across stages of the criminal sanctioning process. For example, Anchorage residents perceive that a person carrying a gun illegally is only slightly more likely than not to get caught by authorities (mean = .085); in contrast, residents perceived a much greater likelihood that legal action will be taken once an offender is caught (mean = 1.278); subsequently, respondents indicated an even greater perceived likelihood that after an offender has been caught and charged with a gun crime, a prosecutor will permit a guilty plea to a less serious offense in order to secure a conviction (mean = 1.349).
These data suggest that Anchorage residents’ consciousness of legal punishment for gun crime is fairly sophisticated and layered: their perceptions of the likelihood of sanctions vary according to which stage of the criminal justice process is being referenced. For some points in the process, respondents reported that they perceived a strong likelihood of legal sanctions (e.g., prosecution once an offender was caught; conviction of a gun offender once charged), but in others there were almost as many skeptics as there were believers in the criminal justice system (e.g., getting caught carrying gun illegally; swift prosecution of offender once caught; lengthy prison term for gun offender once convicted). Moreover, respondents demonstrated an awareness of one of the most fundamental aspects of criminal prosecution—plea agreements.
A re-examination of the seven perceived deterrence measures as a function of the demographic characteristics of respondents thought to shape people’s awareness and understanding of the legal system helps to “unpack” Anchorage residents’ perceptions.
Table 3 presents all of the sample means for each of the seven measures included in the survey, according to age, gender, race, education, individual income and current work status. Differences in perceptions about the criminal justice system can be detected by reading across rows. For example, the first row shows the mean scores for respondents under the age of 18. Respondents in this age group thought the event with the greatest likelihood of occurrence was a person being prosecuted for a gun crime once caught by authorities; meanwhile, they perceived a timely prosecution as the least likely to occur of all events. On the other hand, differences between categories of independent variables (e.g., age, gender, race, education, income, and work status categories) can be seen by reading down each column. For example, the last column (“a person convicted of a gun crime receiving a long prison sentence”) reveals a difference between male and female respondents in the perceived likelihood that a person who is convicted of a gun offense will be given a lengthy prison sentence. Females are less likely, on average, than males to think a person who is found guilty of a gun crime will be punished severely.
A procedure called one-way analysis of variance (ANOVA) was used to compare the sample means for each measure of perceived deterrence within each independent variable (age, gender, race, education, income, work status). The results of this analysis are presented in Table 4.
Of the six independent variables examined, significant differences in mean deterrence scores were found most consistently across age categories (significant differences detected for 5 measures), followed by gender (significant differences detected for 4 measures), income (significant differences detected for 4 measures), work status (significant differences detected for 4 measures), race/ethnicity (significant differences detected for 2 measures), and finally education (significant differences detected for 2 measures). In their totality, these findings show that individual-level perceptions of legal sanctions for gun crime vary according to a person’s demographic characteristics, but not every demographic characteristic is related to every measure of perceived deterrence. What follows is a brief overview of those differences that were found to be highly significant (p < .01).
The only highly significant mean differences across age categories were found for respondents’ views about the likelihood that a prosecutor would permit a gun offender to plead guilty to a less serious offense in order to secure a conviction. Follow-up analysis revealed that the mean difference detected by the ANOVA procedure was due to the difference between 18-19 year olds and all other age groups. Respondents in the 18-19 age group thought it less likely than others that a prosecutor would allow a gun offender to plead guilty to a less serious offense.
Men were more likely than women to believe that a person caught committing a gun crime would be prosecuted for that offense and that a person convicted of a gun crime would receive a lengthy prison sentence. Notably, women in this sample demonstrated a bit more skepticism of the criminal justice system in that, on average, they thought it more unlikely than likely that a person convicted of a gun crime would receive a long prison sentence.
ANOVA results revealed very specific race/ethnicity effects on individual perceptions of legal sanctions for the one deterrence measure meeting our p < .01 threshold: the likelihood of a gun offender being allowed to plead guilty to a less serious offense. The mean score for Asian respondents differed from all other race/ethnicity categories in that they were less likely to think that a prosecutor would plea bargain with a gun offender.
A respondent’s level of education was found to significantly influence two dimensions of deterrence: the likelihood that a person who carried a gun illegally would be caught, and the likelihood that a prosecutor would allow a gun offender to plead guilty to a less serious offense in order to secure a conviction. The perceptions of high school-educated respondents were different than those with lower levels of education (no degree) as well as those with more formal education (four-year degree or higher). Interestingly, respondents on the ends of the educational spectrum (very low and very high) actually perceived that detection was more unlikely than likely to occur.
The second deterrence measure that demonstrated significant mean differences across education levels was the likelihood that a prosecutor would plea bargain with a gun offender. Additional analysis showed that the primary difference in perceptions was between those without any sort of educational degree and all other educational categories, except those with a high school degree or equivalent. In all cases, those with the lowest level of education perceived a greater likelihood of a plea agreement.
Income was found to be highly influential in shaping individual perceptions of criminal sanctioning. The significant differences detected highlight how economic level shapes respondents’ view of the criminal justice process. In almost all instances, the mean differences detected were between those with the lowest and those with the highest incomes.
Respondents earning less than $12,000 last year thought it much more likely that a person would be prosecuted swiftly for a crime involving a gun than those who earned over $75,000. In addition, the perceptions of those who earned less than $12,000 were distinguished from all others in terms of the likelihood that a prosecutor would enter into a plea agreement. This time, however, poor respondents were more skeptical than others. Finally, the perceptions of those at the bottom of the economic ladder and of those at the top were contrasted with respect to the probability of a long prison sentence. Those who reported earning $40,000 or more thought it much less likely that a convicted offender would receive a long prison sentence. In sum, the poorest respondents had the most “positive” view of the criminal justice process, while the richest respondents in the survey tended to be more skeptical of it.
The power of class position to shape perceptions of criminal justice processes is highlighted even more when respondents’ current work status is examined. Employment status appears to influence perceptions in a fashion similar to income. In comparison to the unemployed, those who are members of the workforce (i.e., employed) thought it: less likely that a person who carries a gun illegally would be caught by authorities; more likely that a prosecutor would plea bargain; and less likely that a convicted gun offender would receive a long prison sentence. As was demonstrated in the analysis of respondent income, the residents at the bottom of the economic ladder—the unemployed—perceive a much greater likelihood of legal sanctioning for the commission of gun crimes than others in the community.
Since the mid-eighteenth century, Western criminal law has been partially based, at least rhetorically, on the notion of deterrence—that it is better to prevent crime than deal with its consequences. Underlying the theory of criminal deterrence is a utilitarian philosophy which, in simple terms, holds that human beings behave in ways that simultaneously maximize benefits and minimize costs. The combination of these two strands—prevention and utilitarianism—results in a theory of crime prevention in which legally prohibited conduct can be forestalled by making its costs higher than its benefits. But, as a sub-structure for this behavioral calculus, the potential offender must first be aware of the sanctions for engaging in legally prohibited behavior and be convinced they are real.
In articulating Anchorage residents’ perceptions of legal sanctions, the data in this article illuminate a critical aspect of criminal deterrence with regard to one category of legally prohibited behavior: gun crime. The data show that the level of perceived deterrence is not uniform; people assign different probabilities to each stage of the criminal justice process. This suggests that while the general concept of deterrence may be fairly straightforward (people will not engage in undesirable conduct if the costs of doing so exceed its benefits), in reality, its operation is extremely complex.
By and large, the evidence presented here suggests that individuals assign probabilities to legal sanctions differentially, according to the stage of the criminal justice process. For example, in general, Anchorage residents do not think it is very likely that a person who carries a gun illegally will be caught by authorities, but once an offender is caught, on average they think it likely that person will be prosecuted.
This study also went beyond general single-variable trends to examine also how people’s perceptions of the likelihood of legal sanctions for gun crime vary as a function of their demographic characteristics. A number of subtleties in perceptions of legal punishments for gun crime emerged through examining six demographic variables (age, gender, race/ethnicity, education, individual income and work status). Each of these variables is associated with individual perceptions for at least two aspects of legal sanctions, with some factors showing a significant association with perceptions in as many as five areas.
Age had its most profound impact on people’s perceptions of the likelihood a prosecutor would allow a gun offender to plead guilty in order to secure a conviction. Gender effects were most prominent with perceptions of the probability that a person caught committing a gun crime would be prosecuted and the chances that a convicted gun offender would receive a lengthy prison sentence. Race/ethnicity was related to perceptions of the likelihood a prosecutor would plea bargain with a gun offender. A person’s education level was shown to play a role in shaping perceptions about the probability of a person who carries a gun illegally being caught as well as the perceived probability of a prosecutor entering into a plea agreement with a gun offender. The socioeconomic factors income and work status were shown to demonstrate a high degree of association with perceptions of a person being caught carrying a gun illegally, the chances of a swift prosecution, a prosecutor plea bargaining, and a convicted gun offender receiving a long prison sentence as punishment.
Beyond showing that distinct demographic factors differentially influence people’s perceptions of the various stages of the criminal process, the ANOVA results show precisely where these differences arise. It’s not simply that “age matters” or “race/ethnicity matters,” but how age and race/ethnicity matter. For example, only particular ages, races/ethnicities, levels of education and income, and categories of workforce participation matter for our understanding of Anchorage residents’ perceptions of plea bargaining.
These analyses also shed light on which aspects of perceived legal sanctions are most susceptible to the influence of demographic factors. Nearly all demographic variables contained significant cross-category differences for perceptions of the likelihood that a person who carries a gun illegally will be caught (gender being the only exception) and the likelihood that a prosecutor will allow a gun offender to plead guilty to a less serious charge (gender being the only exception). For other deterrence measures—such as the likelihood that a person caught committing a gun crime would be prosecuted for that offense, and the likelihood of a person prosecuted for a gun crime being convicted (in either state or federal court)—almost none of the demographic characteristics examined showed significance; that is, perceptions did not vary significantly across levels of these variables.
These results suggest that the notion of generalized deterrence may be an illusion. People’s perceptions about the likelihood of sanctions—their understanding of the “costs” of crime—differ according to the stage of the criminal justice process.
Assuming deterrence theorists are correct when they assert that behavior is guided by perceptions of likely sanctions, these data suggest that deterrence strategies might have to focus on a particular aspect of the criminal process (detection v. prosecution v. punishment), or be directed toward specific target audiences, to be effective. “One-size-fits-all” deterrence strategies, in which a single policy or practice is intended for an entire population (for example, mandatory minimum sentences), may miss the mark.
Brad Myrstol is a research associate with the Justice Center.