Does Crime Cause Punitiveness?

Abstract

Why are Americans so punitive toward criminals? Some theories of punitiveness suggest that exposure to crime makes people more supportive of punitive policies toward criminals. We analyzed national survey data and found that neither support for longer prison sentences for four different crimes nor support for the death penalty had a significant positive association with crime rates, prior victimization, vicarious victimization, higher perceived risk of victimization, or fear of crime. Instead, punitiveness was related to how often people watched local TV news, the percent Republican of the person’s county, and race. Support for harsh treatment of criminals therefore appears to be more a product of race, ideology, and news media presentations of crime than of the reality of crime.

Keywords

punishment punitiveness punitivity

Introduction

Why are Americans so punitive toward criminals? Most Americans favor the death penalty for murderers, mandatory minimum prison sentences for a variety of crimes, three-strikes laws, longer prison sentences in general, treating some juvenile offenders as adults, and thereby making them subject to more severe penalties, and a long list of other highly punitive policies (Costelloe, Chiricos, & Gertz, 2009; Cullen, Fisher, & Applegate, 2000). U.S. policy toward criminals is correspondingly harsh, and supporters of punitive policies commonly justify their positions by asserting that these are the policies favored by most Americans (e.g., DiIulio, 1997). The implied argument is that democracy is functioning as it should, with mass public opinion affecting the positions of the people’s elected representatives and thereby determining, at least in a general way, criminal justice policy.

If the majority’s attitudes are driving harsher policies, why then are Americans’ attitudes so favorable to harsh treatment of criminals, and so much more supportive of punitive policies than the citizens of other developed nations? One common explanation is that, in one sense or another, crime causes punitiveness, so that the higher rates of crime in the United States cause more people to favor harsher punishment. If public support for punitive policies really is driven by crime, policymakers who might otherwise be persuaded to reduce the harshness of the criminal justice system could plausibly reply, “What can I do? Crime is high, and my constituents demand that I deal with it by getting tough on criminals.” This line of defense implies that the public is realistically reacting to the reality of crime, and pragmatically supporting punitive policies that they believe will be effective in reducing crime. Is this in fact an accurate description of the link between crime and supportive for harsh criminal sanctions?

Theory

There are many explanations of punitiveness, pointing to conservative political ideology, attribution of blame for criminal behavior, media influences, racial animus and racial threat, economic threat, religious fundamentalism, education, income, gender, and misperceptions of criminal justice as more lenient than it really is. Our focus is a narrower one. We seek to explore how exposure to crime, broadly construed, affects support for harsher punishment of crime.

One prominent theory of punitivity is what Unnever and Cullen (2010) called the “escalating crime-distrust model” (see also the similar emphasis on “crime salience” in Costelloe et al., 2009; Taylor, Scheppele, & Stinchcombe, 1979). This theory argues that punitivity in the United States and other Western nations grew as perceptions of rising crime and disorder spread, against a background of growing distrust in government and its ability and willingness to protect its citizens from crime (Simon, 2007). This theory’s stress on perceptions of a growing crime threat suggests that there may be a considerable gap between perceptions and reality. Although these perceptions might be partially attributable to a reality of rising crime rates, they can also be encouraged by news about crime and the criminal justice system, or by political propaganda. The volume of news about crime can increase even when crime rates are declining, and exposure to an endless stream of accurate, yet unrepresentative, news accounts of violent crime can create a vision of the world as a dangerous place, even for people who are largely sheltered from the risks of victimization.

This conception of exposure to crime may, however, be unduly narrow. There are many other ways in which crime could influence punitiveness besides a perception of rising crime rates. First, emotional reactions to crime may influence punitiveness independently of beliefs about crime. Thus, the fear of victimization may matter, not just the belief that its risks are rising. Second, even within the category of perceptions or beliefs about crime, perceptions of the current level of crime risks may matter, independent of perceptions that those risks are increasing. Certainly intense, high risks of victimization could motivate people to support short-term crime control strategies that promise quick results, such as “getting tough on criminals,” even if that risk level was not rising past an already intolerably high level. Third, it is possible that support for punitive measures is driven by a belief (accurate or not) that crime in one’s immediate environment is intolerably high, relative to other places, and thus, one is in greater need than other Americans of policies that promised to quickly reduce crime. Similarly, some people may believe, accurately or not, that they are at higher risk of future victimization than others, regardless of whether crime rates in general are rising.

Conversely, a focus solely on subjective reactions to crime might also be unduly restrictive. Some people are more likely than others to have actual personal experience with crime as victims, or to have become “vicarious victims” through the experiences of family members, friends, neighbors, and other significant others (Borg, 1998). Thus, there may be a reality of direct and indirect personal experience with crime underlying their perceptions and feelings of risk. Likewise, some people live in areas where the general area-wide risk of criminal victimization is, to the extent that official crime rates can measure that risk, genuinely higher than average. Thus, for some, there is an objective reality about crime that could motivate support for harsher punishment of criminals, apart from the effects of exaggerated perceptions of future risk and excessive fear stimulated by misleading news coverage and political rhetoric about crime. If so, prior victimization of one’s self or of significant others, and local crime rates could affect get-tough attitudes. The personal experience of victimization might stimulate in some people an intensified support for policies perceived to be effective tools for reducing the risk of future victimization. Alternatively, it could trigger in others a thirst for vengeance or retribution, not only against the specific persons responsible for the individual’s victimization but also against others perceived to be like them and the larger class of criminals in general.

Support for harsher crime control measures could also be influenced by relatively unemotional, intellectual assessments of the risk of future victimization, rather than just emotional reactions to prior victimization. Even in the absence of fear, if one believed that punitive measures are effective in quickly reducing crime, those who anticipated a higher probability of future victimization for themselves could be more supportive of such tactics simply because they perceive themselves to be more in need of measures that would reduce crime. To be sure, these estimations of risk are subject to considerable error, but even inaccurate estimates could influence attitudes toward the treatment of criminals. Thus, perceptions of future crime risk, as well as fear of crime, merit a place in a theory of punitivity.

There are some fairly clear causal chains implied by these separate hypotheses. High or rising crime rates can affect the probability of personal victimization for any given individual, and for people they know. These can in turn affect (along with media influences and political rhetoric) perceptions of the risk of future victimization, of trends in crime rates, and of the risk of crime in one area relative to others. These risk perceptions can in turn directly affect punitiveness, or can affect fear of crime, which then affects punitiveness. It is, however, also possible that levels or trends in crime rates can influence punitiveness in heretofore undiscovered ways beyond their effects on individual and vicarious victimization, risk perceptions, and fear. If so, a more complete test of the impact of crime on punitiveness would have to include crime rates even if the model already included these other factors that are presumed to mediate the effect of crime rates.

To summarize, support for harsher treatment of criminals could be influenced by (a) personal prior victimization, (b) vicariously victimization of others known to the person, (c) a higher perceived individual risk of future victimization, (d) higher crime rates prevailing in one’s area, (e) perceptions of rising crime rates, in a context of distrust about governments’ willingness and ability to reduce crime, (f) the belief that crime rates are relatively higher in one’s area compared with other areas, or (g) greater fear of crime.

Methods of Prior Research

Previous work on the effects of exposure to crime on punitiveness has suffered from a number of limitations, including constricted conceptions of exposure to crime, unduly narrow measures of punitiveness, use of local or nonprobability samples of limited generalizability, and measures of punitiveness of doubtful validity. Exposure to crime has been conceptualized in most previous studies exclusively as personal experience of victimization or fear of crime (e.g., Taylor et al., 1979; Unnever, Cullen, & Fisher, 2007), and in most studies, only one or two of these factors are addressed. This focus neglects the possibility the crime rates can influence an individual’s attitudes toward punishment independent of his or her personal experiences as victims or fear of crime. Certainly, people can be aware of high rates of crime without experiencing it themselves. People may likewise also be affected by the victimization of those they know, that is, by “vicarious victimization” (Borg, 1998). And to the extent that they are influenced by factors pertaining to themselves, it could be cognitive assessments of future risk of victimization that matter, not just prior experiences as victims or the emotion of fear.

Much of the prior work in this area has also been unduly narrow in the kinds of punitive attitudes that are measured. Most studies have focused on just two attitudes: (a) support for the death penalty for murder (e.g., Baumer, Messner, & Rosenfeld, 2003), and (b) the belief that local courts are too lenient in their sentencing (e.g., Taylor et al., 1979). Attitudes toward the death penalty may be influenced by attributes that are unique to that particular penalty, such as its irreversibility, so that findings pertaining to capital punishment may not be generalizable to other punishments. In any case, only a tiny fraction of sentences for crime are death sentences; serious offenses are far more likely to be punished by prison sentences.

Regarding sentencing in general, some studies have used a General Social Surveys (GSS) question inquiring whether the respondent considers local courts to be too lenient or too harsh, and interpreted endorsement of the “too lenient” response as indicative of punitiveness. The chief problem with this measure is that it implicitly assumes that respondents know something about local sentencing practices, because the respondent’s preferences for harsher sentencing are expressed solely in terms of a comparison with those perceived sentencing practices. Given the average person’s ignorance of either sentencing practices or criminal justice realities in general, this assumption is implausible. Consequently, a “too lenient” response could reflect an underestimation of local court harshness rather than a preference for harsher penalties (Roberts, 1992).

Measures of fear of crime and perceived risk of future victimization have been typically limited to single items. For example, fear is commonly measured using the GSS item asking whether there is an area near the respondent where they would be afraid to walk alone at night. This measure is problematic because of its hypothetical character—although there might well be places where the respondent would be fearful if he or she walked there at night, if they do not do so, they may actually have very little fear of crime. Using measures that directly ask people to rate their fear levels may be more appropriate.

Measures of fear and perceived risk are often limited to questions pertaining to a single crime, although these perceptions may be very crime-specific, varying considerably from one crime type to another. Using multiple items to measure fear and perceived risk with regard to multiple offense types would seem to be preferable.

Findings of Prior Research

The findings of past studies have been very inconsistent regarding the impact of most crime-related variables on punitiveness. First, there is little research on the impact of perceived risk of future victimization, and it has yielded inconsistent results. Pickett and Chiricos (2012) found a significant positive association, but it applied only to attitudes toward the punishment of juveniles. However, King and Maruna (2009) and Pickett and Baker (2014) found no significant association between perceived risk and general punitive attitudes. The results of the latter two studies, however, have limited generalizability: King and Maruna studied only residents of six nonrandomly selected local areas in England, whereas Pickett and Baker based their research on a nonprobability volunteer sample.

In contrast, there is a wealth of research bearing on the effect of fear of crime on punitive attitudes, but the findings are highly inconsistent. Of 18 studies we located, nine obtained results supporting a positive effect of fear on punitiveness (Costelloe et al., 2009; Hogan, Chiricos, & Gertz, 2005; Johnson, 2006; Rankin, 1979; Sprott & Doob, 1997; Thomas & Foster, 1975; Tufts & Roberts, 2002; Unnever et al., 2007; Wu, Sun, & Wu, 2011) and nine found no support (Chiricos, Welch, & Gertz, 2004; King & Wheelock, 2007; Leverentz, 2011; Messner, Baumer, & Rosenfeld, 2006; Ouimet & Coyle, 1991; Stack, 2003; Taylor et al., 1979; Tyler & Weber, 1982; Welch, Payne, Chiricos, & Gertz, 2011). There is no clear methodological superiority of either of these two conflicting bodies of research, so it remains very much an unsettled matter whether fear of crime motivates people to favor harsher treatment of criminals.

There has been far more agreement regarding the influence of individuals’ personal victimization experiences—researchers generally find no impact. Of 19 studies we located, 16 found no significant positive association of prior victimization with punitive attitudes (Borg, 1998; Costelloe et al., 2009; Cullen, Clark, Cullen, & Mathers, 1985; Hartnagel & Templeton, 2012; Johnson, 2006; King & Maruna, 2009; Leverentz, 2011; Ouimet & Coyle, 1991; Pickett & Baker, 2014; Pickett & Chiricos, 2012; Seltzer & McCormick, 1987; Sprott & Dobb, 1997; Stack, 2003; Taylor et al., 1979; Tufts & Roberts, 2002; Tyler & Weber, 1982; Unnever et al., 2007). Only three studies found significant positive associations (Hanslmaier, 2013; Rich & Sampson, 1990 [for two of three offense types]; Wu et al., 2011). The Rich and Sampson findings were based on a local sample of Chicago-area residents, the Wu et al. study addressed only students at four nonrandomly selected universities (three of them in China), and the Hanslmaier study was based on a national sample of German residents. Thus, the supportive results of these three studies may be reflective of effects unique to these samples, but not applicable to the general U.S. population.

Research on the effect of crime rates on individual support for harsh punishment of criminals is considerably more modest, and generally quite narrow in its focus. Three of the studies address only the effect of one crime type, homicide, on support for one punitive policy, capital punishment. Messner et al. (2006) found no significant association, whereas Baumer, Messner, and Rosenfeld (2003) found a significant positive association and Soss, Langbein, and Metelko (2003) observed the same association among Whites. Four studies analyzed more general measures of punitiveness, and all found no significant association with crime rates, although each of them was significantly limited in some way. Ousey and Unnever (2012) found no effect of national violent crime rates on individuals’ support for more severe punishment of criminals in 26 European Union nations. People may, however, only be aware of local crime rates, which might affect their attitudes toward punishment. The same problem affects the null findings of Welch et al. (2011), who measured only state-level property and violent crime rates. King and Wheelock (2007) measured county crime rates, but they assessed only the effects of the homicide rate, finding no significant association with punitiveness. Finally, Pickett and Chiricos (2012) also used county crime data, but only for the single combined Index crime rate. They found no association between punitiveness and this combined measured, but use of this measure could have concealed positive effects of rarer crimes such as homicide or rape that were counterbalanced by null or negative effects of more common crimes such as larceny.

Finally, there is some evidence that attitudes favoring harsher punishment of criminals are affected by the perception—accurate or not—that crime is increasing. Sometimes, what researchers vaguely describe as “crime salience” or “concern” about crime is actually measured by survey items that tap into the perception of rising crime (e.g., two of four items in Leverentz’ measure of crime salience measured perceptions of increasing crime, and two of five items addressing “salience of crime” in Cullen et al., 1985, likewise measured perceptions of increased crime). Other scholars have specifically measured perceived crime trends (Hogan et al., 2005; Pickett & Baker, 2014; Thomas & Foster, 1975; Unnever & Cullen, 2010). All but one of these six studies (Cullen et al., 1985) found punitiveness to be positively and significantly related to the perception that crime is increasing. The supportive studies, however, are characterized by the use of nonprobability (Pickett & Baker, 2014) or local samples (Hogan et al., 2005; Leverentz, 2011; Thomas & Foster, 1975), or by a dubious measure of punitiveness. Unnever and Cullen (2010) studied a national probability sample, but measured “punitiveness” by asking forced-choice questions in which survey respondents were asked to choose between (a) punitive policies for reducing crime or (b) policies that “address the social problems that cause crime like bad schools, poverty, and joblessness.” A high score of this punitiveness index in this study therefore may have reflected only a lack of support for the nonpunitive policies rather than support for punitive policies. As a result, there is no evidence on the impact of perceptions of increasing crime on punitiveness that can be reliably generalized to any national population.

In sum, prior work has produced highly conflicting results concerning the impact of fear of crime on punitive attitudes, little generalizable research on the impact of crime rates on punitive attitudes in general (as opposed to just those pertaining to the death penalty or punishment of juveniles) or of the effects of perceptions that crime is increasing, and only a handful of conflicting studies of the effect of perceived risk of future victimization. A large body of research overwhelmingly indicates that prior victimization is unrelated to punitive attitudes.

There nevertheless is at least some research supporting the hypotheses that crime is affected by any of five distinct crime-related factors: crime rates, perceived risk of victimization, fear of crime, prior victimization, and a perception of increasing crime. All of these variables are conceptually distinct, they all could directly affect punitive attitudes independent of effects mediated through the other variables, and they are positively correlated with each other (Hogan et al., 2005; Johnson, 2006). Thus, the effect on punitiveness of any one of them could be confounded with the effects of the others if the others are not statistically controlled. Yet, none of these studies has controlled for all five types of variables, nor have any studies controlled for even four of them. Even the best studies in this regard simultaneously estimated the effects of just three of these five types of crime-related variables (Leverentz, 2011; Pickett & Baker, 2014; Pickett & Chiricos, 2012; Welch et al., 2011). As a consequence, all extant estimates of the effects of each of these factors are probably distorted by omitted variables bias.

Our study contributes to the literature on the effects of crime on support for harsher punishment of criminals in the following ways:

We measure and simultaneously estimate the effects of all five of the aforementioned types of crime-related variables, as well as two others: vicarious victimization (Borg, 1998) and the perceived level of crime in one’s own area compared with that of other areas.

We investigate the possible effects of understudied aspects of exposure to crime such as perceived risk of future victimization.

We use improved multi-item measures of the more commonly studied variables such as prior victimization and fear of crime.

We use a nationally representative probability sample of the U.S. urban population, providing a formal basis for generalizing its findings to the population most exposed to crime.

We directly measure a wide array of sentencing preferences regarding prison sentence lengths for multiple offenses, as well as support or opposition to the death penalty.

Finally, we carry out crime-specific analyses of the potential effects of crime rates, prior victimization, fear, and perceived risk on preferred punishment severity for the corresponding crime types, testing the proposition that it is exposure to specific types of crime that influences attitudes toward how harshly those same crimes should be punished.

Hypotheses

Based on the aforementioned theoretical considerations and the lacunae in prior empirical research, we seek to test the following specific hypotheses concerning the impact of exposure to crime on punitive attitudes:

Hypothesis 1: Higher crime rates in the area where one resides increase support for more punitive treatment of criminals.

Hypothesis 2: A perception that crime rates are increasing causes greater support for more punitive treatment of criminals.

Hypothesis 3: A perception that crime rates are relatively higher in the area where one resides increases support for more punitive treatment of criminals.

Hypothesis 4: Personal experience of crime victimization increases support for more punitive treatment of criminals.

Hypothesis 5: Vicarious victimization (knowing a person who has been a victim of serious crime) increases support for more punitive treatment of criminals.

Hypothesis 6: Perceiving a higher personal risk of future victimization increases support for more punitive treatment of criminals.

Hypothesis 7: Greater fear of crime increases support for more punitive treatment of criminals.

Method

The Sample

Our individual-level data were obtained in a national telephone survey of a probability sample of 1,500 adults (age 18 and above) living in a probability sample of 54 large urban counties. The sample was selected using random digit dialing procedures. This sampling method allowed access to the 95% of the U.S. households with a telephone, including those with unlisted numbers. The sample was drawn exclusively from the 54 largest counties participating in the National Judicial Reporting Program (NJRP). These 54 counties were selected by NJRP staff to be representative of the 75 largest (by population) counties in the United States. In 1998, these 75 counties accounted for 50.2% of the nation’s murders, 61.9% of robberies, and 51.4% of all violent crimes known to the police (analysis of U.S. Department of Justice, Federal Bureau of Investigation, 2000). Consequently, our results can be generalized to the noninstitutionalized adult residents of the 75 largest urban counties that account for most of the nation’s violent crime.

Telephone numbers were randomly generated for each county, using the area code and residential prefixes operative in each county, with sample sizes for each county proportional to the population size of the county. Within each household contacted, an adult R was randomly selected by the interviewer, by asking to speak to the resident age 18 or older who had most recently celebrated a birthday. Rs were weighted by the inverse of the number of different telephone numbers in their household. Telephone interviews were conducted byResearch Network, a professional polling firm that has conducted hundreds of telephone surveys, including many concerning crime.

The interviews were conducted in April and May of 1998. Although levels of punitivity and its potential determinants have changed since 1998, we are not aware of any evidence or sound rationale for believing that the relationships between these variables or the underlying causal structure of punitivity has changed since then. In any case, no more recent data set has nearly as rich a set of measures of exposure to crime as this one.

Measurement of the Variables

Dependent variables—Measures of punitiveness

We measured the dependent variable, support for punitive policies toward criminals, in six ways: (1a) a binary measure of support for the death penalty as a punishment for murder, (b) the length of preferred prison sentences for convicted murderers, (c) the length of preferred prison sentences for convicted robbers, (d) the length of preferred prison sentences for persons convicted of aggravated assault, (e) the length of preferred prison sentences for convicted burglars, and (f) an index combining the four preferred-sentence measures. We asked Rs how long a sentence they preferred for persons convicted of each of these four crimes. Sentences have been converted from years into number of months when necessary. When a death sentence or life sentence was the preferred sentence for murder, it was treated as 540 months (45 years)—the average remaining life expectancy for a male of the median age of murder arrestees. The exact wording of the relevant survey questions can be found in the appendix.

Principal independent variables—Measures of exposure to crime

The principal independent variables of interest were measures of exposure to crime. We measured prior victimization by asking questions about (a) having been a victim of robbery in the past year, (b) having been a victim of burglary in the past year, (c) having been a victim of an assault since becoming an adult, and (d) knowing a person who has been a victim of a serious crime in the past year (vicarious victimization). We measured perceived risk of future victimization by asking questions about (e) the perceived likelihood that one will be murdered in the next 12 months, (f) the perceived likelihood that one will be robbed or mugged in the next 12 months, and (g) the perceived likelihood that one will be burglarized in the next 12 months. We measured crime rates in areas in which Rs resided using official police-based data on (h) the local (county) murder rate, (i) the local robbery rate, (j) the local aggravated assault rate, and (k) the local burglary rate. We measured fear of crime by asking the R to rate (l) how much (on a 0-10 scale) they feared being murdered, (m) how much they feared being robbed or mugged, and (n) how much they feared having their home broken into. Finally, we asked questions regarding the Rs (o) perception that crime in their county was increasing, and their (p) perception that crime in their county was higher/lower than the national average.

Control variables

We controlled for many attributes of both individuals and their environments that theory and prior research suggest could affect support for punitive crime control policies, and that might also be correlated with measures of exposure to crime. Specifically, we controlled for the individual attributes of age, sex, race, income, education, marital status, the number of times per week the person watches local television news, the number of times per week the person watches national television news, whether the person had ever been arrested, and whether the person or a member of his or her family had ever worked in the criminal justice system. Race is a particularly important control given prior evidence that Whites are more punitive, and that racial animus toward Blacks lies behind much of this punitiveness (Chiricos et al., 2004; Unnever & Cullen, 2010).

We also recognized that punitiveness might be affected by perceptions of attributes of one’s environment that are popularly associated, correctly or not, with crime. In particular, the racial threat explanation of punitiveness asserts that people (certainly White people, and possibly members of racial minorities as well) will feel more threatened if they believe that a large share of the population in their area are Black (Pickett, Chiricos, Golden, & Gertz, 2012). More broadly, crime threat theory suggests that people will feel more threatened if they live where a large share of the population belongs to other groups thought to commit crimes at a higher rate than average, such as the poor, unemployed people, the uneducated, urban dwellers, and so on (Chiricos, McEntire, & Gertz, 2001). Greater numbers of unemployed persons and other “threatening populations” could increase public anxiety over disorder, leading to increased support for more punishment of criminals as a way of alleviating this anxiety, and of protecting themselves from this perceived danger. We therefore controlled for the following county-level measures of threatening populations, as of 1990: percent Black, the percent of families under the poverty line, and percent of the civilian labor force unemployed. We also controlled attributes of areas that might induce some to perceive social disorder that needed to be addressed: divorces per 1,000 marriages, percent with a college degree, and population size (U.S. Bureau of the Census, 1994).

Finally, prior multi-level analysis has indicated that punitive attitudes are increased by the prevalence of conservative political views in their area (Baumer, Messner, & Rosenfeld, 2003), so we controlled for the percent who voted Republican in the 1996 presidential election as a measure of the county population’s political conservatism (the individual’s political ideology was not measured in the survey), and for whether the county was located in the South (1 if in a state of the Confederacy, 0 otherwise).

Estimation procedures

We estimated models of support for the death penalty using logistic regression, as this was a binary dependent variable, and estimated models of the other measures of punitiveness using ordinary least squares (OLS) regression, as they were continuous variables. Because there were considerable missing data on the income variable, we used multiple imputation methods to impute missing data, to avoid the sample bias and loss of statistical power that could result from list wise deletion of cases with missing data. Results concerning effects of exposure to crime, however, were generally substantively identical when list wise deletion was used. The STATA Version 13 statistical package, using its mi command for multiple imputation of missing values, was used to compute estimates. In addition to the estimates shown here, we also estimated standard errors using the STATA cluster command to take account of the clustering of respondents by county. Some estimates were higher, some were lower, but all signs and significance of all measures of exposure to crime were unchanged (estimates available from senior author).

Results

Table 1 lists the full set of variables used in our analyses, along with their sample means and standard deviations. Table 2 displays the estimates that test our hypotheses.¹ Histograms in the appendix display the distributions of the dependent variables, that is, the measures of punitiveness.

Table 1.

Descriptive Statistics.

Variable	Description	M	SD	Original N
Supports the death penalty	R supports the death penalty for murder	0.74	0.43	1,257
Preferred sentence, murder	Preferred months, homicide	475.64	134.60	1,359
Preferred sentence, robbery	Preferred months, robbery	146.91	128.66	1,302
Preferred sentence for assault	Preferred months, assault	141.73	133.59	1,320
Preferred sentence, burglary	Preferred months, burglary	106.07	108.83	1,307
Average level of punitiveness	Standardized index of average preferred sentence	0.00	0.76	1,401
Murder rate	Murders per 100,000	11.36	7.79	1,501
Robbery rate	Robberies per 100,000	372.09	229.23	1,501
Assault rate	Aggravated assaults/100,000	522.61	256.14	1,501
Burglary rate	Burglaries/100,000	1,073.83	392.27	1,501
Personal burglary victimization	Burglarized within the past year (2 = yes, 1 = no)	1.07	0.25	1,491
Personal robbery victimization	Robbed within the past year (2 = yes, 1 = no)	1.04	0.20	1,493
Personal assault victimization	Assaulted as adult (2 = yes, 1 = no)	1.25	0.43	1,487
Perception of increasing crime	Perception of increasing crime (1 = decreasing, 2 = same, 3 = increasing)	2.15	0.77	1,424
Perceived area is high-crime area	Perception that crime is higher in own area	2.94	1.11	1,428
Vicarious victimization	Personally knows a victim of a serious crime	0.36	0.48	1,484
Perceived risk of murder	Perceived likelihood of being murdered	1.50	2.23	1,424
Perceived risk of robbery	Perceived likelihood of being robbed	2.49	2.61	1,437
Perceived risk of burglary	Perceived likelihood of being burglarized	2.72	2.73	1,437
Fear of being murdered	Fear of being murdered (0-10)	2.77	2.90	1,470
Fear of being robbed	Fear of being robbed (0-10)	3.97	2.96	1,473
Fear of being burglarized	Fear of being burglarize (0-10)	4.16	3.11	1,475
Frequency watch local news	Times per week R watches local TV news	5.93	4.06	1,482
Frequency watch national news	Times per week R watches national TV news	4.54	3.46	1,469
Age	Age in years	44.12	17.10	1,451
Sex	Sex (Male = 1)	0.46	0.50	1,469
Race	Race (Black = 1)	0.13	0.34	1,501
Level of education	Education (7-point scale)	4.76	1.60	1,475
Income	Income (5-point scale)	3.19	1.28	1,132
Marital status	Married = 1, other status = 0	0.52	0.50	1,473
Ever arrested	R has been arrested for non-traffic offense	0.13	0.33	1,453
Family member works for CJS	Family member works in CJS	0.22	0.42	1,475
% below poverty line in county	% county population below poverty line, 1990	13.27	4.63	1,501
% Black in county	% county population Black, 1990	16.66	12.60	1,501
% Republican in county	% 1996 presidential votes for Republican	35.8	10.24	1,501
County divorce rate	Divorces/1,000 married population, 1990	18.94	7.33	1,501
County population (1998)	1998 County Population (1,000s)	2,754.44	2,685.74	1,501
% college grads in county	% college graduates, 1990	28.11	6.76	1,501
County located in the south	County is in state of confederacy	0.25	0.43	1,501

Table 2.

The Effect of Exposure to Crime on Punitiveness^*.

Variable	Supports the Death Penalty				Preferred Sentence for Murder
Variable	b	OR	t	p	b	t	p
Murder Rate	4.8E-4	1.00	.03	.98	.96	.97	.33
Robbery Rate	-1.3E-3	1.00	-1.39	.16	.04	.77	.44
Assault Rate	8.3E-5	1.00	.14	.89	-.01	-.35	.73
Burglary Rate	-1.7E-4	1.00	-.50	.62	-.04	-1.81	.07
Personal Burglary Victimization	.08	1.08	. 26	.80	5.03	.29	.77
Personal Robbery Victimization	.13	1.14	.32	.75	-40.15	-1.63	.10
Personal Assault Victimization	2.4E-3	1.00	.01	.99	11.66	1.21	.23
Perception of Increasing Crime	.18	1.00	1.86	.06	7.34	1.31	.19
Perceive Area is High Crime Area	4.2E-3	1.00	.06	.95	2.33	.62	.54
Vicarious Victimization	-.07	.93	-.47	.64	8.78	1.02	.31
Perceived Risk of Murder	-.07	.94	-1.27	.21	.19	.06	.95
Perceived Risk of Robbery	.06	1.06	1.06	.29	-1.00	-.32	.75
Perceived Risk of Burglary	-.03	.97	-.69	.49	.11	.04	.97
Fear of Being Murdered	.01	.97	.26	.80	-3.45	-1.55	.12
Fear of Being Robbed	.06	1.07	1.46	.14	3.73	1.58	.11
Fear of Being Burglarized	-.04	.96	-1.25	.21	-3.06	-1.50	.14
Frequently Watch Local News	.05	1.05	1.52	.13	2.42	2.24	.03
Frequently Watch National News	-.02	.98	-.83	.41	-1.78	-1.31	.19
Age	-2.1E-3	1.00	-.42	.67	-.15	-.55	.59
Sex (Male = 1)	.23	1.26	1.54	.12	-9.88	-1.19	.24
Race (Black = 1)	-.66	.52	-3.41	.00	-37.05	-2.68	.01
Education	-.19	.83	-3.46	.00	5.55	1.93	.05
Income	.18	1.20	2.36	.02	-.62	-.15	.88
Marital Status	.01	1.01	.05	.96	10.52	1.23	.22
Ever Arrested	-.17	.84	-.76	.45	.02	.00	.99
Family Member Works for CJS	.06	1.06	.37	.71	.35	.04	.97
% Below Poverty Line in County	3.7E-3	1.00	.10	.92	-2.20	-1.05	.30
% Black in County	.03	1.03	2.61	.01	-.32	-.44	.66
% Republican in County	.03	1.03	2.45	.01	.34	.50	.62
County Adjusted Divorce Rate	3.0E-3	1.00	.25	.81	.59	.78	.43
County Population (1998)	4.1E-5	1.00	.88	.38	-1.4E-3	-.53	.60
% College Grads in County	.01	1.01	.36	.72	-1.14	-1.12	.26
County Located in the South	-.32	.73	-1.29	.20	15.25	1.16	.25
Constant	-.37	.69	-.32	.75	512.84	7.56	.00
Adjusted R²	N/A				.02
N	1501				1501
Variable	Preferred Sentence for Robbery				Preferred Sentence for Assault
Variable	b		t	P	b	t	p
Murder Rate	.48		.53	.60	2.05	2.32	.02
Robbery Rate	.06		1.14	.26	.08	1.35	.18
Assault Rate	-.06		-1.68	.09	-.05	-1.40	.16
Burglary Rate	.02		.89	.37	-.03	-1.82	.07
Personal Burglary Victimization	25.23		1.63	.10	11.24	.72	.47
Personal Robbery Victimization	-51.95		-3.72	.00	-35.20	-2.10	.04
Personal Assault Victimization	7.92		.93	.35	4.37	.46	.65
Perception of Increasing Crime	-3.39		-.62	.53	-5.87	-1.09	.28
Perceive Area is High Crime Area	1.20		.32	.75	2.75	.73	.47
Vicarious Victimization	1.30		.16	.87	17.89	2.15	.03
Perceived Risk of Murder	5.33		2.15	.03	3.26	1.28	.20
Perceived Risk of Robbery	-1.65		-.63	.53	-1.49	-.56	.58
Perceived Risk of Burglary	-1.73		-.68	.50	-2.20	-.86	.39
Fear of Being Murdered	-3.04		-1.62	.11	-.67	-.36	.72
Fear of Being Robbed	1.26		.53	.59	.13	.06	.95
Fear of Being Burglarized	.07		.03	.97	.53	.28	.78
Frequently Watch Local News	2.94		2.01	.04	3.20	2.13	.03
Frequently Watch National News	-1.25		-.88	.38	-1.85	-1.24	.22
Age	.04		.16	.88	-.06	-.22	.83
Sex (Male = 1)	12.49		1.61	.11	-11.73	-1.47	.14
Race (Black = 1)	-16.12		-1.57	.12	-26.29	-2.44	.02
Education	-.26		-.10	.92	3.02	1.06	.29
Income	7.29		1.97	.05	6.58	1.55	.12
Marital Status	12.72		1.54	.12	6.25	.71	.48
Ever Arrested	3.57		.29	.77	-20.98	-1.78	.08
Family Member Works for CJS	9.12		1.00	.32	-5.66	-.60	.55
% Below Poverty Line in County	-1.56		-.78	.44	-2.35	-1.18	.24
% Black in County	.81		1.08	.28	-.09	-.12	.90
% Republican in County	1.70		2.86	.00	1.60	2.56	.01
County Adjusted Divorce Rate	.23		.34	.74	.86	1.23	.22
County Population (1998)	3.3E-3		1.26	.21	8.2E-4	.32	.75
% College Grads in County	1.12		1.14	.25	.47	.51	.61
County Located in the South	6.63		.55	.58	11.43	.51	.61
Constant	3.64		.06	.95	58.70	.96	.34
Adjusted R²	.03				.03
N	1501				1501
Variable	Preferred Sentence for Burglary				Average Level of Punitiveness
Variable	b		t	p	b	t	p
Murder Rate	1.00		1.34	.18	.01	1.75	.08
Robbery Rate	.05		.97	.33	4.7E-4	1.41	.16
Assault Rate	-.03		-1.09	.28	-2.9E-4	-1.55	.12
Burglary Rate	-9.50E-4		-.06	.95	-1.2E-4	-1.11	.27
Personal Burglary Victimization	8.94		.67	.50	.08	.85	.39
Personal Robbery Victimization	-21.69		-1.64	.10	-.30	-3.44	.00
Personal Assault Victimization	5.95		.74	.49	.06	1.24	.22
Perception of Increasing Crime	.80		.18	.86	4.3E-3	.14	.89
Perceive Area is High Crime Area	.70		.21	.83	.02	.77	.44
Vicarious Victimization	4.79		.68	.49	.07	1.51	.13
Perceived Risk of Murder	1.72		.82	.41	.02	1.47	.14
Perceived Risk of Robbery	-2.06		-.91	.36	-.01	-.76	.45
Perceived Risk of Burglary	-2.05		-.95	.34	-.01	-.79	.43
Fear of Being Murdered	1.40		.86	.39	-.01	-.98	.33
Fear of Being Robbed	-.26		-.14	.89	3.6E-3	.29	.77
Fear of Being Burglarized	-.67		-.40	.69	5.3E-3	.46	.64
Frequently Watch Local News	2.65		2.15	.03	.02	2.65	.01
Frequently Watch National News	-.74		-.55	.58	-.01	-1.12	.26
Age	.13		.62	.54	7.2E-5	.05	.96
Sex (Male = 1)	5.05		.74	.46	-.01	-.27	.79
Race (Black = 1)	-12.41		-1.25	.21	-.19	-3.05	.00
Education	-.45		.20	.85	.01	.95	.34
Income	1.63		.47	.64	.03	1.20	.23
Marital Status	6.95		.94	.35	.06	1.28	.20
Ever Arrested	-7.38		-.66	.51	-.05	-.75	.45
Family Member Works for CJS	6.17		.78	.44	.03	.61	.54
% Below Poverty Line in County	-2.13		-1.20	.23	-.02	-1.28	.20
% Black in County	.58		.85	.39	1.2E-3	.28	.78
% Republican in County	1.08		2.22	.03	.01	2.68	.01
County Adjusted Divorce Rate	.74		1.16	.25	4.6E-3	1.05	.29
County Population (1998)	1.6E-3		.71	.48	6.0E-6	.40	.69
% College Grads in County	.47		.54	.59	1.0E-3	.17	.87
County Located in the South	1.00		.09	.93	.08	1.18	.24
Constant	19.69		.35	.73	-.48	-1.33	.18
Adjusted R²	.01				.03
N	1501				1501

Note. *b = unstandardized OLS/Logistic regression coefficient; OR = odds ratio; t = ratio of coefficient over its standard error; p = 2-tailed significance.

Put simply, exposure to crime, regardless of how it was measured, showed no significant association with support for punitive measures, no matter how they were measured. With 16 measures of exposure to crime and six different measures of punitiveness, there were a total of 96 tests of the general hypothesis that exposure to crime increases support for punitiveness toward criminals. Table 3 summarizes the results of all of the tests of the hypothesis. Of the tests, only three yielded statistically significant (at the .05 significance level, two-tailed) positive associations consistent with the hypothesis, one yielded a significant negative association, contrary to the hypothesis, and 92 (96%) were not significantly different from zero. Because one would expect about five of 96 results to be significant purely on the basis of random chance, even if there was no actual effect of exposure to crime on punitiveness, the three seemingly supportive findings are likely to be chance findings. This interpretation is supported by the fact that two of these three findings were nonsensical, seemingly indicating that (a) a higher murder rate causes support for more severe punishment of aggravated assault, but not of murder, and that (b) a higher perceived risk of being murdered causes support for more severe punishment of robbery but not of murder. We hypothesized that crime rates could affect punitiveness indirectly, via individual victimization, perceived risk of future victimization, or fear of crime, but because these variables have no direct effects of their own, they cannot mediate any indirect effects of crime rates.

Table 3.

The Effects of Exposure to Crime on Punitiveness: Summary of Findings.

	Preferred sentence length for				Favors
	Homicide	Robbery	Assault	Burglary	Death penalty
Crime rates
Murder	ns	ns	+, p = .07	ns	+, p = .07
Robbery	ns	−, p < .01	ns	ns	ns
Assault	ns	ns	ns	ns	ns
Burglary	−, p = .07	ns	−, p = .10	ns	ns
Victimization
Robbery	ns	−, p < .01	−, p = .04	−, p = .09	ns
Assault	ns	ns	ns	ns	ns
Burglary	ns	ns	ns	ns	ns
Know victim	ns	ns	+, p = .04	ns	ns
Perceived risk
Murder	ns	+, p = .08	ns	ns	ns
Robbery	ns	ns	ns	ns	ns
Burglary	ns	ns	ns	ns	ns
Rising crime	ns	ns	ns	ns	+, p = .05
Above average	ns	ns	ns	ns	ns
Fear of crime
Murder	ns	−, p = 07	ns	ns	ns
Robbery	ns	ns	ns	ns	ns
Burglary	ns	ns	ns	ns	ns

Note. Findings in bold are findings in the predicted positive direction and significant at the .05 level, two-tailed. p = two-tailed significance of coefficient; ns = two-tailed significance > .10.

What, then, does affect punitiveness? Our findings indicate that people are more punitive if they are White, live in politically conservative areas, and view more hours of local television news programs. Rather than being a response to real crime levels or actual exposure to crime, punitiveness is more likely to be a response to exposure to news media coverage of crime and a politically conservative context.

In particular, watching more hours of local television news appears to increase viewers’ support for harsh treatment of criminals. The fact that local news but not national news has this apparent effect is consistent with prior research indicating that local television news is especially focused on crime and likely to project an image of the world as a dangerous place (Chiricos, Padgett, & Gertz, 2000). We did not, however, find that fear mediates the effect of local TV viewing, so it remains for future research to explore how this effect works.

People are significantly more likely to support the death penalty for murder, though not longer prison sentences, if they live in places where a larger share of the residents are Black. Because we controlled for crime rates, this finding provides partial support for the racial threat hypothesis that larger Black populations per se, independent of actual crime levels, represent a perceived threat that increases support for harsher measures to control crime (Chiricos, Hogan, & Gertz, 1997). Furthermore, as did previous researchers, we consistently found that Whites were more supportive of harsh punishment of criminals than African Americans.

Conclusion

It is impossible to prove a negative, but it is fair to say that our findings overwhelmingly fail to support the broad hypothesis that exposure to crime causes punitiveness. Higher crime rates do not cause increased support for harsher punishment of criminals, nor does personal experience as a crime victim, vicarious victimization through the experiences of others, higher perceived risk of becoming a victim in the future, a perception of rising crime rates, a perception of a local crime problem that is worse than average, or greater fear of crime.

In sum, support for punitive measures does not appear to be grounded in the reality of higher crime rates, prior victimization, or vicarious victimization, or even subjective reactions to these realities such as increased fear of crime or perceived risks of future victimization. Instead, it appears to be affected by factors with much weaker connections to the reality of crime. One variable that was consistently related to a variety of measures of punitiveness was the number of hours of local television news watched per week. Prior research indicates that viewing local TV news encourages fear of crime, quite apart from actual crime rates or personal contact with crime, and this effect is stronger than is the viewing of national news (Chiricos et al., 2000). Our findings, however, indicate that fear does not affect punitiveness; rather, viewing of local TV news somehow increases punitiveness independent of fear, local crime rates, victimization risk, and prior victimization. Perhaps repeated exposure to crime on television inspires generalized feelings of anger toward offenders rather than (or in addition to) fear of victimization.

Past research indicates that punitiveness can be generated by racial animus, especially among Whites willing to typify criminals as Black (Chiricos et al., 2004; Unnever & Cullen, 2010). Although we did not have a measure of racial typification, our findings did show that greater punitiveness prevailed among Whites, and among persons living in places with a larger Black share of the population, independent of local crime rates. Considerable prior research supports the view that, among Whites who engage in racial typification, harsh treatment of racial minorities is justified by the need to control crime (Chiricos & Eschholz, 2002; Chiricos et al., 2004; Costelloe et al., 2009; Unnever & Cullen, 2010). Because punitiveness is not a response to the reality of crime, however, this appears to be more of a rationalization than an accurate explanation of support for harsh treatment of criminals.

Because our individual-level results indicate that punitiveness is not a response to crime, this suggests that at the societal level, the rise of punitive criminal justice policies in the 1980s and 1990s likewise was not likely to have been due to a popular response to higher or rising crime rates, increased fear of crime, or greater perceived risk of victimization, even if those factors did increase. Therefore, an alternative explanation for this development is needed. It has been argued instead that conservative policy elites of that period who favored harsher treatment of criminals for their own ideological reasons deployed rhetoric to shift public opinion in a more punitive direction. Thus, elite opinion may have driven mass public opinion, rather than the reverse. If so, it was the “tough-on-crime” rhetoric wielded by conservative elites, more than the reality of crime, that drove increased public concern about crime (Beckett, 1997, pp. 28-43).

Footnotes

Appendix

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Gary Kleck is the David J. Bordua professor of criminology and criminal justice at Florida State University. He received his doctorate in sociology from the University of Illinois in 1979, where he received the University of Illinois Foundation Fellowship in Sociology. He has been at Florida State since 1978. His research has focused on the topics of the impact of firearms and gun control on violence, deterrence, crime control, and violence. He is the author of Point Blank: Guns and Violence in America, which won the 1993 Michael J. Hindelang Award of the American Society of Criminology, awarded to the book of the previous several years that “made the most outstanding contribution to criminology.” He also wrote Targeting Guns (1997) and, with Don B. Kates, Jr., The Great American Gun Debate (1997) and Armed (2001). His articles have been published in the American Sociological Review, American Journal of Sociology, Social Forces, Social Problems, Criminology, Journal of Criminal Law and Criminology, Law & Society Review, Journal of Research in Crime and Delinquency, Journal of Quantitative Criminology, Crime and Delinquency, UCLA Law Review, the Journal of the American Medical Association, and many other journals.

Dylan Baker Jackson is an assistant professor at the University of Texas, San Antonio. His research explores the biosocial underpinnings of antisocial behaviors. His recent work has examined the influence of neuropsychological deficits on low self-control and misconduct as well as the impact of children’s antisocial traits and behaviors on measures of family risk.

References

Baumer

Messner

S. F.

Rosenfeld

(2003). Explaining spatial variation in support for capital punishment: A multilevel analysis. American Journal of Sociology, 108, 844-875.

Beckett

(1997). Making crime pay: Law and order in contemporary American Politics. New York, NY: Oxford University Press.

Borg

M. J.

(1998). Vicarious homicide victimization and support for capital punishment: A test of Black’s theory of law. Criminology, 36, 537-567.

Chiricos

Eschholz

(2002). The racial and ethnic typification of crime and the criminal typification of race and ethnicity in local television news. Journal of Research in Crime & Delinquency, 39, 400-420.

Chiricos

Hogan

Gertz

(1997). Racial composition of neighborhood and fear of crime. Criminology, 35, 107-131.

Chiricos

McEntire

Gertz

(2001). Perceived racial and ethnic composition of neighborhood and perceived risk of crime. Social Problems, 48, 322-340.

Chiricos

Padgett

Gertz

(2000). Fear, TV news and the reality of crime. Criminology, 38, 755-786.

Chiricos

Welch

Gertz

(2004). The racial typification of crime and support for punitive measures. Criminology, 42, 359-390.

Costelloe

Chiricos

Gertz

(2009). Punitive attitudes toward criminals: Exploring the relevance of crime salience and economic insecurity. Punishment & Society, 11, 25-49.

10.

Cullen

F. T.

Clark

G. A.

Cullen

J. B.

Mathers

(1985). Attribution, salience, and attitudes toward criminal sanctioning. Criminal Justice and Behavior, 12, 305-331.

11.

Cullen

F. T.

Fisher

B. S.

Applegate

B. K.

(2000). Public opinion about punishment and corrections. In Tonry

(Ed.), Crime and justice: A review of the research (Vol. 27, pp. 1-79).Chicago, IL: University of Chicago Press.

12.

DiIulio

J. J.

Jr. (1997). Are voters fools? Crime, public opinion, and representative democracy. Corrections Management Quarterly, 1, 1-5.

13.

Hanslmaier

(2013). Crime, fear and subjective well-being. European Journal of Criminology, 10, 515-533.

14.

Hartnagel

T. F.

Templeton

L. T.

(2012). Emotions about crime and attitudes to punishment. Punishment & Society, 14, 452-474.

15.

Hogan

M. J.

Chiricos

Gertz

(2005). Economic insecurity, blame, and punitive attitudes. Justice Quarterly, 22, 392-412.

16.

Johnson

(2006). Crime salience, perceived racial bias, and blacks’ punitive attitudes. Journal of Ethnicity in Criminal Justice, 4, 1-18.

17.

King

Maruna

(2009). Is a conservative just a liberal who has been mugged? Punishment & Society, 11, 137-169.

18.

King

R. D.

Wheelock

(2007). Group threat and social control: Race, perceptions of minorities and the desire to punish. Social Forces, 85, 1255-1280.

19.

Leverentz

(2011). Neighborhood context of attitudes toward crime and reentry. Punishment & Society, 13, 64-92.

20.

Messner

S. F.

Baumer

E. P.

Rosenfeld

(2006). Distrust of government, the vigilante tradition, and support for capital punishment. Law & Society Review, 40, 559-590.

21.

Ouimet

Coyle

E. J.

(1991). Fear of crime and sentencing punitiveness. Canadian Journal of Criminology, 33, 149-162.

22.

Ousey

G. C.

Unnever

J. D.

(2012). Racial-ethnic threat, out-group tolerance, and support for punishing criminals. Criminology, 50, 565-603.

23.

Pickett

J. T.

Baker

(2014). The pragmatic American: Empirical reality or methodological artifact? Criminology, 52, 195-222.

24.

Pickett

J. T.

Chiricos

(2012). Controlling other people’s children. Criminology, 50, 673-710.

25.

Pickett

J. T.

Chiricos

Golden

K. M.

Gertz

(2012). Reconsidering the relationship between perceived neighborhood racial composition and whites’ perceptions of victimization risk: Do racial stereotypes matter? Criminology, 50, 145-186.

26.

Rankin

J. H.

(1979). Changing attitudes towards capital punishment. Social Forces, 58, 194-211.

27.

Rich

R. F.

Sampson

R. J.

(1990). Public perceptions of criminal justice policy. Violence and Victims, 5, 109-118.

28.

Roberts

J. V.

(1992). Public opinion, crime, and criminal justice. Crime and Justice, 16, 99-180.

29.

Seltzer

McCormick

J. P.

(1987). The impact of crime victimization and fear of crime on attitudes toward death penalty defendants. Violence and Victims, 2, 99-114.

30.

Simon

(2007). Governing through crime: How the war on crime transformed American Democracy and created a culture of fear. New York, NY: Oxford University Press.

31.

Soss

Langbein

Metelko

A. R.

(2003). Why do white Americans support the death penalty. Journal of Politics, 65, 397-421.

32.

Sprott

J. B.

Doob

A. N.

(1997). Fear, victimization, and attitudes to sentencing, the courts, and the police. Canadian Journal of Criminology, 39, 275-291.

33.

Stack

(2003). Authoritarianism and support for the death penalty: A multivariate analysis. Sociological Focus, 36, 333-352.

34.

Taylor

D. G.

Scheppele

K. L.

Stinchcombe

A. L.

(1979). Salience of crime and support for harsher criminal sanctions. Social Problems, 26, 413-424.

35.

Thomas

C. W.

Foster

S. C.

(1975). A sociological perspective on public support for capital punishment. American Journal of Orthopsychiatry, 45, 641-657.

36.

Tufts

Roberts

J. V.

(2002). Sentencing juvenile offenders: Comparing public preferences and judicial practice. Criminal Justice Policy Review, 13, 46-64.

37.

Tyler

T. R.

Weber

(1982). Support for the death penalty. Law & Society Review, 17, 21-45.

38.

Unnever

J. D.

Cullen

F. T.

(2010). The social sources of Americans’ punitiveness. Criminology, 48, 99-129.

39.

Unnever

J. D.

Cullen

F. T.

Fisher

B. S.

(2007). A liberal is someone who has not been mugged’’: Criminal victimization and political beliefs. Justice Quarterly, 24, 309-334.

40.

U.S. Bureau of the Census. (1994). County and city data book 1994. Washington, DC: U.S. Government Printing Office.

41.

U.S. Department of Justice, Federal Bureau of Investigation. (2000). Uniform crime reporting data [United States]: County-level detailed arrest and offense data, 1998 [Computer file]. 2nd ICPSR ed. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [producer and distributor].

42.

Welch

Payne

A. A.

Chiricos

Gertz

(2011). The typification of Hispanics as criminals and support for punitive crime control policies. Social Science Research, 40, 822-840.

43.

Sun

I. Y.

(2011). Support for the death penalty: Chinese and American college students compared. Punishment & Society, 13, 354-376.